Advances in Japanese Business and Economics 27
Akira Shimizu
New Consumer
Behavior
Theories
from Japan
Advances in Japanese Business and Economics
Volume 27
Editor-in-Chief
Ryuzo Sato, C.V. Starr Professor Emeritus of Economics, Stern School of Business, New York
University, New York, NY, USA
Senior Editor
KAZUO MINO
Professor Emeritus, Kyoto University; Professor of Economics, Doshisha University
Managing Editors
HAJIME HORI
Professor Emeritus, Tohoku University
HIROSHI YOSHIKAWA
Professor Emeritus, The University of Tokyo; President, Rissho University
TOSHIHIRO IHORI
Professor Emeritus, The University of Tokyo; Professor, GRIPS
Editorial Board
YUZO HONDA
Professor Emeritus, Osaka University; Professor, Osaka Gakuin University
JOTA ISHIKAWA
Professor, Hitotsubashi University
KUNIO ITO
Professor Emeritus, Hitotsubashi University
KATSUHITO IWAI
Professor Emeritus, The University of Tokyo; Visiting Professor, International Christian University
TAKASHI NEGISHI
Professor Emeritus, The University of Tokyo; Fellow, The Japan Academy
KIYOHIKO NISHIMURA
Professor Emeritus, The University of Tokyo; Professor, GRIPS
TETSUJI OKAZAKI
Professor, The University of Tokyo
YOSHIYASU ONO
Professor, Osaka University
JUNJIRO SHINTAKU
Professor, The University of Tokyo
MEGUMI SUTO
Professor Emeritus, Waseda University
EIICHI TOMIURA
Professor, Hitotsubashi University
KAZUO YAMAGUCHI
Ralph Lewis Professor of Sociology, University of Chicago
Advances in Japanese Business and Economics (AJBE) showcases the work of
Japanese and non-Japanese scholars researching the Japanese economy and
Japanese businesses. Published in English, the series highlights for a global
readership the unique perspectives of Japan’s most distinguished and emerging
scholars of business and economics. It covers research of either theoretical or
empirical nature, in both authored and edited volumes, regardless of the
sub-discipline or geographical coverage, including, but not limited to, such topics
as macroeconomics, microeconomics, industrial relations, innovation, regional
development, entrepreneurship, international trade, globalization, financial markets,
technology management, and business strategy. At the same time, as a series of
volumes written by Japanese and non-Japanese scholars studying Japan, it includes
research on the issues of the Japanese economy, industry, management practice,
and policy, such as the economic policies and business innovations before and after
the Japanese “bubble” burst in the 1990s.
AJBE endeavors to overcome a historical deficit in the dissemination of Japanese
economic theory, research methodology, and analysis. The volumes in the series
contribute not only to a deeper understanding of Japanese business and economics
but to revealing underlying universal principles.
Overseen by a panel of renowned scholars led by Editor-in-Chief Professor
Ryuzo Sato, AJBE employs a single-blind review process in which the
Editor-in-Chief, together with the Managing Editors and specialized scholars
designated by the Editor-in-Chief or Managing Editors, rigorously reviews each
proposal and manuscript to ensure that every submission is a valuable contribution
to the global scholarly readership.
All books and chapters in AJBE are indexed in Scopus.
More information about this series at http://www.springer.com/series/11682
Akira Shimizu
New Consumer Behavior
Theories from Japan
123
Akira Shimizu
Faculty of Business and Commerce
Keio University
Minato-ku, Japan
ISSN 2197-8859
ISSN 2197-8867 (electronic)
Advances in Japanese Business and Economics
ISBN 978-981-16-1126-1
ISBN 978-981-16-1127-8 (eBook)
https://doi.org/10.1007/978-981-16-1127-8
© Springer Nature Singapore Pte Ltd. 2021
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Preface
Modern marketing was first introduced to Japan from the United States in the
1950s. Especially as consumers’ disposable income increased with Japan’s rapid
economic growth and they were able to choose products according to their own
preferences, the concept of marketing was seen as important in differentiating one’s
products from those of other companies, and was adopted by many companies.
Japanese consumer electronics and automobile manufacturers have been successful
in their overseas expansions by analyzing consumer preferences and product distribution channels in detail, precisely because they practice marketing. At universities, marketing is considered a core subject, along with business administration
and accounting, in faculties such as business administration and commerce, and
many young people are studying marketing before entering the workforce.
Thus, in Japan, marketing has penetrated companies and consumers in just less
than 70 years since it was introduced, and it has achieved a great deal of success
both domestically and internationally. However, Japanese marketing research has
rarely been introduced to foreign countries or attracted the attention of foreign
companies and researchers. This is in contrast to business administration, which,
like marketing, was introduced from the United States and is now being studied
under the title “Japanese-style management,” and is also attracting attention
overseas.
In the course of this research, I have learned that there are many phenomena in
the Japanese market that cannot be explained by foreign theories and that it is
possible to construct a new marketing theory that takes advantage of the characteristics of the Japanese market, especially Japanese consumers.
With this in mind, this book was originally published by Chikura Shobo in
Japanese and has been modified to make it suitable for use overseas. I am very
grateful to Chikura Shobo for his willingness to allow this book to be published in
English. During the writing of the English version of the book, the professors on the
editorial board gave me specific advice and helped me to elevate the book to greater
sophistication. I am grateful to them for doing so and thereby enabling the book to
be accepted overseas.
v
vi
Preface
I am also very grateful to Ryuzo Sato, Editor-in-Chief of the Advances in
Japanese Business and Economics book series, for giving me the opportunity to
write this book. No matter how much one has proposed an original theory in Japan,
it will remain unnoticed by overseas researchers if it has not been published in
English. I would like to repay Professor Sato for his kindness by using this book as
an opportunity for me to play an active role overseas.
Finally, I am grateful to Juno Kawakami of Springer for her help in editing this
book.
Tokyo, Japan
Akira Shimizu
Contents
1
2
3
Japanese Consumers and Media Usage . . . . . . . . . . . . . . . .
1.1 Media Usage Among Consumers . . . . . . . . . . . . . . . . .
1.2 The Way Consumers Use Different Media for Different
Purposes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3 Changes in Consumer Clusters Over Time . . . . . . . . . .
1.4 Conclusion and Future Implication . . . . . . . . . . . . . . . .
......
......
1
1
......
......
......
6
8
13
Evolution of the Comprehensive Decision-Making Process:
Emergence of Outspoken Consumers . . . . . . . . . . . . . . . . . . . .
2.1 Stimulus-Response Model and Information Processing
Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 Significance of a Comprehensive Model of Decision
Making That Considers Outspoken Consumers . . . . . . . . . .
2.3 Comprehensive Model that Considers the Effect of Word
of Mouth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4 Demonstration of the AISAS® Theory . . . . . . . . . . . . . . . .
2.5 Situations in Which the Customer Would Like to Influence .
2.6 Media Contact that Leads to Satisfaction and Information
Sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.7 Conclusion and Future Implication . . . . . . . . . . . . . . . . . . .
Measuring the Impact of a Blog: Quantitative and Qualitative
Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1 Review of Previous Studies on Consumer Behaviors
Related to Information on the Internet . . . . . . . . . . . . . . . .
3.2 Quantitative Measurement of the Effect of Blogs . . . . . . . .
3.3 Qualitative Measurement of the Effect of Blogs . . . . . . . . .
3.4 Conclusion and Future Implication . . . . . . . . . . . . . . . . . . .
...
17
...
17
...
21
...
...
...
25
27
29
...
...
31
36
...
41
.
.
.
.
41
45
48
54
.
.
.
.
.
.
.
.
vii
viii
Contents
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
57
57
61
62
63
67
72
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
75
76
78
80
82
93
6
Emergence of Communication-Oriented Consumers . . . . . . . . .
6.1 The Idea of CRM and Its Development . . . . . . . . . . . . . . .
6.2 What Is a “Communication-Oriented Consumer”? . . . . . . . .
6.3 Characteristics of Communication-Oriented Consumers . . . .
6.4 Purpose of Managing Communication-Oriented Consumers .
6.5 Conclusion and Future Implication . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
97
98
102
104
110
113
7
Research on Uncertain Listeners . . . . . . . . . . . . . . . . . . . .
7.1 Lifestyle Research Is Flourishing . . . . . . . . . . . . . . . .
7.2 Relationship Between Lifestyles and Brands . . . . . . . .
7.3 Possibilities for Constructing Segments Using Brands .
7.4 Data Used to Create the Segments . . . . . . . . . . . . . . .
7.5 Segments and the Profiles . . . . . . . . . . . . . . . . . . . . .
7.6 Segments and Brand Evaluations . . . . . . . . . . . . . . . .
7.7 Conclusion and Future Implication . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
117
118
121
122
124
125
131
136
8
Studies on Connoisseurs . . . . . . . . . . . . . . . . .
8.1 Demand Forecast for New Products . . . . .
8.2 Studies on Trendsetters . . . . . . . . . . . . . .
8.3 Mechanism of the Study on Connoisseurs
8.4 Analysis Results . . . . . . . . . . . . . . . . . . .
8.5 Profile of Connoisseurs . . . . . . . . . . . . . .
8.6 Conclusion and Future Implication . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
141
142
145
150
152
158
160
9
Brand Rating in the Age of Information Sharing . . . . . . .
9.1 Setting of the Kikimimi Panel . . . . . . . . . . . . . . . . . .
9.2 Brand Rating Using the Kikimimi Panel . . . . . . . . . .
9.3 Effect of Product Line Extension and Its Measurement
9.4 Conclusion and Future Implication . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
167
167
170
178
185
4
Studies on Commitment . . . . . . . . . . . . . . . . . . . . . .
4.1 Current State of Retailing in Supermarkets . . . . .
4.2 Loyalty and Commitment . . . . . . . . . . . . . . . . .
4.3 Explanation and Analysis of the Data . . . . . . . .
4.4 Analysis on Customers of Long Sellers . . . . . . .
4.5 Chronological Change in Affective Commitment
4.6 Conclusion and Future Implication . . . . . . . . . . .
5
Mechanism of Attitude Formation for Consumers Who
Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1 The ELM Concept . . . . . . . . . . . . . . . . . . . . . . . .
5.2 Direction of the ELM Application . . . . . . . . . . . . .
5.3 Explanation of the Analyzed Data . . . . . . . . . . . . .
5.4 Analysis Results . . . . . . . . . . . . . . . . . . . . . . . . . .
5.5 Conclusion and Future Implication . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Convey
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Contents
10 A New Decision-Making Process—A Circulating-Type
Communications Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.1 What Has Been Clarified in This Chapter Up to This Point .
10.2 A New Theory—Toward Circulating-Type Marketing . . . . .
10.3 Overview and Results of the Experiment Survey . . . . . . . .
10.4 Conclusion and Future Implication . . . . . . . . . . . . . . . . . . .
ix
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
189
190
193
196
208
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
About the Author
Akira Shimizu joined the Faculty of Commerce at Keio University as a Professor
of Marketing in April 2009. Before joining Keio University, Prof. Shimizu worked
at Meiji Gakuin University as an assistant professor from 1991 to 1993, an associate
professor from 1994 to 1999, and a professor from 2000 to 2009. He received his
B.S. (1986) and M.S. (1988) in marketing from Keio University, and also his Ph.D.
in marketing from the Graduate School of Commerce at Keio University (2004).
Professor Shimizu’s research focuses on consumer behavior and marketing
strategy, with particular interest in the decision-making process of consumers, sales
estimation using innovative consumers, and applying consumer behavior theories
for marketing strategy.
He received the Best Book Award from the Japan Society of Marketing and
Distribution in 2005, and also the Best Paper Award at the Global Marketing
Conference in 2012 and 2018.
Professor Shimizu was the former Chief Editor of the Journal of Marketing and
Distribution, published by the Japan Society of Marketing and Distribution, and the
former President of the Japan Association for Consumer Studies.
xi
Chapter 1
Japanese Consumers and Media Usage
The environment in which we live, particularly the way we access the media and
obtain information, has changed considerably since the advent of the Internet. It is
quite evident that this environmental change could significantly affect the formulation
of marketing strategies in terms of targeting consumers and their behaviors. Using
various types of data, this chapter will explain the facts associated with and changes
in the way Japanese consumers access the media.
1.1 Media Usage Among Consumers
First, an overview of the current usage of four existing media, i.e., TV, newspaper,
radio, and magazine, will be provided.
The National Media Access and Evaluation Survey, a large study conducted once
every two years by the Japan Newspaper Publishers & Editors Association (Nihon
Shinbun Kyokai or NSK), indicates that the most frequently used media by Japanese
people is TV.1 Those who watch TV everyday account for 86.6% of the overall
population. The proportion is the lowest at 73.2% among those in their 20s, but
is higher at 96.4% among those in their 70s. The average daily viewing time was
201.6 min for the overall population. The average viewing time, which is the lowest
at 149.9 min among those aged 15–19, increases as people become older and is
the highest at 247.8 min among those in their 70s. However, when the average is
drawn regardless of age, people watch TV for about three hours a day. Since the
survey conducted four years prior to this one showed that 89.1% of the respondents
watched TV every day and the average viewing time was 209.0 min, we can say that
there are still several people who watch many hours of TV.
Although 49.1% of the respondents said that they read newspapers (morning
edition) every day, the percentage was extremely low at 13.2% among those aged
15–19. The survey conducted six years ago showed that 30.0% among those aged
15–19 read newspapers every day, indicating that younger people are moving away
© Springer Nature Singapore Pte Ltd. 2021
A. Shimizu, New Consumer Behavior Theories from Japan, Advances in Japanese
Business and Economics 27, https://doi.org/10.1007/978-981-16-1127-8_1
1
2
1 Japanese Consumers and Media Usage
from newspapers. Conversely, the percentage of those who subscribe to newspapers
every day is 86.2% among those in their 70s, showing a large variance in terms of age
group. Similarly, since 62.7 and 60.7% of them were reading newspapers every day
according to the survey conducted four years ago and six years ago, respectively, it
can be said that the percentage of newspaper readership is declining rapidly, mainly
among younger people.
For magazines, 35.2% of the respondents were not interested in reading them,
and the percentage was extremely high at 55.1% among those aged 15–19. The
survey conducted six years ago showed that 21.2% of all respondents and 22.5% of
those aged 15–19 did not read magazines. Similar to newspapers, it can be said that
the percentage of magazine readership is also declining rapidly, mainly among the
younger generation.
The overall average time spent listening to the radio is 94.8 min per day. Similar
to newspapers, the difference by age group is large, as the average is 56.3 min among
those in their 20s and 103.2 min among those in their 60s. Since the overall average
in the survey conducted six years ago was 96 min, the time spent on listening to a
radio has not changed much.
A trend similar to the one described above was observed in a survey conducted
by Yomiko Advertising Inc.2 According to this survey, only 1.9% of the respondents
do not watch TV on weekdays; in other words, almost everyone is watching TV.
Moreover, 43.8% said that they watch TV for more than three hours per day. The
data on viewing time by age group is also similar to the results obtained by the NSK:
51.6% of men in their 70s and 68.9% of women in their 70s said that they watch
TV for more than three hours per day on average. These figures almost match those
of a similar survey conducted four years ago. With regard to newspapers, 40.2%
people said that they do not read newspapers; this figure is lower than the 30.2%
figure recorded four years ago. This factor differs in the NSK survey. However, the
data by age group shows that while 13.9% of men and 16.1% of women in their
70s do not read newspapers, the figures are 60.8 and 69.6% for men and women in
their 20s, respectively, indicating that younger people are largely moving away from
newspapers. Those who do not read magazines account for 56.2% of the population.
Around 61.1% of men in their 70s do not read magazines, while 57.5% of men in
their 20s do not read magazines. Considering the fact that 43.0% were those who do
not read magazines according to the 2012 survey, we can say magazine readers are
definitely decreasing as well. The overall percentage of those who do not listen to
radio is 63.9%. Things have not changed much for radio as, according to the 2012
survey, 61.4% were those who do not listen to it.
Although there are slight differences in the figures cited in the NSK survey and
the Yomiko survey, the overall trend in the demographic attributes of media users
is similar. In other words, TV is a medium watched by many people regardless of
age and Japanese people are watching it about three hours per day on average. As
for time spent listening to radio, it has not changed much in the past several years.
The figure that is declining greatly is the proportion of those who read newspapers
and magazines. Although they are still well read among older people, a considerable
percentage of younger people are moving away from newspapers. Likewise, the
1.1 Media Usage Among Consumers
3
percentage of people who do not read magazines has increased a lot in the past several
years. In other words, paper-based media is declining mainly because younger people
have stopped reading them.
Next, the study will explore in detail the reality of Internet usage.
According to a white paper on information communication published by the
Ministry of Internal Affairs and Communications (MIC), the Internet user population
was over 100 million at the end of 2016, with a penetration rate of 83.5%.3 Since
the Internet penetration rate was only 9.2% when the same survey was conducted
in 1997, we can see that the penetration rate has grown significantly during the past
two decades. The breakdown of devices used for accessing the Internet is as follows:
Personal Computer (PC) 58.6%, smartphone 57.9%, and tablet 23.6%. While the
figures for PC and smartphone are vying with each other, smartphone has by far
shown a higher growth rate and is expected to surpass the PC figure by next fiscal
year. The Internet usage rate by age group exceeds 95% among those aged 13–49, is
93.0% among those in their 50s, 83.3% among those in their early 60s, and 53.6%
among those in their 70s. This shows that Internet usage is quite common among
people in their early 60s and below, and is lower among people in their late 60s and
above. The Internet usage has not changed much among those aged 49 or younger
compared to five years ago; however, we can see it has dramatically increased among
older segments, considering it was 70.1% among those in their early 60s and below
and 39.2% among those in their early 70s.
According to the NSK survey, 54.9% are using the Internet every day, which is
an increase of more than 10% points over 44.4% five years ago. In line with the
white paper on information communication, while 85.8% of those in their 20s are
using the Internet every day, the rate is 54.8 and 36.5% among those in their 50s
and 60s, respectively. We can see that the variance by age group is large and the
Internet is a medium that is regularly used by younger people. Likewise, the Yomiko
Survey shows that 77.1% of men and 71.3% of women use the Internet on weekdays.
Furthermore, the corresponding proportion is 94.8 and 94.6% for men and women in
their 20s and merely 59.2 and 39.6% for men and women in their 60s. Considering
this in conjunction with the survey conducted by the MIC, we can say that the Internet
has not become a part of the daily lives of older people, even though the penetration
rate among them is high.
The Internet is gaining attention not only for its penetration rate but also for
the overwhelming volume of information it provides.4 The Institute for Information
and Communications Policy of the MIC measured and compared the volume of
information each media provided in FY 2001 to the volume in FY 2009. Albeit
being slightly old data, according to their report, while broadcasting and postal mail
have grown by 1.98 times and 2.57 times, respectively, the volume of information
offered by the Internet has actually grown by 71.63 times in the past decade or
so. In terms of the actual volume of consumption, the Internet has grown by 2.37
times while the broadcast and postal media have only grown by 1.04 times and 0.90
times, respectively. This indicates that the actual consumption of information on
the Internet—the level of consumers’ dependency on the Internet—is increasing in
addition to the amount of information circulated by the Internet relative to other
media.
4
1 Japanese Consumers and Media Usage
Next, of all information on the Internet, this study will examine the ones posted
by individuals. In its white paper on information, the MIC looked at six SNSs that
can be compared over the years, namely LINE, Facebook, Twitter, mixi, Mobage,
and GREE, and studied the changes in their usage rates. Based on this, it became
clear that the percentage of people who use one of these six SNS s has increased
from 41.4% in 2012 to 71.2% in 2016 in conjunction with the spread of smartphones.
Looking at it by age group, 81.3% of those in their 20s were already using some sort
of SNS as of 2012 and, as the figure reached 97.7% in 2016, it became clear that
almost everyone was using SNS. The growth rate was higher among those in their
40s and 50s; while their usage rate was 37.1% and 20.6%, respectively, it had risen
to 78.3% and 60.8%, respectively, by 2016. SNS, therefore, is no longer just a tool
for young people.
Among those six SNSs in question, the one for which the usage rate increased
the most from 2012 to 2016 was LINE: its overall rate rose from 20.3 to 67%. It
was followed by Facebook (16.6–32.3%) and Twitter (15.7–27.5%). While LINE is
used evenly by all age groups, from those in their 10s–40s, Facebook is most used
by those in their 20s and 30s and Twitter is most used by those in their 10s and 20s.
The remaining three SNSs—mixi, Mobage, and GREE—are not used much today.
Instead, the usage rates have increased for SNSs such as Google+ and Instagram. As
of 2016, their overall usage rates are 26.3% and 20.5%, respectively. It is characteristic that the usage rate of Instagram is remarkably high among those in their 20s
(45.2%) while Google+ is used regardless of age.
Changes in the penetration rate of each media over time are shown in Figs. 1.1
and 1.2. These charts are cited from the “Study on Communication Touchpoints”
included in the Media Value Survey conducted by Dai Nippon Printing Co. every
year since 2001. Figure 1.1 shows the changes in the media usage rates and Fig. 1.2
shows the changes in channel usage rates. From Fig. 1.1, the percentage of Internet
users has increased from 41% in 2002 to 95% in 2016 and the usage of mobile phone
Internet has grown at an extremely rapid rate: the usage rate of mobile phone Internet
has increased from 31% in 2003 to 92% in 2016 and that of SNS has increased from
50% in 2010 to 75% in 2017. Meanwhile, comparing the years 2002 and 2016 in
terms of the four traditional media types, the usage rate has not declined dramatically
for TV, although the usage rate of radio, newspapers, and magazines has declined by
10% points. In particular, the figure declined drastically for newspapers as it reached
60% in 2016 even though 90% of people were accessing newspapers in 2002. In
other words, people are still using these media except newspapers at a high rate even
though Internet usage rate has risen, and it is unlikely that they will be replaced by
the Internet. From the previously discussed results of the NSK and Yomiko studies,
we can see that although the time spent on these three media has definitely declined,
the drop is small. Furthermore, we can conclude that the Internet is spreading today
because people are using it along with other media or spending time on the Internet
by redirecting other spare time, and not because they have stopped using other media
and redirected that time to use the Internet. However, as far as the younger generations
are concerned, the decline in the newspaper and radio usage has been drastic, and it
seems that the Internet is in fact replacing these media.
1.1 Media Usage Among Consumers
5
Fig. 1.1 Changes in media usage rate
Next, we can see from Fig. 1.2 of Internet-related channels has gradually increased
since 2001 when the survey began. Online shopping has increased from 15% in
2001 to 49% in 2016 and mobile shopping has increased significantly from 7% in
2001 to 48% in 2016. Since the usage rate has declined drastically for department
stores and specialty retailers, from 83% in 2001 to 65% in 2016 and from almost
98% in 2001 to 75% in 2016, respectively, while supermarkets and convenience
stores maintained high usage rates of more than 90%, it seems that the places for
purchasing the kind of products sold at department stores and specialty retailers have
shifted to Internet-related channels.
As described, the spread of the Internet has increased its importance not only as
a means of collecting information but also as purchasing channels.
6
1 Japanese Consumers and Media Usage
Fig. 1.2 Changes in channel usage rate
1.2 The Way Consumers Use Different Media for Different
Purposes
As explained, the Internet usage rate has increased exponentially in the past 15 years,
we can surmise that the Internet will be an alternative to TV, newspapers, radio, and
magazines. Though the time spent on these traditional media has decreased, their
usage rates have not decreased drastically except among younger generations. This
is because many consumers are choosing a different media for different purposes.
According to a joint study conducted by the author with NSK in 2006 regarding
the way the Internet and traditional media are used differently according to the
purpose, although people use the Internet when they “learn more about the products
and other things that I become interested in (61.4%),” “obtain information that is
useful for my hobby (67.8%),” and so on, what prompts them to notice a product is
newspapers (40.2%), company websites (35.6%), TV (31.3%), and consumer review
sites (19.8%). Back then, the Internet was a media for researching about a product;
although the information disseminated by manufacturers was used at the time of
1.2 The Way Consumers Use Different Media for Different Purposes
7
Table 1.1 Consumer decision-making process and media contact
Utilization ratio
Product
recognition and
interest
Searching about
the product
Considering the
purchase
Newspaper articles 58.1
20.9
7.9
6.3
Newspaper ads
25.3
10.5
8.6
36.8
TV programs
70.8
38.5
8.9
7.6
TV commercials
60.6
55.7
10.6
11.8
Radio
17.8
6.7
1.4
1.2
Magazines
21.3
18.8
9.5
6.8
Company websites 10.5
6.8
25.3
14.7
9.1
8.1
11.5
15.1
11.6
Banner ads
Consumer review
sites
23
6.1
18.3
product recognition, the influence of SNS was relatively small.5 Newspapers and
TV still had the influence during the stage of product recognition and the Internet
was often used for learning more about those products.
A summary of the different uses of media in consumers’ each decision-making
process, which was prepared based on the aforementioned 2011 NSK survey, is
provided in Table 1.1. It indicates that while newspaper articles, newspaper ads, TV
programs, and TV commercials are more likely to prompt product recognition and
interest, when it comes to the stage for searching about the product and considering
the purchase, the use of Internet sites such as company websites and consumer review
sites increases. Although the direct comparison of the data is not possible because
the survey methods differ, the overall structure has not changed much since 2006.
However, the usage rate of consumer review sites, as a medium that prompts interest,
has reached the same level as that of the Internet for researching about products and
referencing during purchase, indicating that people are beginning to use the Internet
as a medium for finding new products rather than just for researching.
Table 1.2 shows the same summary for those aged between 15 and 29. It indicates
that the effect of TV programs and commercials on product recognition and interest
is about the same as the overall data; however, the role of newspaper articles and ads
is very low compared to the overall data and the role of magazines and consumer
review sites is very high instead. Specifically, the usage of consumer review sites is
very high in all decision-making stages among younger people, indicating that such
sites have become very important sources of information for them.
Based on these results, we can see in general that even though the relative importance of the four traditional media types has declined, consumers are using different
media for different purposes and actually consider TV, newspapers, and magazines
unnecessary for those aged over 30. However, it has also become clear that the
younger people are more dependent on SNS among all the Internet media. Also,
we should note that the use of consumer review sites as a media to first notice a
8
1 Japanese Consumers and Media Usage
Table 1.2 Consumer decision-making process and contact (under 30)
Utilization ratio
Newspaper articles 29.9
Newspaper ads
17.7
Product
recognition and
interest
8.9
11
Searching about
the product
Considering the
purchase
3.9
2
4.2
3.2
TV programs
73.3
46.1
13.5
7.7
TV commercials
67.6
58.4
12
9.9
Radio
11.3
Magazines
32
3.8
2.1
1.4
31.4
14.6
10.1
Company websites 15.2
11.6
34.7
20.6
Banner ads
19.8
14.7
15.3
8.7
Consumer review
sites
42.3
32.3
45.2
37.4
product has been increasing recently across all aged segments. As is clear from the
above summary, instances in which young people use SNS are increasing. In particular, since the survey was conducted 10 years ago, the use of the Internet and social
networking sites is likely to be on the rise among a larger number of generations
today. Here, the big difference is the fact that they are using SNS not only to search
for information but also to find a new product.
1.3 Changes in Consumer Clusters Over Time
As described earlier, the media usage among consumers has changed due to the
emergence of the Internet. However, all consumers are not exclusively using the
Internet-based media; the existing media are still used in parallel, especially over
30 years old. It is probably reasonable to think that the emergence of the Internet has
increased the patterns for using different media for different purposes.
Therefore, by using the data collected in the aforementioned Media Value Survey
and dividing the consumers into clusters, I have analyzed how consumer categories
based on the way they use media change over time. The Media Value Survey investigates the usage rates and average days of use per four weeks for 18 media. Clusters
were created based on the usage rate of these 18 media. Figures 1.3, 1.4, and 1.5
show the survey results for FY 2002, FY2012 and FY 2017, respectively.6 Of these,
the usage rate of the network media—or the Internet-based media—is shown on the
vertical axis, while the usage rate of the mass media is shown on the horizontal axis
and the number of people belonging to a given cluster is shown by the size of the
circle.
For FY 2002, when the Internet penetration rate was about 50%, six clusters
were created. The first “network media” type is a cluster of consumers with a high
Internet-based media usage rate and a high rate of information access, accounting
1.3 Changes in Consumer Clusters Over Time
9
usage rate of the network media
network media
mobile phone and email
printed media
Avg.
media indifference mass media
knformation-active
usage rate of the mass media
Fig. 1.3 Clusters based on the usage rate of 18 media (2002)
usage rate of the network media
social media-centric
all media
mobile phone-centric
PC-centric
actual word-of-mouth
mass media
and PC
mass media
direct media
usage rate of the mass media
Fig. 1.4 Clusters based on the usage rate of 18 media (2012)
for 24.7% of the overall population. This is a cluster of “most up-to-date consumers”
so to speak; they have no resistance to purchasing products on the Internet and are
interested in new tools such as mobile devices. The “mobile phone and email” type
is a cluster of consumers whose heavy use of mobile phone stands out of all Internetbased media usage. They are characterized by not being interested in other media. In
terms of age group, many teenagers belong to this cluster, which accounts for 4.6% of
the overall population. Following these two clusters in terms of high network media
usage is the “printed media” type. Currently, they often tend to use print media such
as magazines and books. Because they perceive the current information environment
10
1 Japanese Consumers and Media Usage
usage rate of the network media
Internet-centric
PC promotion media
smartphone-centric
PC-centric
mobile
promotion media
real word-of-mouth
newspaper-centric
low media engagement
usage rate of the mass media
Fig. 1.5 Clusters based on the usage rate of 18 media (2017)
to be inadequate and do not spare the money to use the Internet or mobile phones, they
can be considered would-be “network media” type consumers. This cluster accounts
for 5.1% of the overall population. The “mass media” type has high rates of using
traditional media such as TV, radio, and newspapers. The usage rate of Internet-based
media among these people is <50%. The cluster, characterized by all generations of
people without any particular age group bias, accounts for 16.7% of the overall
population. The “information-active” type currently uses almost no Internet-based
media; however, they have a high level of information access as well as intention
to use the Internet. The cluster accounts for 18% of the overall population and its
percentage of woman is as high as 61%. The last cluster, “media indifference” type,
has a low level of interest in the Internet as well as information access and accounts
for 31% of the overall population. The average age is also the highest among all these
six clusters.
As also shown in the figure, the FY 2002 survey indicated that the “network media”
types who use Internet-based media are also using mass media in parallel at a high
rate; therefore, they can be considered a cluster with a large amount of information.
As of 2002, the usage of network-based media was not high among other clusters even
though their mass media usage was high. However, since the “printed media” types
were dissatisfied with the current information environment and favorable toward the
Internet, and the “information-active” types were also highly capable of collecting
information, it seems these clusters adapted to use the Internet over time. When
thinking of it this way, the reason why the usage rates are declining for magazines
and newspapers can be understood.
In the 2012 survey, the population was first divided into two major clusters based
on whether the individual actively uses the Internet or not and each of those clusters
was then further divided into four clusters.
1.3 Changes in Consumer Clusters Over Time
11
First, those who actively use the Internet are divided into four clusters: the “all
media” type, “PC-centric” type, “mobile phone-centric” type, and “social mediacentric” type. The “all media” type includes people who are the most up-to-date;
this is a cluster of people who utilize not only information on the Internet but also
any useful information on media such as TV and flyers. Accounting for 9.3% of
the overall population, their average age is 30 and includes more men than women
with the male-female ratio of 6:4. The “PC-centric” type includes many married
men and women in their 40s and 50s, with the highest average age (41.6) of the
four clusters of consumers who use the Internet. They are characterized by the way
they obtain shopping information such as online supermarkets, online flyers, and
coupons, and they shop efficiently while comparing. They account for 12.8% of the
population. The “mobile phone-centric” type includes a large percentage of women
at the female-male ratio of 7:3, most of which is accounted by women in their 10s
and 20s. Their average age is also young at 25. They tend to collect shopping-related
information and communicate only by mobile phones without using PC most of
the time. Around 7.5% of the overall population falls under this cluster. While the
“social media” type also includes many young people just like the “mobile phonecentric” type and the average age is the same at 25, the male-female ratio does not
change. These people actively post information on SNS, microblog, video sharing
sites, etc. and seek to connect with friends, but are not active in shopping. However,
they frequently purchase contents such as downloadable music. The cluster accounts
for 10.1% of the overall population.
Next, those who do not actively use the Internet were divided into four types:
mass media, direct media, mass media and PC, and actual word-of-mouth. The “mass
media” type mainly consists of men and women in their 50s and 60s. There are slightly
more men in terms of composition and the average age is 47.3. Their media usage
is generally low except for TV and newspaper. They are also uninterested in matters
other than health and post-retirement years. Around 24.5% of the entire population
falls under this cluster. The “direct media” type values information touch points
such as direct mails, flyers, pamphlets, and mail order catalogs. Many of them are
homemakers, and 11.1% of the overall population falls under this cluster. It consists
mainly of housewives in their 50 and 60s and three-quarters of the population are
women. Their average age is also the highest at 52. They are highly likely to frequent
drugstores and supermarkets. In contrast, two-thirds of the “mass media & PC” type
are men and they are highly likely to use TV, newspaper, and car card advertising.
Their average age is 43. Around 14.4% falls under this cluster. The average age
of “in-person word-of-mouth” type is 36 and two-thirds are women. Rarely using
media other than actual word-of-mouth, these are people who are reluctant in posting
comments. They value reputation and opinion coming from people around them and
purchase what is recommended by their friends, family, and store clerks. Around
10.2% of the overall population falls under this cluster.
As it became clear here, the 2012 survey showed that (i) Internet users can be
divided into four clusters, and there are digital divides among them; (ii) the clusters
are divided based on whether they post information or not and whether the device
is a mobile phone or PC; (iii) the “social media” type whose purpose is to connect
12
1 Japanese Consumers and Media Usage
with friends rather than actively collecting information even when using Internetbased media has emerged; and (iv) there are clusters like the “direct media” type
of consumers who actively try to obtain information even with real information
sources. In other words, we can say that a new cluster that is quite different from the
conventional stereotype (i.e., those who use the Internet are more likely to actively
seek information) is emerging now.
It is noteworthy that those referred to as the “mass media” type and the “real
word-of-mouth” type clusters, which have not shifted emphasis on Internet-based
media as of 2012, account for 34.7% of the overall population. In addition, there is
a cluster named “direct media type” which consumes a lot of information although
it values information from flyers more than the Internet for shopping. These individuals account for 11.1% of the population. Based on this, we see that although
the Internet penetration rate had reached 90% according to the Media Value Survey,
it is incorrect to assume that everyone is benefiting from it and that the amount of
information consumption increases through the use of the Internet. We can say that
the development of the Internet has diversified the way of accessing information,
widening the digital divide.
The 2017 survey featured five clusters (the “Internet-centric” type, “mobile
promotion media” type, “PC promotion media” type, “PC-centric” type, and
“smartphone-centric” type) based on the state of Internet-based media usage. Meanwhile, those who do not use the Internet much were divided into three clusters (the
“real word-of-mouth” type, “low media engagement” type, and “newspaper-centric”
type). It is characteristic that the Internet-based media was further divided and those
who use the Internet-based media are becoming older.
First, 13.6% of the overall population falls under the “Internet-centric” type. The
average age is 37 and men account for approximately two-thirds. These are people
who obtain information by fully utilizing PC and mobile devices and are characterized by being well-informed about information from brick-and-mortar stores. The
“mobile promotion media” type accounts for 10.9% and, unlike the “Internet-centric”
type, 70% of those who fall under this cluster are women. The average age is 40 and
30% are from dual income households, raising children. Around 90% of them read
promotional media such as magazines and direct mails and they also check information on flyers and brochures on their smartphones. Their online banking usage is
also high. Around 40% of the “PC promotion media” type is accounted by those in
their 60s. The male-female ratio is almost half and the average age is 50. They shop
at brick-and-mortar stores such as grocery stores and drugstores, are highly likely
to use a PC than smartphone, and collect information by using shopping sites and
manufacturer’s sites. Around 7.6% of the overall population falls under this cluster.
The “PC-centric” type accounts for 8.4% of the overall population. The average age
is 47 and 70% are men. Most of them are businessmen in their 50s and 60s who read
newspapers, online news, and product review sites by using PC and tablet. Their
SNS usage rate is low and they use the Internet mainly for collecting information.
Around 80% of them are also purchasing products by using a PC. The “smartphonecentric” type accounts for 10.6%. Their average age is 31, 40% of them are women
in their 20 and 30s, and the cluster mainly consists of students and office workers.
The percentage of women is 70%. These people use a smartphone to collect a wide
1.3 Changes in Consumer Clusters Over Time
13
range of product information on websites such as shopping sites and manufacturer’s
sites and their usage of newspapers and flyers is low. They also do not really shop
by using smartphones and are less likely to shop at brick-and-mortar stores such as
supermarket and drugstores.
In contrast, the three clusters with a relatively low rate of Internet use are characterized as follows. First, 9.1% falls under the “real world-of-mouth” type. The
average age is 33 and women account for 60% while students account for one-fourth.
Although chatting with friends and family is their main source of information, half of
them are also interacting on the Internet by posting comments on SNS, etc. They are
characterized by referring to word-of-mouth information from people around them,
such as friends and family, as well as information obtained from sales clerks at the
storefront when shopping. The average age of the “low media engagement” type is
34. About 20% are distributed to each age group from the 10 to 30s and 19.3% of the
overall population falls under this cluster. This is an information-passive segment
that does not access media other than TV; although about half of them use SNS, they
mainly browse and rarely post information themselves. The “newspaper-centric” type
accounts for 20.5% of the overall population. The average age is 55 and half of them
are men and the other half are women. While almost 100% of them use newspaper
and TV and 80% also use flyers, they do not use other media. In particular, the rate
of their Internet usage is the lowest at 81.5%.
Considering that the 2017 survey indicated the Internet usage rate hit 81.5% even
for the “newspaper-centric” type with the lowest Internet usage rate, it has become
clear that the number of people who do not use a smartphone or PC reduced a
lot and the Internet has widely spread as an infrastructure among Japanese people.
Furthermore, the age of people using the Internet also increased, revealing that PCs
and smartphones are used by people of all ages. The fact that people who use a
smartphone as a means of collecting information (the “low media engagement” type)
exist even in the segment of relatively young people is probably another proof of the
widespread infrastructure of the Internet. This result indicates that we need to explore
even deeper in terms of how people use PCs and smartphones, and how this is linked
to shopping, etc. rather than whether people have a PC or smartphone as a device.
1.4 Conclusion and Future Implication
As described above, the change in Japanese consumers in terms of their media access
since the emergence of the Internet is quite considerable; we can see that the Internet
has a huge impact on how consumers access media and obtain product information.
The speed of change is remarkable: only a few years ago, consumers were considered
to have the most up-to-date information if they were using the Internet-based media;
however, it was shown that, as of 2017, consumers could be further divided into
clusters based on the way they use Internet-based media, and there were clusters of
Internet users who were not so up-to-date. It should also be noted that this is affecting
not only information collection but also purchasing channels.
14
1 Japanese Consumers and Media Usage
Nevertheless, as described above, the usage rates of TV have not changed. The
usage rate for existing stores such as supermarkets and convenience stores has not
changed either. In other words, these were not something to be replaced by the
Internet and some media and channels coexisted. As the cluster analysis revealed,
this is because the options for collecting information and shopping increased due to
the emergence of the Internet. It is not that everything has changed because of the
Internet and that is something we need to continue taking into account.
The joint study with the NSK has revealed that the profile of people and the way
they think vary depending on which of these three media, i.e., newspapers, TV, and
the Internet, is used as the primary media and how it is used. It also showed that few
people use only one of those media.7 We introduced a concept of “base media” in
the joint research and investigated which of the above-mentioned three media was
essential in consumers’ daily life and society on a scale of 10. We then defined the
media that scored 10 as a base media for the consumers and explored the difference
in base media as well as in their profile and the way of thinking.
Some consumers had no base media and few others had all of the above three
media as base media. When the difference between consumers who assigned a score
of 10 and those who assigned 9 was explored for each of the above three media,
the attributes commonly seen among the three media were “I can sufficiently obtain
necessary information using only this media,” “I am focused when I watch/read
this media,” and “I watch/read this media at the beginning of the day.” Because
consumers would probably give a rating of 10 for the media that is central in their
lives, we presumed that this concept of “base media” would be useful in that sense.
Figure 1.6 shows the distribution of the respondents who identified each media as
their base media.
Looking at the differing factors among the respondents, consumers who use newspaper as their base media (47.5%) were highly interested in the society, and were
highly environmentally conscious as well. In contrast, consumers who use TV as
their base media (61.8%) prefer popular and major things and tend to go with the
Fig. 1.6 Segmentation
based on ‘base media’
Newspaper
214
others 902
696
85
755
437
TV
207
387
Internet
1.4 Conclusion and Future Implication
15
flow of the world around them, and consumers who use the Internet as their base
media (38.9%) have a strong fixation about themselves.
Next, by looking at individuals who have multiple base media, we found that
while the average age of those who consider all media—newspaper, the Internet,
and TV—as their base media was 46, with an average annual household income of
JPY7.02 million, those who use a combination of newspapers and TV were 53 years
old on average, with 40% of them in their 60s and an average annual household
income of JPY5.7 million. Meanwhile, the percentage of those with a managerial
job was higher among consumers who said newspaper and the Internet were their
base media. Their average age was 46 and their average annual household income
was the highest at JPY8.59 million among all combinations. Considering the fact that
newspaper is a so-called “broad media” that provides a wide range of information
while the Internet is a media with depth for exploring things one became interested
in, it is understandable that those in a managerial profession regard both as their base
media. Either way, because demographic factors and the way of thinking differ based
on the difference in the media usage, this result indicates that the way consumers
access the media and use information can be used as a basis for segmentation, which
is a foundation for consumer behavior.
In any case, the Internet, which has grown rapidly since 2000, has changed the
behavior of consumers considerably. There is no doubt that we need to rethink the
conventional theory of consumer behavior as well as marketing strategies associated
with it. From the next chapter on, the study will use the insights obtained here and
embark on a new direction in understanding consumer-centric marketing strategies
while providing overviews on the past studies on consumer behavior.
Notes
1.
2.
3.
4.
5.
6.
The 2015 National Media Access and Evaluation Survey, 2011 National Media
Access and Evaluation Survey, and 2009 National Media Access and Evaluation
Survey published by the Japan Newspaper Publishers & Editors Association.
The survey is conducted nationwide once biennially by randomly selecting men
and women aged between 15 and 79. The number of valid respondents was 3845
in 2015, 4092 in 2011, and 3683 in 2009.
2016 YOMIKO Resident Survey: CANVASS) published by Yomiko Advertising
Inc. This survey is conducted every year since 2000 among 1715 ordinary men
and women aged between 13 and 69 and residing within 30 km of the Tokyo
Metropolitan Area or 20 km of the Osaka Metropolitan Area, selected by using
the area sampling method (written in Japanese).
Ministry of Internal Affairs and Communications (2017), 2017 White Paper:
Information and Communications in Japan.
http://www.soumu.go.jp/menu_news/s-news/16188.html.
From 2006 Online Survey on Media and Consumer Behavior by the Japan
Newspaper Publishers & Editors Association.
DNP Media Value Research Team, Akira Shimizu (2007), Identify communication-oriented consumers, Nikkei BP. The following 18 media were used
16
7.
1 Japanese Consumers and Media Usage
in the survey. Newspapers, television (terrestrial wave), television (BS/CS),
radio, magazines, street advertisements, flyers, Brochures and free papers,
mail-order catalogs, direct mail, mail magazines, Manufacturers’ and stores’
websites on the Internet (via computer), social networking sites (via computer),
mobile phones Internet information services on the phone, in-store promotions, information from shoppers and salespeople, information from friends and
acquaintances, and information from family members (written in Japanese).
Listed above: 2009 National Media Access and Evaluation Survey published
by the Japan Newspaper Publishers & Editors Association.
Chapter 2
Evolution of the Comprehensive
Decision-Making Process: Emergence
of Outspoken Consumers
How you profile your target consumers is extremely important when it comes to
a marketing strategy. In economics, some theories regard consumers as “homo
economicus” and assume that consumers who are advantaged or in an economically advantageous position would purchase expensive products. The research on
marketing, particularly on consumer behavior, has widened its scope and now takes
into consideration the comprehensive process of consumer decision making. It
divides the cases into a “stimulus-response model,” where consumers are regarded
as being passive and reactive to external stimulus, and an “information processing
model,” where consumers are regarded as being active, addressing concerns by
collecting information on problems that they seek to resolve. This chapter will
briefly discuss these two comprehensive decision-making processes as well as the
comprehensive models of decision making that are adapted in accordance to the
environmental changes described in Chap. 1.
2.1 Stimulus-Response Model and Information Processing
Model
In the history of studies on consumer behavior, it was in the late 1950s that researchers
began studying individual consumers. Prior to that, economic psychology to understand the trend in overall consumption and studies that classified consumers into
several segments to explore their characteristics constituted the mainstream. Studies
evolved as segmentation became essential in a marketing strategy, particularly after
the segmentation, targeting, positioning (STP) marketing approach was proposed,
raising the expectations that consumer behavior studies will identify remarkable
perspectives to divide segments (Shimizu 1999).1
Amid these circumstances, studies were conducted on the decision-making
process of individual consumers based on the premise that segmentation will eventually boil down to individuals measuring external stimuli such as advertising, and
© Springer Nature Singapore Pte Ltd. 2021
A. Shimizu, New Consumer Behavior Theories from Japan, Advances in Japanese
Business and Economics 27, https://doi.org/10.1007/978-981-16-1127-8_2
17
18
2 Evolution of the Comprehensive Decision-Making Process …
promotion will require one to know how they are processed in the mind of consumers,
and so on. Since these studies were carried out in a way to incorporate consumer’s
decision-making stages, ranging from recognizing a problem to making a purchase,
as well as the factors that affect each stage, these models are called the comprehensive
models of decision making. The early comprehensive decision-making processes are
based on the idea of neo-behaviorism by C. L. Hull, which was well established in the
field of psychology at the time. Hull expanded the idea of behaviorism, which tried to
explain human behavior based on stimulus (S) and response (R). It is characterized
by organism (O), which was added between the S and R to represent the human
getting the stimulus. The idea contends that external stimuli such as advertising, instore promotion, and discount drive consumers (or the subjects), and consequently
purchases. Based on this idea, many stimulus-response models of the comprehensive
decision-making model, such as the Nicosia model, Engel-Kollat-Blackwell (EKB)
model, and Howard-Sheth model, were formulated by the early 1970s (Nicosia 1966,
Engel, Kollat, Blackwell 1968, Howard, Sheth 1969).2
In these stimulus-response decision-making processes, consumers are assumed to
be passive entities that are driven to act by external information rather than being the
ones who collect information on their own. Thus, this is a valid theory for products for
which consumers make a decision based on traditional media, allowing information
to come in naturally rather than going forward and proactively collecting information.
For example, the vast majority of purchases made by consumers at supermarkets and
drugstores in Japan are unplanned purchases, wherein the decision is made at the
store; it is said that 70% of all purchases are unplanned (Shimizu 2004).3 This is
because most of the products sold at supermarkets and drugstores in Japan have
been commoditized even though they are high quality due to intense competitions
among manufacturers; furthermore, manufacturers drive consumers to purchase by
frequently launching sales promotions. Therefore, consumers are often prompted to
decide on a purchase by stimuli such as in-store end island promotion and various
discounts. Thus, it seems these products are purchased according to the stimulusresponse model of decision making.
In contrast, one of the characteristics of the information processing model of the
comprehensive decision-making process—a theory that emerged in the late 1960s—
lies in the fact that consumers are regarded as people who proactively collect information to achieve own goals. Of all information processing models, the model proposed
by Bettman (1969) is a typical example.4 Its underlying idea is the “decision net”
proposed by Newell-Shaw-Simon, which had been successful at the time as a way to
solve decision-making problems in economics (Newell, Shaw, Simon 1958).5 In this
theory, what drives a purchase is the objective or goal that the consumer first thinks
of to resolve; here, a large difference from the stimulus-response model lies in the
fact that the driver is generated internally rather than from external stimulus. To take
the Bettman model as an example to illustrate the process, once the goal or objective
is determined, a consumer collects information accordingly, makes up his/her mind,
and eventually makes a purchase. The entire process is performed proactively because
the consumer is internally motivated. It is said that consumers use this information
2.1 Stimulus-Response Model and Information Processing Model
19
processing model of the comprehensive decision-making process when purchasing
a product that requires careful consideration, such as relatively expensive products
and products that are risky or require a lot of information.
Another characteristic of the information processing model is the assumption that
consumer’s information processing ability controls the decision-making process. For
this reason, the middle process is simplified to make decisions when the consumer’s
information processing ability is not strong or when there are too many pieces of
information to be processed at once. The elaboration likelihood model by Petty
and Cacioppo (1983), which advanced the information processing model, assumes
that when consumers process information, they make up their mind by using two
routes6 —the central route in which they make up their mind through cognitive
thinking and the peripheral route in which they make up their mind through emotional
thinking. It seems that when the information processing ability is weak, the decision making would weigh more on the peripheral route which is less burdensome
rather than the central route which is highly burdensome in terms of information
processing. In other words, consumers are not always making logical decisions even
with expensive and high-risk products; they are incorporating emotional judgment as
well. The idea that the mind is made up through these two paths—that is, cognitive
and emotional routes—in turn leading to the ultimate purchase, was also applied in
the study on the comprehensive decision-making model by MacInnis et al. which
was published after the study by Petty et al. (MacInnis, Jaworski 1989, MacInnis,
Moorman, Jaworski 1991).7 It has also become a common idea in the information
processing model of the decision-making process.
As described earlier, the theory of comprehensive process of consumer decision making presumes that consumers use one of the two decision-making rules—
the stimulus-response model and information processing model—depending on the
situation in which they are placed under. According to previous studies, products
are almost clearly divided into the ones purchased by following the rules of the
stimulus-response model and the ones purchased under the information processing
model. However, with the development of Internet-based media, it does not seem
sufficient to try to understand consumer’s decision-making process based only on
the stimulus-response model and information processing model.
This is because, as shown in Chap. 1, consumers who frequently use Internetbased media and come in contact with a lot of information and those who rely on
traditional mass media and do not come in contact with a lot of information use
completely different media in each stage of decision making even when purchasing
the same product. In particular, the segment of heavy smartphone users obtains information instantaneously and does not mind taking that extra step; they are probably
using the information processing model of the decision-making process in most
product-selection settings. It had been thought in the past that a consumer uses the
stimulus-response model or information processing model depending on the situation. However, now that there is a wide variety of information contacts among
consumers and differences are emerging, it is logical to think that the decision-making
process would differ by how proactive the consumer is about information contacts
20
2 Evolution of the Comprehensive Decision-Making Process …
even when purchasing the same product. In other words, the conventional assumption that a particular decision-making process is employed based on the situation the
individual is placed under no longer holds.
Second, both the stimulus-response model and information processing model
show consumer’s process flow up to the product purchase but do not refer to the postpurchase influence on other people. As outlined in Chap. 1, it has been shown that
word of mouth—especially the one via the Internet such as blogs, Twitter, Facebook,
and Instagram—is a rapidly growing source among young people in today’s Japan
and such online consumer reviews play a significant role in decision making. Now
that new SNS tools are appearing one after another, the influence of SNS is believed
to grow more and more. In this situation, the traditional decision-making process
that ends with product purchase is insufficient; post-purchase information sharing
must be taken into consideration as well.
However, comprehensive decision-making process studies that examine all the
ways to post-purchase information sharing are barely found in Europe and the United
States. Darley, Blankson, and Luethge (2008) argued for the need of a comprehensive
model of consumer’s online behavior and reviewed past studies on consumer behavior
on the Internet.8 They created detailed categories for the decision-making process and
factors that affect the process, and examined 52 studies on consumer’s online behavior
that have been published in major marketing academic journals between 2001 and
2008. According to this analysis, while most studies dealt with the effect of external
factors on the decision-making process, there was no study that comprehensively
dealt with everything, including problem recognition, searching, consumption, and
subsequent behavior. There was only one study that dealt with post-purchasing and
discussed cognitive dissonance. Based on these results, they concluded that studies
that explain the relationship between consumer’s decision-making process and online
behavior are in the preliminary stage and it will become essential to reexamine
information sharing and consumer’s comprehensive decision-making process model.
Third, the assumption that the decision-making process presumed in the traditional
comprehensive model of decision making is completed within the individual must be
put down once post-purchase information shared by individuals becomes the source
of information for other potential customers. The past studies on the comprehensive
model of decision making have focused on individual decision making and, while the
possibility of a purchase experience affecting the person’s next purchase had been
presumed, they did not assume people would share their purchase experience with an
unspecified number of people via the Internet. This issue will be discussed further in
Chap. 10. Meanwhile, the present chapter will consider the difference in the decisionmaking process by consumer’s media contacts as well as a comprehensive model of
decision making that incorporates two specific points on information sharing.
2.2 Significance of a Comprehensive Model of Decision Making …
21
2.2 Significance of a Comprehensive Model of Decision
Making That Considers Outspoken Consumers
The importance of having consumers recommend a specific manufacturer or brand—
that is, human communication—was mentioned as the effect of the word of mouth
even before the emergence of the Internet. According to Hamaoka (1993), word of
mouth is effective when it comes from a trusted individual in a group one belongs to
(reference group) and it is particularly effective for products with uncertainty.9
Figure 2.1, which is from a study that was conducted before the emergence of
the Internet, plots products by using the type of information required at the time of
product purchase (degree of ambiguity) and the corresponding information acquisition behavior among consumers (proactive vs. passive) as two axes (Shimizu 1995).10
To follow the above argument by Hamaoka, out of these products, the ones suitable
for word of mouth are the products consumers proactively collect information on,
particularly the items such as cosmetics and sanitary products that require ambiguous
information. It is often mentioned that people use the information processing model
of a decision-making process—which requires many parts of information in order to
avoid risk—for products such as PCs and automobiles, which are relatively expensive and cannot be replaced easily once purchased. However, since the performance
and usage method of these products can be inferred based on objective information
called specs, ambiguous and subtle nuances obtained from word-of-mouth information do not really become necessary except for some enthusiasts. In contrast, since
cosmetics and sanitary products are very delicate issues related to one’s own skin
type and physical constitution and not the kind of information you can ask anyone,
people often consulted others when the Internet did not exist. For this reason, word
of mouth by trusted individuals who are knowledgeable about cosmetics or sanitary
products and are also in one’s own reference group was important.
Passive
Toothpaste
Beer
*Shampoo
Cleaner
Tissue
Instant noodle
Television set
Detergent
Canned coffee
Energy Drink
Cold medicine
Air Conditioner
Personal Computer
Washing
machine
VCR
Automobile
Sanitary
Cosmetic
Proactive
Ambiguous
Information
Specific
Information
Fig. 2.1 Type of Information required and Information acquisition behavior
22
2 Evolution of the Comprehensive Decision-Making Process …
The development of the Internet environment is expanding the extent of this
human communication’s influence from a specific reference group to the entire world.
Trusov, Bucklin, and Pauwels (2009) examined the effect of online consumer reviews
on acquiring new customers by comparing it to the existing marketing measures with
real data.11 It was found that consumer reviews on the Internet were 20 to 30 times
more effective than existing marketing measures in terms of acquiring new customers
in the long run. This is because the communication on the Internet spreads not only
within the existing formal group but also to loosely-connected but not completely
unrelated communities (having similar hobbies, for example). In other words, while
the coverage of word-of-mouth communication in the past was limited to the group
one belonged to, the development of the Internet made it possible to also influence
people who are loosely connected.
If so, approximately when did the notion of “online consumer reviews with such
characteristics affect consumer’s decision making” emerged?
According to a survey conducted in 2005 by Nikkei Inc. (Yamaoka 2005),12
among consumers who have browsed internet communities, the percentage of those
who referred to information such as posts on the Internet and purchased a product
or decided to use a service reached 85.2%. Other sources of information valued
by online community viewers include word of mouth from acquaintances (51.2%),
review information posted on the Internet (34.9%), catalogs of products, etc. (29.7%),
and newspaper and magazine articles (25.8%), indicating that word-of-mouth reviews
have been playing a large role in product selection since that time.
According to a joint survey conducted by the author with the Japan Newspaper
Publishers & Editors Association (Nihon Shinbun Kyokai, or NSK) in November
2006, product recognition is mainly done by newspapers (40.2%) and company
website (36.6%). On the other hand, product recognition on consumer review sites
and comparison sites is only at 19.8%. Consumer review sites and comparison sites
are rated in terms of “the information is accurate (49.3%),” “it became a habit to
look at review sites (44.6%),” “it is indispensable in my daily life (45.4%),” and “I
can research the information I want until I’m convinced (58%).”13 Although the role
of consumer review sites at the stage of product and service recognition was not
large then, we can see that consumer review sites had become common among many
Japanese people as there were many consumers who agreed on their usefulness.
By 2011, the Japan Direct Marketing Association (JDMA) had conducted a survey
on mail order usage four times.14 When it asked online shoppers about information
sources that are useful for product recognition and purchasing, the results of the
2010 survey indicated the top three information sources for product recognition as
“email newsletter (30.8%),” “company website (28.6%),” and “consumer review site
(19.2%),” followed by “seeing at the physical store” and “TV program.” As for
information sources that prompted purchasing, it included “consumer review site
(28.6%),” “company website (23.5%),” and “seeing at the physical store (15.6%),”
indicating that consumer review site is an important source of information for product
recognition and purchase for those who shop online. In the 2012 survey, shopping
site was the number one source of information for prompting product recognition
and purchase (44.9% and 35.2%, respectively), followed by consumer review site
2.2 Significance of a Comprehensive Model of Decision Making …
23
(27.3% and 30.2%, respectively). Since the survey was conducted among online
shoppers, the results probably overstate the degree of Internet dependency compared
to the results of surveys conducted among consumers in general. However, we can
see that consumer review sites have had a large effect on product recognition and
buying motive year after year.
Thus, it can be said that since more than ten years, Japanese people have been
reading consumer reviews on the Internet and using them as a reference when making
purchases, and this number is increasing. However, the understanding in the past
was that not many people were proactively posting information on the Internet. For
example, Shibuya (2004) points out that the core participants on the Internet are
read-only members (ROMs) and it is only some participants that actually communicate with other participants.15 In addition, a study by Miyata (2001) indicates that
participants in an online community are those who seek professional information to
meet own needs or large amounts of new information as well as those who frequently
look at user reviews. It states that their online community usage includes (1) monitoring the information on products and services of interest by regularly browsing,
even if there is no plan for purchasing; (2) using the amount of positive vs. negative
feedback (distribution of opinion) on the comment to determine whether the quality
of information is good; (3) supporting each other; and (4) making demands to the
company.16 Based on these studies, we can infer that not many people posted reviews
on the Internet even though there were many consumers who are influenced by online
reviews.
In fact, there had been research results indicating that Japanese people do not
share opinions online even though they do so offline. Hamaoka and Satomura (2009)
presented the results of a survey on behavior after watching movies.17 According
to this survey, which was conducted among university students, although the vast
majority told others about the movie they watched and only 9.5% did not, the
percentage of those who did so on the Internet was just 1.9%. However, there were
quite a few who referred to online reviews such as internet forums before watching
a movie and they were more likely to post reviews on the Internet than those who
watched a movie based on offline word of mouth. In addition, it was found that
those influenced by offline word of mouth were likely to make comments offline.
It highlighted the state where many people share comments offline but only some
people share comments online and, despite the situation, many people refer to online
reviews.
If so, why is it that many consumers were referring to the kind of information
that only some people shared? Miyata and Ikeda (2008) introduces a concept called
“market maven.”18 Market maven, which is a theory published in the 1987 by Feick
and Price, refers to an individual with multifaceted ability to bring information.
When market mavens share opinions online, they drastically increase the amount
of information and become a reference for purchasing.19 The results of actual data
analysis showed that market mavens touch traditional media and online media, and
provide information in person more often than opinion leaders. They also send and
receive more emails and read more genres of email newsletters. It seems many
24
2 Evolution of the Comprehensive Decision-Making Process …
consumers use online reviews because they as readers implicitly know that the writer
of the particular online review is a market maven. There are only a few online review
writers because only the market mavens post information.
As described earlier, a small number of individuals referred to as “market mavens”
were sharing information on the Internet and many consumers were referring to that
information when making a purchase in Japan in the past. However, the situation has
been changing drastically in the past 10 years because the number of people who
post information online has increased sharply thanks to the introduction of new tools
such as Twitter, Facebook and LINE. According to the white paper on information,
the MIC that I explained at Chap. 1, the percentage of people who use SNSs has
increased from 41.4% in 2012 to 71.2% in 2016 in conjunction with the spread of
smartphones. In that case, we must consider the quality of the consumer review since
several non-market mavens are posting their comments.
In fact, consumer reviews were not something everyone necessarily posted and
their influence on purchase behavior was not so large even in the United States about
up until ten years ago. According to a survey, conducted by Riegner (2007) among
4000 broadband users in the United States,20 on the state of online consumer review
usage and its influence on purchasing, those aged 24 or younger are more likely to
use online consumer reviews. As for the way of use, 31% write product reviews, 25%
submit to forums, and 15% create personal webpages. There is a relationship between
the percentage of individuals who post information and their age; 50% of males aged
34 or younger post information by using message boards, chat, and Wikipedia, while
50% of females aged 17 or younger post information in a similar way. However,
since the percentage of respondents who indicated that they use SNS as a reference
for making a purchase was 9%, it was reported that the United States users use SNS
as a means of communication and do not consider it as a source of information for
making a purchase.
Huang, Shen, Lin, and Chang (2007), based on a survey of 311 bloggers, reported
on the motivation and behavior of individuals who write blogs.21 The average age of
respondents was 23 and 70% of respondents were aged between 16 and 24. First of
all, motivations for writing a blog included expressing self, describing life, wanting to
write comments, intending to participate in a forum, and anticipating the information
being searched via a trackback. The purposes included striking a conversation with
other bloggers, acquiring information, and expanding the network; not many respondents mentioned making a purchase. In other words, the study showed that bloggers
post information for the purpose of expanding their network rather than posting information useful for buyers and they do not think about sharing information for others
to refer to at the time of making a purchase.
In response to these studies, Berger (2014) summarized motivations for posting
and using consumer reviews into five categories based on reviews of past consumer
review studies.22 These are: (1) Impression Management (to create your image
perceived by others), (2) Emotion Regulation (to control your emotions), (3) Information Acquisition (to acquire information on products to purchase), (4) Social Bonding
(to make a connection with others), and (5) Persuasion (as a tool to persuade others).
In other words, the motivations for posting consumer reviews include building own
2.2 Significance of a Comprehensive Model of Decision Making …
25
image, controlling emotions, and connecting with others. For this reason, there are
many contents with exaggeration on the premise of being read by people and contents
launched for the sense of fellowship. Thus, this study shows that we need to be
cautious in dealing with the contents.
Based on the above information, it is certain that SNS used to have a large influence
on purchases in Japan until ten years ago because a small number of market mavens
were posting useful information. However, later on, not just market mavens but
ordinary people too began posting information as they gained access to tools for
posting information, and smartphones became common. We also need to be careful
about the fact that there are various motivations for posting information, such as
self-assertion and network building. This is done without any particular intention to
provide information useful for others in making a purchase. It can be said that the
stage of blindly accepting the information on SNS and use it for purchasing is over
and we are at the stage where we must carefully consider the information contents.
In any case, SNS will definitely play some role in consumer decision making and
we need a decision-making process model that takes this part into account. In what
follows, an overview will be offered on the decision-making models that have been
developed in Japan by incorporating a consumer’s post-purchase effect, and then
their usefulness with data will be verified.
2.3 Comprehensive Model that Considers the Effect
of Word of Mouth
While the development of consumer decision-making models is as described in
Sect. 2.1, there is an area apart from this where the comprehensive information flow
was studied from the perspective of delivering advertisement messages to consumers.
According to Bruce, Kay, and Prasad (2012), studies on the advertising effect can
be roughly divided into the ones that look at the effect of advertising on sales by
separating it into long term and short term, and the ones that consider the flow
as to how advertising affects consumer’s recognition, emotion, and experience.23
This section deals with the latter, i.e., studies that suggested a hierarchical flow in
which advertising messages lead consumers all the way to purchasing, as typified by
AIDMA.
According to a study by Barry and Howard (1990), which reviewed studies on
advertisement messages through the 1990s, past studies on the hierarchical effect
were based on Attention, Interest, Desire (AID; proposed in 1898) and evolved
into Attention, Interest, Desire, Memory, Action (AIDMA; proposed in 1956).24
According to them, there are various opinions on the order of the three factors “recognition → emotion → action.” However, they list benefits such as “there is no question
of exploring the effect of advertisement based on these three factors,” “exploring the
effect of advertisement by basing it on hierarchical flows such as AID and AIDMA
makes it possible to predict consumer behavior,” “it can provide guidelines on what
26
2 Evolution of the Comprehensive Decision-Making Process …
kind of advertisement strategy to take at which stage of the hierarchy,” and “it helps
companies in planning and considering concepts.” The theory of AIDMA became
widely known in Japan through the advertising industry, being interpreted as “running many advertisements leads to more purchases.” However, empirical analysis
has rarely been performed. Introduced below, the model that incorporates posting of
information, which has been discussed recently in Japan, has been developed in the
course of these studies.
First, as a consumer decision-making model that takes the effect of word of mouth
into account, we need to have a look at Katahira’s model. The researcher mainly
advocates AIDEES from the perspective of brand building.25 While the traditional
marketing considered companies to be the driving force to distribute products, the
researcher presumes that the driving force of product popularization is with the
customer, now that the Internet has advanced. Katahira then proposed AIDEES,
which incorporated excitement and fascination on the premise that being wowed by
a product is essential for customers to popularize the item. With the content being
Attention → Interest → Desire → Experience (experiencing) → Enthusiasm (being
excited and fascinated) → Share (information sharing), we can say that Enthusiasm
is incorporated as the driver of information sharing.
Yamamoto and Katahira (2008) demonstrated the validity of AIDEES with actual
data.26 Their study measured the effect of consumer reviews by the spread and quality
of network among consumers. It was then established that although the attainment
rates to the AIDE stages of AIDEES are extremely high at 90% or more with all
products surveyed (automobiles, PC peripherals, snacks sold at convenience stores,
music), among those who received word of mouth, the rate considerably drops at the
next E, or Enthusiasm. The rates ranged from 7 to 8% for snacks sold at convenience
stores and music, and 1 to 2% for automobiles and PC peripherals. This shows that
while consumer reviews can push customers through the flow up to product experience, it cannot bring them to the stage of being wowed by the product. Furthermore,
the level of difficulty for Enthusiasm is high when AIDEES is actually applied, even
though AIDEES is excellent as a concept.
Among the models that represent information sharing after purchasing, the most
frequently cited one is AISAS®.27 Proposed by Akiyama and Sugiyama in 2004,
AISAS® is characterized by the fact that it improved the consumer decision-making
process by taking the influence of the Internet into account. AIDMA—which had
long been used as the theory of advertising effect measurement—assumes passive
recipients of stimulus and shows a process as to how they respond to the stimulus by senders. However, AISAS® is characterized by the assumption of proactive
recipients and presumes the flow of “directing attention to the product (Attention),
becoming interested (Interest), searching on the Internet and other sources (Search),
subsequently purchasing the product or service (Action), and posting a feedback
on the Internet regarding the product purchase (Share).” Building upon the idea of
the comprehensive model of consumers, it adds information sharing to the information processing model. Here, “Share” refers to the action of writing comments and
concerns on the SNS regarding the purchased products. Most of the “Search” is done
through search engines on the Internet. However, Hamaoka (2006) has shown that
2.3 Comprehensive Model that Considers the Effect of Word of Mouth
27
the percentage of people who mention consumer review sites as an important source
of information at this search stage goes up to 40%, which is at the same level as word
of mouth from friends and families.28 In other words, the circulation of information is
created where posts shared on the SNS by strangers can be searched just as easily as
on company’s official websites. Morioka, Hasegawa, and Yamakawa (2006) named
this cycle “AISAS® feedback loop,” deeming it as a proof that the theory of AISAS®
is valid.29
Taking this further, another idea called SIPS emerged to regard “Share” as the
starting point for information recognition. The concept, which was proposed by the
Dentsu Modern Communication Laboratory, states that it is necessary to consider
communication centered on SNS as SNS evolves.30 SIPS comprised Sympathize
→ Identify → Participate → Share and Spread. The starting point in the previous
models, whether it is the stimulus-response model or AISAS®, was information
provided by the company side. However, SIPS is characterized by the fact that
Sympathize, which includes information on SNS, is the starting point of the decisionmaking process. While SNS in the past had been somewhat of an information search
medium open to the general public, Facebook and Twitter are networks of groups
that are close to one’s own reference group connected in a closed world, intended to
be more of a triggering media than search media. In SIPS, the focus is on evangelists
who are particularly capable of disseminating information. Similar to the concept of
market maven explained earlier, it shows that not everyone could become the starting
point.
We can see based on these studies on advertising communication that explaining
a consumer’s decision-making process by the stimulus-response model and information processing model is not sufficient and it should include post-purchase communication to understand the behavior of consumers in today’s Japan where Internet-based
media is found everywhere. In what follows, real data to check the existence of the
stages in the AISAS® theory will be analyzed.
2.4 Demonstration of the AISAS® Theory
As described above, the emergence of Internet-based media has shown that the stage
of information sharing after purchasing is necessary in a consumer’s decision-making
process. If so, does the flow that goes up to information sharing as shown by the new
advertising communication model really exist?
As shown in the previous section, cases of empirical studies on the advertising communication models that consider all the way down to post purchases are
extremely rare although it is a topic of main discussion in Japan (Hamaoka, Satomura
2009).31 In this chapter, I will have a look at AISAS® and show a case in which the
theory was validated with real data. The study was conducted in collaboration with the
NSK over the period of November 2006.32 An online survey was conducted among
male and female Internet users aged between 15 and 69 and the results were tabulated.
28
2 Evolution of the Comprehensive Decision-Making Process …
The number of samples was 3803, and the surveyed products ranged across 12 categories, including beverage/beer/alcoholic beverage, cosmetics/toiletries, home appliances/audiovisual equipment, automobile, books/magazines, distribution/retailing,
finance/insurance, travel/transportation, real estate/housing equipment, educational
services, fashion/brand items, and digital equipment.
The AISAS® theory does not explicitly list specific criteria for stages such as
recognition and interest. Therefore, I defined each stage as follows: the recognition/interest stage if any of “I learned about a new product,” “my interest or concern
increased,” “there was a product that caught my attention,” “there was a product that
might be suitable for me,” or “there is a product that made me want to purchase/use”
was applicable; the information searching stage if any of “I think about it when I
look at information,” “I asked family, friends, acquaintances, etc. about the product,”
“I searched about it on the Internet,” “I looked at the company’s website,” or “I
began actively collecting information” was applicable; the purchasing stage if any
of “there is a product that I looked into in person at the store,” “there is a product
that I purchased for the first time,” “there is a product I purchased again,” “there is a
product I purchase more frequently or use more,” or “there is a product that became
my favorite or I came to like more” was applicable; and information sharing stage
if any of “I talked about the product that I purchased,” “I commented on the product
on an internet forum or blog,” or “I recommended the product or gave advice” was
applicable. Subsequently, whether each item was applicable to the respondents or
not was tabulated and all surveyed products were aggregated in accordance with the
flow of AISAS®. Figure 2.2 illustrates the changes in the percentage of respondents.
As a note, the figure shows those who said “yes” above those who said “no” under
each stage. For example, the percentage of those who recognized and took interest
was 27%, while the percentage of those who did not was 73%.
4.8
10.6
5.8
16.8
6.2
27.0
0.5
10.2
7.3
2.4
Total
0.3
2.1
12.8
10.7
1.8
0.2
73.0
2.7
60.2
2.5
57.5
Recognition
/Interest
Information
searching
Fig. 2.2 Demonstration of the AISAS® theory
Purchasing
Information
Shareing
2.4 Demonstration of the AISAS® Theory
29
Based on this, we can see that 5.8% of the entire purchaser population was
in the role of an influencer at the time based on the definition of the AISAS®
theory. The corresponding figure, which varies by product category, is 1.7% for
finance/insurance, 9.6% for home appliances/audiovisual equipment, and 12% for
beverage/beer/alcoholic beverage. The percentage of people who share information—or influencers—differs by product categories.
While these figures themselves are only for our reference since the survey is old,
what should be noted is that those who took the purchase action with recognition
and interest are more likely to become influencers than those who did not. As we can
see in Fig. 2.2, while those who made a purchase with recognition and interest and
became influencers accounted for 5.3% of the entire population, those who purchased
without recognition and interest and became influencers accounted only for 0.5% of
the entire population. Given that 27% of people had recognition and interest and 73%
did not, it means approximately 20% of those who made a purchase with recognition
and interest became influencers while only about one in 150 of those who did not have
recognition and interest ended up becoming influencers. In other words, it suggests
that people are less likely to become influencers unless they have a firm recognition
and interest to search for information and make a purchase.
With regard to the decision-making process model, when consumers purchase
products under the stimulus-response model of the decision-making process, they are
less likely to become influencers because they ultimately make a purchase in response
to an external stimulus rather than as a result of voluntary information search based
on recognition and interest. Today, many studies indicate that promotion techniques
such as discounts and end island promotion frequently offered at supermarkets, drugstore, etc., are effective means to secure sales since they are instantaneous and can
easily make consumers to respond. The breakdown of actual advertising expenditure
also indicates that the percentage of promotion tends to be maintained even when
online advertising increases and other advertising expenditure decreases (Shimizu
2004).33 However, this kind of promotion is not a way of selling by making consumers
recognize and take interest. It is clear that companies must sell in a way wherein they
do not rely on traditional promotions if they want to nurture influencers and have
them post positive reviews on their products.
Based on this empirical analysis, we can see there is a value to the theory of
AISAS® as a basis for the comprehensive model of consumer decision making in
an era where the Internet is heavily used.
2.5 Situations in Which the Customer Would Like
to Influence
While the above results showed that those who searched for information with recognition and interest were likely to become influencers, it is not the case that everyone
who makes a purchase after going through that process will become an influencer.
30
2 Evolution of the Comprehensive Decision-Making Process …
Even in the above empirical analysis, less than 20% of consumers who made a
purchase based on recognition and interest became influencers. This is probably
because consumers would not spread the word after making a purchase unless there
is some kind of emotional connection as shown in the model by Katahira. In fact,
previous studies have also indicated that emotional factors after making a purchase
influence the word of mouth.
After stating there are only a few studies that take into account the relationship
between word of mouth and the emotion after using the product, Ladhari (2007)
used real data to link satisfaction and emotional rating and examined the process in
which people watch a movie and then spread the word.34 According to this study,
the emotional excitement after watching a movie influenced satisfaction and led
to spreading of the word. This study shows that an emotional rating gained from
watching a movie generates word of mouth via satisfaction.
Brown, Barry, Dacin, and Gunst (2005) elucidated the mechanism by which
car buyers spread the word about their cars.35 The samples were 3000 consumers
randomly selected from the dealer’s customer list and the mechanism was explored
by using Structural Equation Modeling (SEM). According to this study, satisfaction
of consumers led to intention to spread the word via their commitment and then
to action. Since “commitment” refers to an emotional connection with the product,
this study indicates the importance in building satisfaction and emotional connection with the product —or Enthusiasm as argued by Katahira—after the purchase for
spreading the word.
Celso Augusto and Carlos Alberto (2008) reviewed 242 previous studies that
conducted empirical analyses on word of mouth, which were published in 28
marketing-related journals. They extracted 127 studies that researched items affecting
word of mouth and summarized those items.36 According to the research, the number
of studies that dealt with satisfaction as something to influence the word of mouth was
the highest (89), followed by loyalty (62), perception (10), credibility and quality (9),
and commitment (8). Of these, the one with the largest influence on word of mouth
was commitment, followed by satisfaction and loyalty. It was indicated, however,
that loyalty had a greater effect on generating a negative word of mouth when the
product did not meet the expectations. In terms of what the word of mouth actually
said, the study also showed that a positive word of mouth affects cognitive evaluation
while a negative word of mouth is emotional and has an immediate effect, likely to
affect behavioral intentions. We can see from this study that consumer satisfaction
and commitment are important if a positive word of mouth is desired.
The Japanese Customer Satisfaction Index (JCSI) also shows that customer satisfaction affects not only loyalty but also word of mouth in the service industry in
Japan.37 The JCSI is the Japanese version of customer satisfaction model developed by referring to various customer satisfaction indices implemented overseas.
It presumes customer satisfaction as the conventional function of expectation and
results indicated in Oliver’s study, and also the behavior after customer satisfaction
takes place, listing word of mouth and loyalty as behaviors occurring after satisfaction. The model collected a huge amount of data on 29 industries to check the
relationship. According to the results, a positive relationship is observed between
2.5 Situations in Which the Customer Would Like to Influence
31
customer satisfaction and word of mouth regardless of the industry even though the
parameters vary, indicating that customer satisfaction can lead to word of mouth in
the service industry in general.
As described, we can see that generating word of mouth is related to the emotional
connection with the product, as typified by satisfaction and commitment. The fact that
Katahira’s model included Enthusiasm (being excited or fascinated) as a prerequisite
for triggering word of mouth supports this argument. As for commitment, this will be
discussed in detail in Chap. 4. In this chapter, consumer satisfaction will be examined,
which is also incorporated into the existing information processing model.
Studies on consumer satisfaction, which are often conducted around the service
industry, have been pointing out the relationship between word of mouth and satisfaction for a long time. The classic study by Oliver (1980), for example, noted the effect
that a satisfied consumer would make positive comments on the company’s brand
and refer potential customers.38 Yi (1990)’s study, which systematically reviewed
consumer satisfaction studies conducted till 1990, noted that the number of potential
customers would decrease because customers who had chosen the products would tell
others about their dissatisfaction in the event when they could not get satisfaction with
the product they chose.39 There is also a research result indicating that information
such as emotional judgment for good and bad is particularly likely to spread via word
of mouth (Westbrook 1987).40 Furthermore, according to a study by Szymanski and
Henard (2001), which comprehensively reviewed past consumer satisfaction studies
conducted till 2000, there is a negative relationship between negative reviews and
consumer satisfaction.41
As described above, how consumer satisfaction relates to the subsequent behavior
of spreading the word is shown not only on the Internet but also with word of mouth
in person. If so, what kind of media contact leads to satisfaction? In what follows, I
will use the theory of AISAS® and analyze how media touches people in a way that
leads to satisfaction and information sharing.
2.6 Media Contact that Leads to Satisfaction
and Information Sharing
The data used for the analysis is the 2009 data from Zenkoku Media Sesshoku Hyōka
Chōsa (The National Media Access and Evaluation Survey) which was conducted
in collaboration with the NSK. The number of samples was 3683 and the survey
was conducted by using the drop-off/pick-up survey method. The criteria for media
used for each stage of AISAS® were: [recognition and interest] media for which the
statements “regularly view or listen to (regarding the product)” and “prompted me to
take interest (in the product)” were selected; [information search] media for which
the statement “used for doing various research on a given product” was selected;
and [leading to the store] media for which the statement “prompted me to go to the
store to see the product (to check the product details, etc.) was selected. Furthermore,
32
2 Evolution of the Comprehensive Decision-Making Process …
the response to the question “did you purchase (replace, use, remodel) the product
(in the past 2–3 years)?” was counted for [purchase behavior] and the response to
the question “were you satisfied after the purchase (use)?” was counted for [satisfaction], and the number of responses to two questions “did you tell your friends
or acquaintances about it after the purchase (use)?” and “was there anything you
wanted to tell others about after the purchase (use)?” was counted for [information
sharing]. There were five surveyed product categories, including “books and magazines,” “mobile phones,” “travel packages,” “automobiles,” and “real estate (detached
homes, condominiums, etc.).
First, I checked to see if the respondents who said “satisfied” were sharing information in the first place. The results showed that 76.2% of those who purchased
a product were satisfied with the purchase and 22.6% were dissatisfied with the
purchase (they do not add up to 100% due to the existence of some unknowns) and,
while 50.8% of satisfied individuals shared the experience, only 4.1% of dissatisfied
individuals shared information. Here, we can see that satisfied people intend to tell
their experience to others and dissatisfied people do not tell them. That is the reason
that the negative comments are extremely rare.
Next, I looked at the relationship between satisfaction and the intention for sharing
information. At first, I chose the sample who made a purchase with recognition
and interest. Then I checked the satisfaction rate, information sharing rate, and the
difference in the media that prompted the recognition and interest. The results showed
that satisfaction and information sharing increased to 80.4 and 54.6% (Compare to
overall sample’s score, 76.2% and 50.8%, respectively). Furthermore, looking at
the media that prompted recognition and interest by breaking them down into 3
types: Only newspaper (newspaper), Only TV (TV), and Both newspaper and TV
(newspaper +TV). The satisfaction and information sharing was 82.8% and 57.6%,
respectively, for newspaper; 82.2% and 59.5%, respectively, for TV; and 83.9% and
60.9%, respectively, for newspaper + TV. The level of satisfaction is higher when
a purchase was made with recognition and interest than otherwise; in particular, the
score for information sharing differs by 10% points when the cases in which the
recognition and interest came from both newspaper and TV are compared to the
overall population. So recognition, interest, and media strategy are very important
to induce word of mouth.
Each of the five surveyed products, Figs. 2.3 through 2.7 summarize the combination of the media that was used for information search and the media that led the
individual to the store, only for the cases when the recognition and interest came from
a newspaper, in terms of the patterns that resulted in the largest number of purchasers,
highest satisfaction, and most information sharing. In these Figs. 2.4, 2.5, 2.6 and 2.7,
the number at Purchase behavior stage indicate the number of people who actually
purchased. The percentage of Satisfaction stage indicates the rate who purchased a
product were satisfied with the purchase, and the percentage of Information sharing
stage indicates the rate of the satisfied individuals shared the experience.
2.6 Media Contact that Leads to Satisfaction and Information Sharing
Information
search
Leading to
the Store
33
Information
Shareing
Purchase
behavior
Satisfaction
Newspaper
246
82.9%
60.2%
Internet
45
93.3%
73.3%
66
90.9
83.3%
the largest number of purchasers
Internet
highest satisfaction
Magazine
most information sharing
Store
family, friends, or
acquaintances
Total sample size: 2841
Fig. 2.3 Books and magazines
Information
search
Leading to
the Store
Information
Shareing
Purchase
behavior
Satisfaction
brochure/
catalog
100
88.0%
50.0%
TV
66
95.5%
69.7%
family, friends, or
acquaintances
40
92.5
75.0%
the largest number of purchasers
brochure/
catalog
highest satisfaction
brochure/
catalog
most information sharing
family, friends, or
acquaintances
Total sample size: 3122
Fig. 2.4 Mobile phones
What is common to the five products is that the percentage of information sharing
is maximized when the act of leading to the store is done via family, friends, or
acquaintances regardless of the information search method. The percentage differs
by the product; travel packages 87% and mobile phones—which had the lowest
percentage—75%, indicating a gap of more than 10 percentage points. In addition,
in the cases of three products (mobile phones, travel packages, and automobiles),
the highest number of purchasers was observed when searching for information and
the act of being led to the store was done via brochures. On the other hand, for
satisfaction, the media that lead to the store is different by products. These figures
34
2 Evolution of the Comprehensive Decision-Making Process …
Information
search
Leading to
the Store
Information
Shareing
Purchase
behavior
Satisfaction
brochure/
catalog
275
89.5%
73.2%
Store
45
97.8%
84.4%
69
95.7
87.0%
the largest number of purchasers
brochure/
catalog
highest satisfaction
Store
most information sharing
brochure/
catalog
family, friends, or
acquaintances
Total sample size: 1621
Fig. 2.5 Travel packages
Information
search
Leading to
the Store
Information
Shareing
Purchase
behavior
Satisfaction
brochure/
catalog
191
93.7%
68.1%
Store
87
96.6%
71.3%
52
96.2
78.8%
the largest number of purchasers
brochure/
catalog
highest satisfaction
brochure/
catalog
most information sharing
family, friends, or
acquaintances
family, friends, or
acquaintances
Total sample size: 3072
Fig. 2.6 Automobiles
indicate that the combination of media with the highest probability for satisfaction
does not necessarily result in the highest percentage of information sharing. Thus,
while the level of satisfaction certainly leads to information sharing, it alone cannot
explain the trend.
Looking at it by product, first with books and magazines, while the combination of
media for searching for information and being led to the store for the largest number of
purchasers is the Internet and newspaper, the combination for the highest satisfaction
is newspaper and the Internet, and the combination for the highest information sharing
is store and family, friends, and acquaintances. Comparing the pattern in which the
2.6 Media Contact that Leads to Satisfaction and Information Sharing
Information
search
Leading to
the Store
35
Information
Shareing
Purchase
behavior
Satisfaction
flyer inserts
119
81.5%
61.3%
Newspaper
58
87.9%
63.8%
55
83.6
76.4%
the largest number of purchasers
flyer inserts
highest satisfaction
flyer inserts
most information sharing
Store
family, friends, or
acquaintances
Total sample size: 1849
Fig. 2.7 Real estate
number of purchasers is maximized with the pattern in which information sharing
is maximized, while the number of purchasers is close to four times larger, the
percentage of information sharing is actually lower by more than 20% points. While
the top objective for companies is to drive consumers to purchase, this result shows
that the media would be different if the objective was information sharing.
With mobile phones, the combination of media for the highest number of
purchasers is brochure/catalog at the stage of searching for information and
brochure/catalog at the stage of being led to the store, while the combination for
the highest satisfaction is brochure/catalog and TV and the combination for the
highest information sharing is family/friends/acquaintances at both stages. Since the
patterns for the largest number of purchasers and the highest information sharing
are exactly the same as the ones for purchasing an automobile, we can see that the
cases of durable consumer goods follow similar patterns. This shows that putting it
through means of human communication is more effective in reaching information
sharing.
The pattern in which the number of purchasers is maximized for travel packages
is the same as in the cases of mobile phones and automobiles and the combination of media used for searching information and prompting to go to the store is
brochure/catalog for both stages. It seems that when it comes to travel packages, the
level of satisfaction is extremely high when the acts of searching for information
and being led to the store are both done at the store. In the case of travel, the fact
that brochures can be downloaded on company’s website, etc. contributed to this. In
case of packages, which are difficult to differentiate because there are many similar
products, it seems important to provide personal communication at the store before
consumers make a purchase in order to build satisfaction.
Just as in the cases of the above four products, the media that is useful in making a
purchase, the media that leads to satisfaction, and the media that leads to information
36
2 Evolution of the Comprehensive Decision-Making Process …
sharing differed in the case of real estate. Unlike other products, real estate is characterized by the fact that leading to the store by using flyer inserts affects the number
of purchasers. However, it does not necessarily result in post-purchase satisfaction.
These survey results revealed that media used at the stage of searching for information and for leading to the store are diverse; the combination that majority of
purchasers currently use, the combination that results in the highest satisfaction, and
the combination that results in the highest percentage of information sharing differ
and they also differ by product. In particular, it was shown that the information
obtained by contacting family, friends, or acquaintances is necessary at the stage of
leading to the store for the percentage of information sharing to increase; consumers
will not share information unless information is shared with their own reference
group. Given that the values shared among people connected by network are acting
as a reference group in recent years, it can be said that this statement showed the
importance of a network. This result also indicates that from the company’s perspective, they need to choose which media to use depending on the purpose, whether they
want to sell, to satisfy customers, or to have customers share information.
2.7 Conclusion and Future Implication
In the past, the comprehensive process of consumer decision making had been understood based on the stimulus-response model and the information processing model.
However, these models did not take post-purchase information sharing into consideration even though they included the stage up to consumer’s purchase. Considering the
fact that the magnitude of the role played by SNS is being talked about given the recent
development of the Internet, a comprehensive decision-making model to take that into
account is necessary. Although the development of these comprehensive models is
almost absent in Europe and the United States, there are some pioneering conceptual
models such as AIDEES and AISAS® derived from advertisement communication
research in Japan. In this chapter, I had a look at the AISAS® theory and verified
it by using real data to confirm whether this concept is a model that can actually be
utilized.
The results showed that the flow that goes down to information sharing in accordance with the theory of AISAS® indeed exists. Furthermore, consumers who
purchased after searching for information with recognition and interest, in particular, are more likely to become influencers than consumers who purchased without
recognition and interest. It showed the certainty of the concept of information sharing
as well as the importance of the stages of recognition and interest—something traditional mass media is good at—in getting to information sharing. It implies that each
media has its own role and it is important to use them well, rather than providing all
information via the Internet, in order to develop influencers.
In response to these results, I considered the process to develop influencers.
Previous studies have shown that people spread the word when they are satisfied
with the results of their purchase or when the commitment derived is strong. Thus,
2.7 Conclusion and Future Implication
37
following the theory of AISAS®, I explored the patterns of media contact that resulted
into actual satisfaction or information sharing. It thus became clear that (1) consumers
are not satisfied when they simply purchase with recognition and interest; the level
of satisfaction does not improve unless they purchase after searching for information
with recognition and interest and (2) based on the results of analysis on the relationship between the level of satisfaction and media contact by product, changing the
contact media such as brochures, friends/family, and store based on the product to
sell will ultimately result in a larger number of purchasers, higher satisfaction, and
higher level of information sharing than otherwise.
Based on these results, the idea based on the theory of AISAS® was shown to
be effective in developing a new, comprehensive model of decision making that
includes word of mouth, although it would be more realistic to insert a step related to
satisfaction between purchase and information sharing, that is, between Action and
Share. The aforementioned AIDEES by Katahira, which listed Experience (to experience) → Enthusiasm (to be excited or fascinated) → Share (to share information),
included the item “to be excited or fascinated” between purchase and information
sharing. Incorporating satisfaction into the theory of AISAS® is almost the same as
the idea of AIDEES; we could say that the effectiveness of Katahira’s model was
also demonstrated.
As described above, the comprehensive models of consumer decision making that
have been discussed in Europe and the United States are insufficient in the Internetbased society. Furthermore, the comprehensive decision-making models that include
steps through information sharing as shown in the studies on advertising communication in Japan are quite important. Since 1980, consumer behavior research had been
based mainly on the information processing model of the comprehensive decisionmaking model and evolved in a way that delves deeply into each of its components. The fact that the post-purchase behavior called information sharing has been
highlighted indicates that we need to add that part to the information processing
model of the comprehensive decision-making model. Meanwhile, we also need to
be careful about the fact that the information people post is not necessarily useful
compared to the time when SNS first emerged. For this reason, not everyone would
take much account of information at the Share stage. I will present a new comprehensive decision-making model in Chap. 10 by addressing this point and verify it
with real data.
Notes
1.
2.
3.
Shimizu, Akira (1999), New Consumer Behavior, Chikura-Shobo (written in
Japanese).
Nicosia, Francesco M. (1966), Consumer Decision Processes, Prentice-Hall.
Engel, James F.; Kollat, David T.; Blackwell, Roger D. (1968), Consumer
Behavior, Rinehart &Winston. Howard, John A; Sheth, Jagdish N. (1969),
The Theory of Buyer Behavior, John Wiley & Sons.
Shimizu, Akira (2004), Retail Strategy Based on The Client Perspective,
Chikura-Shobo (written in Japanese).
38
2 Evolution of the Comprehensive Decision-Making Process …
4.
Bettman, James R. (1979), An Information Processing Theory of Consumer
Choice, Addison-Wesley Publishing Company.
Newell, Allen; Shaw, J.C.; Simon, Herbert A. (1958), ‘Elements of a Theory of
Human Problem Solving’, Psychological Review, Vol. 65, May, pp. 161–166.
Petty, Richard E.; Cacioppo, John T. (1983), ‘Central and peripheral routes
to advertising effectiveness: The moderating role of involvement’, Journal of
Consumer Research, Vol. 10, pp. 135–146.
MacInnis, Deborah J.; Jaworski, Bernard J. (1989), ‘Information Processing
from Advertisements: Toward an Integrative Framework’, Journal of
Marketing, Vol. 53, October, pp. 1–23, and MacInnis, Deborah J.; Moorman,
Christine; Jaworski, Bernard J. (1991), ‘Enhancing and Measuring Consumers’
Motivation, Opportunity, and Ability to Process Brand Information From Ads’,
Journal of Marketing, Vol. 55, October, pp. 32–53.
Darley, William K.; Blankson, Charles; Luethge Denise J. (2010), ‘Toward an
Integrated Framework for Online Consumer Behavior and Decision Making
Process: A Review’, Psychology & Marketing, Vol. 27, pp. 94–116.
Hamaoka, Yutaka (1993), ‘Consumer Interaction and Dependence’, Marketing
Science, Japan Institute of Marketing Science, Vol. 2, No. 1, 2, pp. 60–85
(written in Japanese).
Shimizu, Akira (1995), ‘Product strategy depends on Consumer’s product
recognition and buying and using tendency’, Monthly forum, Nikkei Research
Institute of Industry and Markets, No. 124, pp. 18–22 (written in Japanese).
Trusov, Michael; Bucklin, Randolph E.; Pauwels, Koen (2009), ‘Effects of
Word-of-Mouth Versus Traditional Marketing: Findings from an Internet
Social Networking Site’, Journal of Marketing, Vol. 73, pp. 90–102.
Yamaoka, Taku (2005), C to C Revolution, Nikkei Consumption Mining,
Research Institute of Research and Regional Economy, July, pp. 4–15 (written
in Japanese).
‘Research of Media usage and Consumer Behavior in 2006’, The Japan
Newspaper Publishers & Editors Association.
Japan Direct Marketing Association http://www.jadma.org/.
Shibuya, Satoru (2004), ‘Opinion Making on the Internet Community and
Marketing Strategy’, Marketing Journal, Japan Marketing Association, Vol.
94, pp. 31–44 (written in Japanese).
Miyata, Kakuko (2001), ‘What does Network community Effect for Consumer
Behavior’, Marketing Strategic Research, No. 677 (written in Japanese).
Hamaoka, Yutaka; Satomura, Takuya (2009), ‘Basic Research of Consumer
Interaction—Focus on WOM’, Keio University Press (written in Japanese).
Miyata, Kakuko; Ikeda, Kenichi (2008), ‘Internet and Consumer Behavior—
Research of word of mouth’, NTT Publishing (written in Japanese).
Feick, Lawrence F.; Price, Linda L. (1987), ‘The Market Maven: A Diffuser
of Marketplace Information’, Journal of Marketing, Vol. 51, pp. 83–97.
Riegner, Cate(2007), ‘Word of Mouth on the Web: The Impact of Web 2.0 on
Consumer Purchase Decisions’, Journal of Advertising Research, December,
pp. 436–447.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
2.7 Conclusion and Future Implication
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
39
Huang, Chun-Yao; Shen Yong-Zheng; Lin Hong-Xiang; Chang Shin-Shin
(2007), ‘Bloggers’ Motivations and Behaviors: A Model’, Journal of Advertising Research, December, pp. 472–484.
Berger, Jonah (2014), ‘Word of mouth and interpersonal communication: A
review and directions for future research’, Journal of Consumer Psychology,
Vol. 24, pp. 586–607.
Bruce, Norris I.; Kay Peters; Prasad A Naik (2012), ‘Discovering How Advertising Grows Sales and Builds Brands’, Journal of Marketing Research, Vol.
49, pp. 793–806.
Barry, Thomas E.; Howard, Daniel J. (1990), ‘A review and critique of the
hierarchy of effects in advertising’, International Journal of Advertising, Vol.
9, pp. 121–135.
Nikkei MJ (2006/4/14) and Nikkei Consumption Mining (2008/4).
Yamamoto, Hikaru; Katahira, Hotaka (2008), ‘To find the Influencer and the
Effect of WOM—Approach of AIDEES model’, Marketing Journal, Japan
Marketing Association, Vol. 109, pp. 4–18 (written in Japanese).
Akiyama, Ryuhei; Sugiyama, Kotaro (2004), ‘Holistic Communication’,
Senden Kaigi (written in Japanese).
Hamaoka, Yutaka (2006), ‘Consumer’s interaction-based on WOM’, Tanaka,
Hiroshi; Shimizu, Akira edit. Consumer and Communication Strategy, Yuhikaku, pp. 57–93 (written in Japanese).
Morioka, Shinji; Hasegawa, Soh; Yamakawa, Shigetaka (2006), ‘Proposal of
WOM Marketing Planning based on the AISAS model’, Marketing Journal,
Japan Marketing Association, Vol. 101, pp. 30–41 (written in Japanese).
http://www.dentsu.co.jp/sips/index.html.
Hamaoka Yutaka, Satomura, Takuya (2009), ibid.
The Japan Newspaper Publishers & Editors Association (2006), ibid.
See Shimizu Akira (2004), ibid.
Ladhari, Riadh(2007), ‘The Effect of Consumption Emotions on Satisfaction and Word-of-Mouth Communications’, Psychology & Marketing, Vol.
24, pp. 1085–1108.
Brown, Tom J.; Barry, Thomas E.; Dacin, Peter A.; Gunst, Richard F. (2005),
‘Spreading the Word: Investigating Antecedents of Consumer’s Positive Wordof-Mouth Intentions and Behaviors in a Retailing Context’, Journal of the
Academy of Marketing Science, Vol. 33, No. 2, pp. 123–138.
Celso Augusto de Matos; Carlos Alberto Vargas Rossi (2008), ‘Word-ofmouth communications in marketing: a meta-analytic review of antecedents
and moderators’, Journal of the Academy of Marketing Science, Vol. 36,
pp. 578–596.
See http://www.service-js.jp/.
Oliver, Richard, L.(1980), ‘A Cognitive Model of the Antecedents and Consequences of Satisfaction Decisions’, Journal of Marketing Research, Vol. 17,
pp. 460–469.
Yi, Youjae(1990), ‘A Critical Review of Consumer Satisfaction’, V. A.
Zeithaml ed., Review of Marketing, AMA, pp. 68–123.
40
40.
41.
2 Evolution of the Comprehensive Decision-Making Process …
Westbrook, Robert A. (1987), ‘Product/Consumption-Based Affective
Responses and Postpurchase Processes’, Journal of Marketing Research, Vol.
24, pp. 258–270.
Szymansky, David M.; Henard, David H. (2001), ‘Customer Satisfaction:
A Meta-Analysis of the Empirical Evidence’, Journal of the Academy of
Marketing Science, Vol. 29, pp. 16–35.
Chapter 3
Measuring the Impact of a Blog:
Quantitative and Qualitative Aspects
In Chap. 1, it was elucidated that SNS is often used by young people at the time
of a product purchase. Using real data, Chap. 2 demonstrated that there are some
pioneering conceptual models of information communication that consider the effect
of SNS in Japan and these concepts are valid. The marketing and science research
group sponsored by the Japan Marketing Association (JMA) measured the effect
of blogs quantitatively as well as qualitatively over two years, that is, FY2006
and FY2007. Given that SNS consisted mainly of blogs prior to the emergence of
Twitter and Facebook, measuring its effect is significant even today when it comes
to reviewing the role of SNS. In this chapter, literature related to the Internet will be
reviewed and the effect that the quality and quantity of blog posts have on purchasing
will be checked using the data from this study group.
3.1 Review of Previous Studies on Consumer Behaviors
Related to Information on the Internet
Ever since the advent of the Internet, its effect has been discussed in many research
works across domains, including the world of marketing. Initially, the main research
areas were information posted by companies and sales on the Internet, such as the
effect of online banner advertisements, studies on e-commerce, and the effect of
homepages. For example, Rebstein (2001)’s study on e-commerce mentioned that
the product categories expected to propel sales were gardening-related and consumer
goods, and that households without children and with an average amount of income
can be considered as promising potential customers.1 In addition, the study by
Briggs and Hollis (2001) on the effect of banner advertisements by apparel brands
This chapter is a revised and corrected version of Akira Shimizu, Yoshiko Takayama, Soh Hasegawa,
(2008), ‘Effects of—thinking Blog about both qualitative and quantitative effects’, Marketing
Journal, Japan Marketing Association, Vol. 110, pp. 63–77 (written in Japanese).
© Springer Nature Singapore Pte Ltd. 2021
A. Shimizu, New Consumer Behavior Theories from Japan, Advances in Japanese
Business and Economics 27, https://doi.org/10.1007/978-981-16-1127-8_3
41
42
3 Measuring the Impact of a Blog: Quantitative …
demonstrated that banner advertisements can firmly communicate the messages they
intended to convey.2 The FY2000 marketing and science research group sponsored
by the JMA studied the difference between Internet users and non-users when it
comes to purchasing a car. It was revealed that users and non-users differed in the
number of car models they listed as final candidates. Furthermore, Internet users had
more candidates and were more likely to have car models they rejected (Shimizu
2006).3 As described, early studies focused on measuring the effect of the information companies post by using new media, such as “how the new information media
known as the Internet affects purchase behavior” and “what segments of people
respond to the Internet,” just like the effects of advertisement and promotion were
measured.
Meanwhile, as consumer networks are built, the mainstream of studies shifted
to communication among consumers, such as online communities and blogs, which
cannot be controlled by companies. If we were to say that studying the effect of
corporate communication in terms of how the information posted on the Internet by
company affects consumers is the first stage of research, considering the interaction
among consumers is the second stage of research. We can say the web is evolving
to become “something of consumers” from “something of companies.” This second
stage of research covers a wide range of topics including the effect of reviews posted
on the site, the effect of blog, and the effect of online communities. The following
are the major studies.
First, as a study that measured the effect of online reviews, there is a study by
Chatterjee (2001) who used students as his subjects and checked their usage of
such reviews.4 Specifically, he divided the students into those who purchased their
textbooks on a familiar site and those who purchased based on price when they
shopped online, and explored what effects negative reviews had on retail sites. The
results indicated that while students who purchase on a familiar site did not read
negative reviews and did not even try to look for the negative information, students
who purchased on an unfamiliar site if the price was low tended to proactively search
for and read negative reviews and switch to another site accordingly. This result
indicates that while negative rumors do not matter much if there is a strong emotional
relationship, they will work as a trigger to switch sites when the relationship is based
only on pricing.
In order to explore how consumers use online recommendations, Senecal and
Nantel (2004) surveyed 487 subjects who use the Internet to see what kind of sources
of recommendations are effective for these three products: computer mouse, calculator, and wine.5 First, it was confirmed there is a relationship between whether
there was a recommendation and online purchase (choice) under all products. Next,
on comparing calculator (which falls under “search goods”) and wine (which falls
under “experience goods”), the effect of recommendation on the experience goods
was larger. Whether the recommendation was by a commercial-based third party or
a non-commercial-based third party had nothing to do with the choice. It was also
confirmed that personalized recommendations are more effective than otherwise.
Chevalier and Mayzlin (2006) used online reviews on Amazon and Barnes &
Noble to see how they affected the fluctuations in sales ranking.6 The books used as
3.1 Review of Previous Studies on Consumer Behaviors Related to Information …
43
samples were those listed as best sellers between 1991 and 2002. According to the
study, the sales ranking is higher when the number of stars in the review is higher;
in particular, five stars helped to increase sales while one star drove the sales down.
Furthermore, it showed that longer reviews contributed more to the sales and, in
terms of the content, the effect of negative reviews driving the ranking down was
much larger than positive reviews driving the ranking up. It indicates that the negative
impact of bad reviews is large because the content of review generates a prospect
theory-type of reaction at the time of decision-making and that peripheral information
like the number of stars and the length of reviews also affects decision-making.
Next, a study by Gruen, Osmonbekov, and Czaplewski (2006) showed the effect
of online communities.7 They surveyed 616 members of a software product forum
(with 5000 registered members), used structural equation modeling, and checked
the effect of participating in online review forums and exchanging opinions. This
study confirmed that exchanging opinions on the forum resulted in a higher overall
rating on the software company as well as a stronger momentum to spread the word
about the product to others, even though it had no relationship with repurchasing the
product.
Studies have been conducted on the effect of online communities in Japan as well.
For example, Kanamori and Nishio (2005) demonstrated that the participation in an
online community affects the attitude toward products.8 They used the elaboration
likelihood model to hypothesize the mechanism of the effect that online community
has on the formation of user’s attitude toward brands and performed an analysis
by fitting it to the structural equation modeling. The sample size is 344 Internet
users. The results of the analysis showed that the perceived amount of comments by
those who used the product affected the potential customer’s cognitive attitude, and
the perceived amount of one-to-one comments which targets individual customer
affected his emotional attitude. Furthermore, while both emotional and cognitive
attitudes affect purchase intention, only the emotional attitude affected the intention
to recommend to others. The study is noteworthy as a study that considered both the
content and the amount of what is discussed in online communities, indicating they
lead not only to purchase but also to subsequent word-of-mouth.
Yamamoto (2005)’s study empirically showed the effect of online communities
by not only looking at the changes in attitude but also combining it with actual online
shopping records.9 In this study, a survey was conducted among @cosme members
with actual purchase history data to verify whether there is a relationship between a
customer’s visit to the website and actual purchase. @cosme is Japan’s largest SNS
site about cosmetics. According to the study, amount of “pre-purchase searches”
among read-only members (ROM) are significantly related to actual purchase. And
the users who visit this site for the purpose of making a fail-proof purchase decision
are becoming excellent customers. In contrast, in the case of members who post
comments at @cosme, the driver for purchase decision is the attachment to this site,
revealing the schematic in which members purchase on the linked site because they
like the @cosme site. This demonstrated the difference between those who post
comments and the ROMs.
44
3 Measuring the Impact of a Blog: Quantitative …
A study by Niederhoffer, Mooth, and Wiesenfeld (2007) measured the effect of
blogs.10 To examine how much influence SNS has on the sales of new consumer
packaged goods, they looked at 70 brands launched in the United States between
mid-2005 and end-2006, and rated them based on each of A.C. Nielsen’s attributes,
including purchase intention, purchase frequency, uniqueness of the product, value,
blog, sales, and media expenditure. First, it showed that the top ten reviews (the
ranking was made by the amount of marketing buzz) of new products accounted for
85% of all buzz. Furthermore, compared to reaching the sales peak, reaching the peak
of buzz (the peak number of buzz of that new product) only took two-thirds of time.
It means the time duration to reach the peak number of buzz is shorter than that of the
sales of the new product. Next, it is company’s advertisement expenditure that mostly
affects the amount of buzz. From the perspective of consumers, the penetration rate
of the category and the uniqueness of the product affected the amount of buzz. The
uniqueness of the product was also related to interest ingness, attractiveness, and
the degree of excitement. The products that had a high buzz ranking were also more
likely to be successful. However, no relationship was observed between the purchase
intention and buzz. There was a relationship between product sales and buzz, but
when it is combined with the score for product concept and the feedback after use,
the predictability for market share considerably improved. Based on this, it seems
there are certain effects of blogs—both in terms of quality and quantity—on sales.
There are also studies that went into depth and discussed the possibility of product
development by consumers. In a series of studies,11 Hamaoka states that although the
idea of marketing in the past is based on an assumption that consumers choose from
the given choices, they do have the ability to develop and create choices in addition
to merely choosing from the ones provided by companies (Hamaoka 2004, 2007).
He then presumes that more complex and sophisticated items are developed when
consumers are connected through online communities and, if this activity grows,
it could have a considerable impact on companies. Calling this “coevolutionary
marketing,” he also expresses the need for a new framework to study consumer
behavior.
As described, while the focus of studies on the Internet has shifted from the effect
of corporate advertisement on the Internet to how consumer communities affect the
sales of products, studies on mobile devices are also attracting attention nowadays.
Lamberton and Andrew (2016) reviewed how studies that combined digital, social
media, and mobile (DSMM) have emerged since 2000.12 They divided and organized
the time period into the first stage of 2000–2004, the second stage of 2005–2010,
and the third stage of 2011–2014. Based on this, in the first stage, DSMM were
studied as (1) platforms for self-expression, (2) a means of searching information at
the time of making a decision, and (3) a new marketing medium. They were studied
in the second stage as (1) the effect of online reviews on marketing and (2) how a
digital network creates information and value. In the third stage, they were studies
as (1) how comments by individuals become a tool to increase corporate marketing
activities, (2) how contents generated by users become a marketing tool, and (3)
how marketing intelligence is understood in a specific media platform. This research
3.1 Review of Previous Studies on Consumer Behaviors Related to Information …
45
review recognizes the fact that the information posted by consumers has become
more important for companies than ever before due to the emergence of mobile
devices, and shows that the effect is actually observed.
Based on these studies, it has become clear that SNS definitely has an impact.
In order to measure that impact, we need to consider not only whether consumers
came across information on SNS or not but also its contents, the interaction between
consumers, and even the devices used for the communication. In addition, past studies
suggest that the level of this impact depends on the product and the advertisements
ran by the company, considering there are synergistic effects.
3.2 Quantitative Measurement of the Effect of Blogs
As is evident from the above literature review, we can say that the information
posted by online communities is a new media that affects consumer’s information
searching behavior and ultimately product purchase. Thus, would the information
posted by such communities—including blog posts in a broad sense—really affect
sales in Japan too? With the above question in mind, the marketing and science
research group at the JMA measured the effect of blog posts both quantitatively and
qualitatively over two fiscal years: FY2006 and FY2007. FY2006 was for the study
on quantitative aspects and FY2007 was for the study on qualitative aspects. First, I
will discuss the quantitative effect.
In the FY2006 study, five product categories sold at supermarkets were considered
and an analysis as to how each brand’s sales trend is related to the number of searches
and the number of blog posts was conducted. The analysis period was 12 weeks, from
November 2006 to January 2007.
First, the data on blog posts was tabulated using Dentsu Buzz Research® . This is
a service that searches, analyzes, and displays word-of-mouth (buzz) on the Internet
such as blog posts and bulletin boards by using an independently developed search
engine. The service became available in the format of an application service provider
(ASP) in November 2005 as the first of its kind in Japan. The target sites for the search
at the time of survey included a total of 20 sites, consisting of 17 leading domestic
blogs and 3 bulletin boards such as 2 channel.13
For the search count data, the number of searches on Yahoo! Search was used.
Yahoo! Search is an online search service of Yahoo! JAPAN, which is an Internet
portal site with 1.6 billion page-views per day, accessed approximately by 50.93
million unique customers per month14 as of the time of survey.
For the sales data, they used the number of units sold per sample store from
the SRI data by INTAGE Inc. SRI is a syndicated data service that surveys 5051
stores nationwide to look at major retail businesses such as general merchandise
stores, supermarkets, convenience stores, pharmacies/drugstores, and home centers.
It provides analysis on their sales trend by collecting the POS data. The number of
units sold per sample store refers to the average number of units sold per surveyed
store.
46
3 Measuring the Impact of a Blog: Quantitative …
First, the relationship between sales and the number of blog posts, and the relationship between number of search and the number of blog post were measured by
using eight brands of chocolate and green tea. Here, the sales figures are standardized by setting the average sales volume for a given product category during the
survey period at 1. Tables 3.1 and 3.2 show the results on chocolate and green tea,
respectively.
First, with chocolate, seven out of the eight brands showed a strong correlation
between sales and the number of blog posts. All correlation coefficients are significant at the 1% level. The correlation reaches about 0.9 especially for new products
and limited seasonal products, indicating that the number of blog posts and sales
are highly correlated. In contrast, while green tea beverages showed a correlation
between sales and the number of blog posts under two brands, there was no correlation at all under the remaining six brands. However, a correlation was observed
between the number of searches and the number of blog posts under four out of the
eight brands. In the case of green tea beverages, blog posts had more influence on
the number of searches than sales volume.
Table 3.1 Correlation between number of blog posts and sales, number of search (chocolate)
Brand type
Brand name
Correlation between Sales
and number of blogs
Correlation between
number of blogs and
number of search
New product & limited
seasonality product
Brand A
0.98
0.94
Brand B
0.91
0.92
Brand C
0.87
0.96
Brand D
0.89
0.88
Brand E
0.86
0.88
Brand F
0.76
0.87
Brand G
0.85
0.80
Brand F
0.28
0.68
Regular product with
campaign
Others
Table 3.2 Correlation between number of blog posts and sales, number of search (green tea)
Brand type
Brand name
Correlation between Sales
and number of blogs
Correlation between
number of blogs and
number of search
Green tea product
Brand I
0.43
0.64
Brand J
0.28
0.19
Brand K
−0.41
0.27
Brand L
0.74
−0.07
Brand M
0.72
0.82
Brand N
0.21
0.32
Brand O
0.11
−0.51
Brand P
0.00
0.04
Healty green tea product
3.2 Quantitative Measurement of the Effect of Blogs
47
As described, it was difficult to make a general statement since the relationship
between the number of blog posts and the number of units sold was weak with
green tea beverages even though the relationship was observed with almost all of
the chocolate products. The survey by the Japan Newspaper Publishers & Editors
Association (Nihon Shinbun Kyokai, or NSK) presented in Chap. 2 showed that those
who refer to SNS at the time of purchase are young people and many of them are
female.15 It might be that the correlation between the number of blog posts and the
sales was observed with chocolate products because the target segment of chocolate
products and the segment of people who refer to blogs was the same.
As for why no relationship was observed for green tea beverages, it might be
because blogs might not be effective to begin with and they are insufficient to explain
the relationship because in store promotion (ISM) has a large impact on the number
of units sold in the case of products sold at supermarkets and drugstores. It might
also be that blog readers and the target segment are not the same since young women
are not the only ones who drink green tea. Furthermore, we would not expect the
number of blog posts to change the overall sales so much for the brands with a certain
level of market share because those brands have loyal users.
Looking at the content of blog posts, while many are about events for a given
brand or comments after tasting a new product when it comes to chocolate, with
green tea, many talk about the actors appearing in the TV commercials. Based on
this, we can see that it is insufficient to merely look at the number of blog posts.
If that is the case, what impact does the number of blog posts have? Of the products
sold at supermarkets, we selected 15 brands with a large number of blog posts and
conducted structural equation modeling (SEM) to explain the increase in sales by
the store-front sales situation, the number of blog posts, and the number of searches.
The increase in sales was set as the dependent variable in order to exclude the effect
of sales generated by loyal users since it accounts for a significant proportion. The
selected brands include six brands of beer and beer-like alcoholic beverages, four
brands of chocolate, and five brands of shampoo. We gathered the data in weekly
base (12 weeks), so total sample size is 180 (15 brands product 12 weeks). Figure 3.1
shows the analysis results. The GFI and AGFI, which indicate the fitness of the model,
are both above 0.9, and the RMSEA is within the acceptable range of 0.1 or less.
Therefore, we can say that this model is appropriate.
Based on this, it became clear that the sales situation—which is composed of
the average selling price and the percentage of stores carrying the product—and
the number of searches affect the increase in sales and the number of blog posts
has a strong impact on the number of searches. When the standardized parameters
are compared, while the sales situation is about three times more effective than the
number of searches, the number of searches is reasonably effective as well. In other
words, it showed that the number of blog posts affects incremental sales via the
number of searches, although the effect is lower than the sales situation. When there
is a large number of blog posts and the number of searches increases accordingly, it
indicates that the product is talked about on the Internet. This result shows that the
fact the product is being talked about on the Internet leads to an increase in sales.
48
3 Measuring the Impact of a Blog: Quantitative …
Fig. 3.1 Relationship between incremental sales and the number of blogs, the number of searches,
and sales situation
Lastly, we calculated the correlation between the number of searches and the
categories created by subjecting the content of blog posts to text mining. It turned out
that while the correlation with the number of searches was 0.342 when the blog post is
about TV commercial, it was 0.472 when the content is about event or campaign and
0.596 when it is about experience in use. It indicates that the content of blogs is also
important and a large number of blog posts by itself will not work; the content based
on consumers’ real experience mostly draws their interest and directs them to do the
search. This result is also consistent with the results of the aforementioned study
by Kanamori et al., which stated that usage experience affects cognitive attitude. In
other words, we can presume that word-of-mouth on purchasing experience increases
searches because it affects cognitive attitude, which is a strong attitude. As shown
in past studies, this suggests we need to understand the content—or the qualitative
aspect—of blog posts rather than just looking at the quantity when analyzing blog
posts.
3.3 Qualitative Measurement of the Effect of Blogs
Next, the qualitative effect of blogs is measured. The FY2007 marketing and science
study group analyzed the qualitative effect that blog posts have on the formation of
brand attitude. Here, we conducted an experimental study so that the effect of blogs
can be extracted and the content of the blog posts being browsed can be identified
simultaneously. In the actual experimental study, we asked the subjects to browse blog
posts after answering questions such as brand image and purchase intention, asked
the same questions again after they finished browsing, and recorded the change. As
3.3 Qualitative Measurement of the Effect of Blogs
49
for the browsing of blog posts, while the subjects were allowed to search and browse
blogs, we specified detailed search criteria and asked them to write down blog URLs
they browsed in order to ensure they browsed qualitatively consistent blogs.
There were two surveyed categories: automobiles and cosmetics. As for surveyed
brands, we chose two new automobile brands released in 2007 and two cosmetic
brands that do not advertise through mass media so that the effect of browsing
blog posts can be extracted as accurately as possible. The number of samples and
allocations (based on age) are as shown in Tables 3.3 and 3.4. For the automobile
survey, people who plan to buy a car within the next year were selected as subject.
For cosmetics survey, we selected female who buy cosmetics at department stores
and also are aware of the brands were selected for the subjects.
Table 3.3 The number of samples and allocations (Automobile)
A-type blog experiment
Brand X
20–29
30–39
40–49
Over 50
Total
Recognition with Brand X
19
19
18
17
73
Non recognition with Brand X
6
7
8
12
33
Brand Y
20–29
30–39
40–49
Over 50
Total
Recognition with Brand Y
21
16
19
19
75
Non recognition with Brand Y
6
5
7
9
27
20–29
30–39
40–49
Over 50
Total
22
26
25
24
97
20–29
30–39
40–49
Over 50
Total
23
25
23
22
93
S_type blog experiment
Brand X
Recognition with Brand X
Brand Y
Recognition with Brand Y
Table 3.4 The number of samples and allocations (Cosmetics)
Brand Q
20–24
25–29
30–34
35–39
40–44
45–49
Total
Experience with Brand Q
24
24
21
20
21
17
127
Non experience with Brand Q
26
22
18
18
12
17
113
20–24
25–29
30–34
35–39
40–44
45–49
Total
Brand P
Experience with Brand P
22
19
21
18
21
17
118
Non experience with Brand P
16
19
18
20
17
16
106
50
3 Measuring the Impact of a Blog: Quantitative …
In analyzing the effect of browsing blog posts, we considered the analysis results
in the previous section and divided the blog posts roughly into 2 types. First type is
mainly about the experience in using/purchasing and second type is reacting to the
information released by the company, such as advertisements and promotions. This is
because the above analysis revealed that personal usage experience is more effective
than the contents such as comments on the information released by the company.
Specifically, there are two groups:
• A-type of blog posts: Blog posts that are generated as a result of advertisement
or promotion. They are posted as general topic by bloggers regardless of their
purchase intention. They post the type of information that falls under “Attention”
in the AISAS purchasing process.
• S-type of blog posts: Blog posts that are established by purchasers where the real
voice of the purchaser is written. They post the type of information that falls under
“Share” in the AISAS purchasing process.
Since the A types deal mainly with the messages released by the manufacturers,
they should have an effect on improving the level of recognition when a new product
is introduced in the market although they would not affect purchase intention. In
contrast, we can presume the S types would affect the site visitor’s purchase intention
since the real voice of purchaser independent of the manufacturer is written. Note
that when reading A-type blogs, we created two groups based on whether or not
they were aware of the product in question, while those who read S-type blogs chose
those who were aware of the brand in question.
First, we checked how a subject’s purchase intention changed after browsing blog
posts. Figure 3.2 shows the scores for purchase intentions for the targeted automobiles
before and after reading the blogs. Subject’s purchase intention are gathered 7-point
scales ranging from 1 (strongly disagree) to 7 (strongly agree). As shown in Fig. 3.2,
when we look at it by the percentage of top two boxes based on a 7-point scale, it
went from 11.3 to 10.3% for Brand X and from 10.8 to 15.7% for Brand Y among
visitors to the A-type of blog posts, while it shifted from 14.4 to 18.5% for Brand
X and from 12.9 to 9.7% for Brand Y among visitors to the S-type of blog posts;
therefore, the changes were not necessarily large. However, when focused on the
percentage of “neither,” it went from 29.2 to 19.8% for Brand X and from 31.4 to
19.6% for Brand Y among visitors to the A-type of blog posts, while it went from
29.9 to 27.8% for Brand X and from 39.8 to 29.0% for Brand Y among visitors to
the S-type of blog posts; thus, the percentage has considerably declined in all except
Brand X among visitors to the S-type of blog posts. We can see that browsing blog
posts has an effect to drive the attitude toward either positive or negative.
Although these interesting results were derived, these results were not statistically supported. For A-type blog posts, difference of the score of purchase intention
between before and after reading blog posts of Brand X is not significant (before
reading blogs: M = 4.01. SD = 1.39, after reading blogs: M = 3.94, SD = 1.55,
t (210) = 0.326, p < n.s.). Also, for Brand Y it is not significant (before reading
blogs: M = 4.07. SD = 1.58, after reading blogs: M = 4.04, SD = 1.68, t (202) =
Fig. 3.2 Purchase intention changed after browsing blog posts (Automobiles)
3.3 Qualitative Measurement of the Effect of Blogs
51
52
3 Measuring the Impact of a Blog: Quantitative …
0.129, p < n.s). For S-Type blog posts, outcome is also same: difference of the score
of purchase intention between before and after reading blog posts of Brand X is not
significant (before reading blogs: M = 4.33. SD = 1.31, after reading blogs: M =
4.33, SD = 1.42, t (192) = 0.455, p < n.s.), and also for Brand Y, it is not significant
(before reading blogs: M = 4.1 SD = 1.43, after reading blogs: M = 3.87, SD =
1.54, t (184) = 1.036, p < n.s.). This is probably because the effect of positive reviews
was canceled out given that these blog posts included not only positive ones for a
given product but also negative ones. East, Hammond, and Lomax (2008) said that
the impact of both positive and negative word-of-mouth (WOM) is strongly related
to the pre-WOM probability of purchases, and the impact of positive WOM is generally greater than negative WOM.16 For this reason, we need to experiment with a
selection of positive blogs to see exactly whether purchase intentions change before
and after reading a blog.
With the experimental survey on cosmetics, we limited blog posts for browsing to
the S types (two cosmetic brands that we chose for this experiment did not advertise
through mass media, so it had no A-type blog posts) with positive reviews on the usage
and divided the subjects into those who have used the given brand and those who
have not. Just like in the case of automobiles, we asked the subjects to autonomously
search and browse among the URLs that contain only specified positive contents and
compared the purchase intention and product rating before and after browsing.
Figure 3.3 shows the change in purchase intention after browsing blog posts,
tabulated among those who have not used and already used each brand. The number
of samples is 113 for Brand Q and 106 for Brand P of non-user, and the number
of samples is 127 for Brand Q and 118 for Brand P of user. Based on the figure,
the percentage of top two boxes out of the seven-point scale, i.e., the percentage
of purchase intention of non-user, has increased after browsing blog posts in both
cases: from 6.2 to 13.3% for Brand Q and from 7.5 to 15.1% for Brand P. For user,
the percentage of purchase intention of top two boxes has increased after browsing
blog posts in both cases: from 38.5 to 51.9% for Brand Q and from 32.2 to 43.2%
for Brand P.
Of these results, the difference in purchase intention scores between before and
after reading the blog was significant for both non-users and users. For non-users,
difference of the score of purchase intention between before and after reading blog
posts of Brand Q is significant (before reading blogs: M = 4.11. SD = 1.07, after
reading blogs: M = 4.53, SD = 1.14, t (224) = −1.884, p < 0.001), and for Brand P
it is also significant (before reading blogs: M = 3.92. SD = 1.17, after reading blogs:
M = 4.45, SD = 1.34, t (210) = −3.060, p < 0.001). Although the significance level
was only at the 10% level, it was confirmed that purchase intentions changed before
and after reading the blog for users: Brand Q (before reading blogs: M = 5.01. SD
= 1.32, after reading blogs: M = 5.31, SD = 1.42, t (252) = −1.780, p < 0.076),
Brand P (before reading blogs: M = 5.03 SD = 1.07, after reading blogs: M = 5.28,
SD = 1.19, t (234) = −1.668, p < 0.097). From here, we can see that the purchase
intention increases by browsing blog posts if the effect of negative reviews can be
Fig. 3.3 Purchase intention changed after browsing blog posts (Cosmetics)
3.3 Qualitative Measurement of the Effect of Blogs
53
54
3 Measuring the Impact of a Blog: Quantitative …
eliminated. And also, the increase in purchase intention scores for non-users was
found to be more pronounced before and after reading the blog posts compared to
users. Therefore, companies are required to have measures to promote positive blog
posts.
3.4 Conclusion and Future Implication
In this chapter, I looked at the role of a blog based on its quantitative and qualitative
aspects. As a result, it became clear from the quantitative aspect that there is a product
category (chocolate) in which a relationship is observed between absolute sales and
the number of blog posts and a product category (green tea beverages) in which the
relationship is not observed. We were unable to identify the rule “increased number
of blog posts = increased sales.” However, when we performed structural equation
modeling for the 15 brands that were mentioned in many blog posts by setting the
increase in sales as an explanatory variable, it became clear that the number of
searches on the given product also affected the increase in sales and the number of
searches was affected by the number of blog posts. From here, it was found that a
large number of blog posts does not necessarily result in sales like promotions do;
there is an indirect effect where the increased number of blog posts generates buzz
on the Internet which increases the number of searches and affects sales. It was also
revealed that the contents become highly correlated with the number of searches when
they are about real, personal experience while the correlation between commercial
reviews and the number of searches does not become strong.
Next, the qualitative aspects confirmed that when the subjects read unscreened
blog posts including positive ones and negative ones, the difference in purchase
intention scores between before and after reading the blog was not significant for
both A-type blog posts and S type blog post. The results show that reading blogs
does not increase consumer purchase intention. We hypothesized that this was due
to a mix of positive and negative blogs, and in a cosmetic experiment, we asked
subjects to read only positive blogs to see if their purchase intentions changed before
and after reading the blogs. The results showed that purchase intention scores after
reading the blog were significantly higher than before reading it, especially for those
who had not used that cosmetic product. In other words, the experiment revealed
that if companies are looking to raise purchase intentions for new products through
consumer word-of-mouth, they need to get consumers to spread positive word-ofmouth to be effective.
The analysis in this chapter showed that when measuring the effect of a blog,
it is useful to understand the aspect quantitatively as well as qualitatively. Going
forward, it is necessary to verify these by running hypothesis tests. In addition,
while this chapter did not go into the details on the relationship with other media,
3.4 Conclusion and Future Implication
55
exploring those relationships should further clarify the difference in the role of media
and the relationship with consumer’s decision-making process. I will take those into
consideration and conduct further analysis in the subsequent chapters.
Notes
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Reibstein, David J., (2001), ‘Internet Buyer’, Wind, Jerry & Mahajan, Vijay
ed. ‘Digital Marketing’, John Wiley & Sons, Inc., pp. 201–225.
Briggs, Rex; Hollis, Nigel (2001), ‘Advertising on the Web: Is There Response
before Click-Through?’, Sheth, Jagdish N; Eshghi, Abdolreza; Krishnan,
Balaji C. ed. ‘Internet Marketing’, Harcourt Collage Publishers, pp. 308–328.
Shimizu, Akira (2006), Strategic Consumer Behavior, Chikura-Shobo (written
in Japanese).
Chatterjee, Patrali (2001), ‘Online Reviews: Do Consumers Use Them?’,
Advances in Consumer Research, Vol. 28, pp. 129–133.
Senecal, Sylvain; Nantel Jacques (2004), ‘The influence of online product
recommendations on consumers’ online choices’, Journal of Retailing, Vol.
80, pp. 159–169.
Chevalier, Judith A.; Mayzlin, Dina (2006), ‘The Effect of Word of Mouth
on Sales: Online Book Reviews’, Journal of Marketing Research, Vol. 38,
pp. 345–354.
Gruen, Thomas W.; Osmonbekov, Talai; Czaplewski, Andrew J. (2006),
‘eWOM: The impact of customer-to-customer online know-how exchange
on customer value and loyalty’, Journal of Business Research, Vol. 59,
pp. 449–456.
Kanamori, Tsuyoshi; Nishio, Chizuru (2005), ‘The Effects of Net Community for Making the Brand Attitude’, Bulletin of Nikkei Advertising Research
Institute, Vol. 221, pp. 65–76 (written in Japanese).
Yamamoto, Hikaru (2005), ‘Understanding the Online Community Member
and Their Purchase Behavior: A Structural Modeling Approach’, Advances in
Consumer Studies, Vol. 11, No. 1/2, pp. 35–49 (written in Japanese).
Niederhoffer, Kate; Mooth, Rob; Wiesenfeld, David; Gordon, Jonathon (2007),
‘The Origin and Impact of SNS New-Product Buzz: Emerging Trends and
Implications’, Journal of Advertising Research, Vol. 47, pp. 421–426.
Hamaoka, Yutaka (2004), ‘Coevolutionary Marketing—New Concept of
Marketing under Age of Information Network’, Mita Business Review, Vol.
47, No. 3, pp. 23–36 (written in Japanese), and Hamaoka, Yutaka (2007), “ Coevolutionary Marketing 2.0: Toward Dynamic Analysis of Community, Social
Network, and Creativity”, Mita Business Review, Vol. 50, No. 2, pp. 67–90
(written in Japanese).
Lamberton, Cait; Andrew T. Stephen (2016), “A Thematic Exploration of
Digital, Social Media, and Mobile Marketing: Research Evolution from 2000
to 2015 and an Agenda for Future Inquiry,” Journal of Marketing, 80 (6),
146–172.
56
13.
14.
15.
16.
3 Measuring the Impact of a Blog: Quantitative …
By accessing with the ID and password issued after signing up, the users
of Dentsu Buzz Reserach® can enter from own PC suitable keywords such as
names of company and product they want to know about in terms of reputations
on SNS and find out the state of buzz almost in real time and going back for a
certain period.
The number of unique visitors for Yahoo! JAPAN was calculated by using
the data “87.5% viewing rate from home” and “89.8% viewing rate from
workplace” published in Nielsen Online’s NetView AMS JP in April 2008 and
assuming that there are approximately 58.12 million Internet users accessing
from home or workplace (Source Neilsen Online “Basic Internet Survey”).
‘Research of Media usage and Consumer Behavior in 2006’, The Japan
Newspaper Publishers & Editors Association.
East, Robert; Hammond, Kathy; Lomax, Wendy (2008), ‘Measuring the impact
of positive and negative word of mouth on brand purchase probability’,
International Journal of Research in Marketing, Vol. 25, pp. 214–225.
Chapter 4
Studies on Commitment
In Chap. 2, it was established that for consumers to launch a positive word-of-mouth
after making a purchase, it is important for them to not only become aware of and
interested in the brand at the initial phase of decision-making but also become satisfied and committed after making the purchase. In this chapter, the concept of commitment and its practical applications are considered. Commitments can be categorized
into several categories based on the source; however, brand commitment, which is a
concept related to satisfaction with and passion for the brand after making a purchase,
as described in Chap. 2, is discussed in this chapter. In this chapter, purchasing products sold at supermarkets is taken as an example and the role played by commitment
is discussed.
4.1 Current State of Retailing in Supermarkets
It is said that the advertising expenditure in Japan is currently approaching almost
JPY6.6 trillion. Of these, advertisements through the TV media incur about JPY1.8
trillion, while online advertisements account for over JPY2.1 trillion and the
remaining JPY2.2 trillion is allocated for sales promotion.1 Even though the growth
of online advertisement is considerable, the expenditure on sales promotion remains
the same as it has been five years ago, and sales promotion continues to be regarded as
a highly effective method. Many previous studies have mentioned that sales promotion is very effective in terms of short-term sales, especially at supermarkets and
drugstores (Totten, Block 1994).2 The impact of sales promotion has already been
reviewed in many studies. The main accomplishments of promotion are categorized
into (i) promotion in general; (ii) promotion tool; and (iii) psychological impact
on consumers (Blattberg and Neslin 1990, Khan and McAlister 1997, Onzo and
Moriguchi 1994, Shimizu 2004).3
Initially, looking at sales promotion in general, it has been shown that the shortterm effect is very large compared to the advertisement on mass media and, although
© Springer Nature Singapore Pte Ltd. 2021
A. Shimizu, New Consumer Behavior Theories from Japan, Advances in Japanese
Business and Economics 27, https://doi.org/10.1007/978-981-16-1127-8_4
57
58
4 Studies on Commitment
it differs by product category, sales increases by about 2.8 times on average during
a promotion. It is said that combining discounts, end island promotion, and flyers is
particularly expected to produce synergetic effects and combining it with advertising
is also effective. In the case of products sold at supermarkets and drugstores, more
promotion leads to higher sales. However, the growth of sales slows down once the
discount rate exceeds 30%.
When a promotion is running, purchase is made not only by the switchers but also
by the brand’s loyal users. In other words, the increase in sales is accounted for by
the purchases made by those who usually do not purchase, and the bulk purchases
made by loyal users. Promotion also evokes interest among consumers to make an
unplanned purchase and increase the money spent per shopping. Given that there are
many full-time homemakers, a lack of large storage spaces at home, and there is a
preference for fresh foods, consumers in Japan head out for shopping very frequently.
In addition, brand competition and promotion always happens in the marketing arena.
For these reason, the unplanned purchase rate is higher in Japan than in the United
States where people shop during weekends to stock up. The unplanned purchase
rate in Japan is estimated to be nearly 80% of all purchased items. Though recently,
US study has indicated that the percentage of unplanned purchases is growing with
increasing promotional expenditure, and 70% of all purchased items are unplanned
purchases today, the rate of unplanned purchase rate in Japan is still higher than in
U.S (Inman 2012).4
Flyers and end island promotions are often studied measuring the effect of a
promotion tool. End island promotions have a larger effect on sales while flyers
have a larger influence in driving consumers to the store than selling products
listed on it. Since these two promotions often run in conjunction with discounts,
there are also many studies on their synergetic effects. Among the promotion tools,
coupons are highly effective in attracting loyal users and driving even consumers
who made a trial purchase to purchase again. Many recent studies have been specifically done to measure the effect of coupon distribution via mobile devices. Various
loyalty programs that use coupons have also been implemented (Danaher, Smith,
Ranasinghe, and Danaher 2015).5
As described, previous studies have shown that the effect of promotion on sales is
significant and valid. On the sales front in Japan, where competition is fierce, there
is always some kind of brand running a promotion. As a result, the percentage of
purchases made by consumers at the time of promotions is increasing more than
the purchases during regular pricing. Another characteristic feature in the Japanese
society is that loyal users often make purchases at the time of promotions.
In that case, the attitude of the consumers when they make a purchase at the time
of promotion should be analyzed. While previous studies on sales promotion have
rarely looked into consumers’ reaction and attitude toward promotion, they have
largely reviewed on discounts on the reference price.
A “reference price” is the price that consumers refer to at the time of purchase.
Internal reference prices are associated with memories from past purchase experiences, whereas external reference prices are formulated from external information
such as price tags at stores and flyers with discount prices presented during a sales
4.1 Current State of Retailing in Supermarkets
59
promotion. While consumers presumably make decisions using both external and
internal reference prices, the weightage depends on the product category. It has been
shown by previous studies that the weightage for an internal reference price is high
for supermarket items on which consumers have many opportunities for repeated
purchase and have a lot of knowledge on price. Conversely, the weightage for an
external reference price is more likely to be used for items such as home appliances
with long purchase intervals and intense technological advancement as the prices at
which consumers purchased them in the past do not help (Shimizu 2004).6
The purchasing behavior of consumers changes based on promotion when the
external reference price is lower than the internal reference price, posing a large
difference. In such a case, consumers perceive the external reference price presented
to them to be cheap and likely to end up purchasing. However, since price is also a
barometer of quality, consumers doubt the quality of the product and do not respond
as much if the external reference price is too low and too far from the internal
reference price. Urbany, Bearden, and Weilbaker (1988) discuss the effect of the
external reference price on the internal reference price.7 They demonstrated that the
external reference price affects the internal reference price, if the quality of product
can be deemed the same as the internal reference price constructed by the consumer.
If that is not the case, the external reference price will not affect the internal reference
price. In other words, when the external reference price is set low for reasons such as
product replacement and expiry of the “best-before” date, consumers conclude that
the external reference price was set due to its slightly inferior product quality and the
internal reference price they already had for that product is unlikely to be updated.
By contrast, when the external reference price is lowered, even though the product
quality is the same, consumers will become slow to respond to the same external
reference price when presented next time because their internal reference price would
have been downwardly updated. In highly competitive product categories, discount
promotions are frequently used without any particular reason for the purpose of
ensuring sales. However, since consumers’ internal reference price is updated during
each promotional cycle, it becomes necessary to present further discounts in order
to maintain the effect of a discount promotion, making the real retail price of the
product decline as a result.
In that case, to what extent can a discount be provided without affecting
consumer’s internal reference price? Kalwani and Yim (1992) discuss the effect that
the frequency and size of discount have on the internal reference price.8 According
to the investigation made by them, the internal reference price does not get updated
if discounts are awarded not more than thrice a week and it also does not get updated
if the discount rate was up to 20%. Conversely, consumers’ internal reference price
would continue to decline once it crosses these lines.
As described, one of the negative effects that promotion has on consumer awareness is that it reduces the internal reference price and slows consumers’ response to
discounts. Another important point is that promotion deteriorates the brand image.
Urbany et al. (1988) also discuss the image that discounts create for the product.9
They have shown that while the perceived value or the sense of bargain increases as
the discount rate increases, the credibility declines with an increase in the discount
rate.
60
4 Studies on Commitment
Suri, Manchanda, and Kohli (2000) empirically verified the relationship between
discount and credibility.10 They indicated that a discount is sufficiently recognized if
it is larger than a certain level. A large discount is rated low in terms of price appeal
and product credibility when compared to a moderate discount. The perceived quality
does not change depending on whether the discount is large or moderate. A large
discount would actually increase the sense of disappointment and also reduce the
perceived value and so on. Based on this, Suri et al. stated that there is an optimal rate
for discounts and when it exceeds that rate, it would lead to a very disadvantageous
situation for the product image.
In this way, promotion activities are effective in ensuring sales in the short term;
however, in terms of consumer awareness, it has been shown to lower their internal
reference price, slow their response to discounts, and negatively affect the credibility
and quality of the product. In other words, retailing based on promotion ensures
sales while shaving off the brand power, and the judgment as to whether it is the
right tactic would be dependent on the short-term and long-term effects. For this
reason, the following studies are conducted: (i) the one that divides the promotional
effect into short-term and long-term effects to explore the long-term effect (Gupta
1988)11 and (ii) the newly emerged, unconventional one that measures the effect of
a non-pricing promotion (Inoue 2012).12
When considering the post-purchase information sharing in addition to these
issues, conventional promotions do not have much effect. This is because unplanned
purchases prompted by promotions are stimulus-reactive purchases that do not really
involve recognition and interest as emphasized in the AISAS® . AISAS® is characterized by the assumption of proactive recipients and presumes the flow of “directing
attention to the product (Attention), becoming interested (Interest), searching on the
Internet and other sources (Search), subsequently purchasing the product or service
(Action), and posting a feedback on the Internet regarding the product purchase
(Share).” As shown in Chap. 2, the likelihood to share information after a purchase is
overwhelmingly higher when the product is purchased with recognition and interest.
Therefore, it is probably unlikely for consumers who were driven to purchase by
promotion to share information. Thus, this can hardly be called an appropriate
strategy when one wants to nurture consumers who post information and have them
on one’s side.
Considering the negative impact of promotions on consumers and brands, there
arises an issue in merely measuring the state of repeat purchase based on consumers’
purchase record without considering whether there was a promotion at the time of
purchase and then regarding it as an important loyalty indicator. Since consumers
generally become more loyal to a brand or manufacturer by making repeat purchases,
loyalty has been regarded as an extremely important scale for companies. This is
also evident from the fact that retailers identify good customers based on the RFM
indicator (i.e., based on the customer’s last visit to the store, frequency of customer
visit, and the amount of money spent during the visit). In fact, many studies on
loyalty programs have shown that consumers who are very loyal to a given retail
store bring great profits to that store (Woolf 1996).13 It has also been observed that
4.1 Current State of Retailing in Supermarkets
61
a brand favored by highly loyal consumers will increase the market share compared
to brands that are not (Shimizu 2006).14
However, as promotions are constantly rolled out on the sales floor of supermarkets
in Japan as described above, it is difficult to identify loyal users for a given brand
only based on purchase history. This is because consumers driven to make purchases
by discount promotions that are constantly taking place are probably included in the
“loyal users” measured solely by purchase history. In other words, loyalty, which is
an indicator of behavior, is insufficient to measure brand power. Indicators to rate
consumers at their attitude level would also become necessary.
4.2 Loyalty and Commitment
It is commitment that attracts attention as an approach for solving this loyalty
problem. Commitment is a concept that was developed during the discussion on relationship marketing. While loyalty captures the relationship with the company from
the behavioral perspective called purchase, commitment is considered to capture
the attitude toward the relationship between the seller and purchaser (Aoki 2004).15
When a consumer is committed to a certain brand, this consumer is loyal to that
brand. However, when a consumer is loyal to a certain brand, it does not necessarily
mean there is a commitment. In other words, commitment measures the emotional
aspects of consumers about the brand. It is, therefore, gaining attention as a concept
that captures the aspects that loyalty cannot.
Although commitment is said to be multi-dimensional, it is generally viewed as
two dimensional in many cases, deeming that it is composed of “affective commitment,” which indicates affirmative feeling and attachment as a motivation to maintain
the relationship, and “calculative commitment,” which signifies the practical feeling
arising after comparing the cost associated with ending or switching the relationship (Kubota 2006).16 Specifically, behaviors such as “I will look for a particular
brand even if it means going to another store” and “I will purchase a particular
brand even if it is somewhat expensive” are behaviors exhibited when the affective
commitment is strong, while behaviors such as “it’s too troublesome to switch to
another brand from the one I have been using,” “I always purchase it because it is
inexpensive,” and “as I am afraid it might not work if I switch to another brand” are
behaviors exhibited when the calculative commitment is strong. When considered
together with the discussions (up to Chap. 3) pertaining to the process of decisionmaking, calculative commitment can be regarded as a commitment that arises when
the decision is made out of habit or based on promotion without going through
the steps of recognition and interest. Previous studies have shown that affective
commitment becomes more important than calculative commitment in developing
long-term relationships (Kubota and Inoue 2004).17 Therefore, it seems important
to consider affective commitment among all commitments in terms of brand power
and subsequent information communication.
62
4 Studies on Commitment
In fact, a series of studies done by Inoue, exploring the relationship of these
commitments with brand power and market performance, led to interesting results
(Inoue 2008, 2009).18 Inoue explored the changes in the market positioning of the
brand and its background using a questionnaire survey. According to the survey, the
purchase frequency is very high when consumers show high commitment to that
brand. It reveals that there was another axis called fascination commitment, besides
affective commitment and calculative commitment, which indicates much stronger
emotion. The brands with large sales get not only fascination commitment but also
strong calculative commitment. Fascination commitment and affective commitment
positively affect the intention of customers to recommend the brand to others. This
series of studies has done by subjective questionnaire survey data, and not related to
real buying behavior data. However, it provides many insights such as (i) affective
commitment is effective at least when considering brand rating; (ii) it is necessary to win consumers with calculative commitment in order to ensure sales; and
(iii) emotional connection—fascination commitment and affective commitment—is
significant with respect to information sharing, and so on. It can be said that the
last point supports Katahira’s AIDEES model which deemed enthusiasm as key in
information sharing.
As described above, commitment, particularly affective commitment, is a very
useful concept in measuring brand power and promoting information sharing. It is a
concept that requires attention when selling frequently-promoted supermarket items
of the current times. While previous studies on commitment have been mainly based
on consumer survey data, real promotion data and purchase history data in addition
to survey data is used in this work (in the next section) to explore the relationship
between affective commitment and promotion and then the relationship between
affective commitment and brand power. Then, pointers to information sharing, which
is a focus of this book, are presented at the end.
4.3 Explanation and Analysis of the Data
The analysis in this chapter is performed using the Consumer Attitude and Behavior
Integration Evaluator (CABIE) which is a database that integrates consumer attitude and purchasing behavior data collected by the Distribution Economics Institute
Japan (DEIJ) (Teramoto 2006).19 Survey companies that collect data on customer’s
purchase history usually do not conduct an attitudinal survey among those customers.
However, in order to explore the relationship between attitude and behavior, samples
of consumers who purchased the surveyed products this time from the panel of data
members (about 1200 samples) were gathered, and the collected data on affective
commitment and purchase intention for those products was extracted by using the
four-point scale questionnaire method. The survey was conducted twice, in May
2005 and January 2007. The purchase history scale including loyalty was prepared
by using purchase history data for the past 12 months from the time the survey was
conducted (May 2004 to April 2005 and January 2006 to December 2006) and the
4.3 Explanation and Analysis of the Data
63
data on promotion was prepared from the POS and causal (state of sales promotion
implementation) data from June 2005 to December 2006. The analysis covered a
total of 44 brands of food, beverage, and seasonings, and so on. Due to the varied
number of purchasers, the most responded brand had 445 respondents while the least
responded brand had 7 respondents.
In this chapter, an analysis is conducted on this data placing focus on two points:
(1) the difference between top-selling brands and long-selling brands in terms of
purchasers and (2) the relationship between affective commitment and the manner
purchases are made. The objective of the first point is to look into the relationship
between the points for rating long-selling brands and the consumers who support them
when brands are becoming increasingly short-lived due to promotions. The goal of
the second point is to measure the effect of promotion on affective commitment to
show why promotions could become obstacles for brands.
4.4 Analysis on Customers of Long Sellers
Initially, I categorized these 44 brands into long-selling brand, category’s top brand,
new product, and others. Here, a “long-selling brand” is defined as a brand that has
been registered in the database from the beginning of the data collection period (July
31th, 1989), and “category’s top brand” signifies a brand that has most respondents
(i.e., purchasers) under the product category, established by the DEIJ. From this, we
selected the sports drink category and the curry roux category to see how consumers’
affective commitment and actual buying patterns differed between the top-selling and
long-selling brands in the category. The two product categories were chosen for two
reasons: (1) the top-selling brand in the category is not a long-selling product, and
(2) there are enough respondents to be analyzed. The variables used for comparing
the top-selling brand and the long-selling brands include the ones related to affective
commitment that can be obtained from the survey data and the ones on promotion
tendency prepared based on the respondent’s actual purchases during the survey
period. The number of samples surveyed was 197 in the sports drink category for the
category top brands, 121 in the long-selling brands, and 194 and 72 in the curry roux
category, respectively. Table 4.1 shows the results of a test of the difference in the
mean value of each item between the top and long-selling brands of sports drinks,
while Table 4.2 shows the results of a similar analysis for the curry roux category.
Based on Table 4.1, it is quite evident that category’s top brand and long-selling
brand differ considerably in terms of the purchase indicator of customers purchasing
those brands: “average buying price of the list price,” “percentage of money spent
at the time of end island promotion,” and “percentage of money spent at the time
of flyer advertising,” are significantly higher than that of long-selling brand. This
means that the top brand sell a higher percentage of their products when they are
promoted than long-selling brand. Although the results in the analysis for curry roux
are not as pronounced as in the analysis for sports drinks, Table 4.2 shows that, as
64
4 Studies on Commitment
Table 4.1 The difference in the mean value between the top and long-selling brands of sports
drinks
Survey data
Long-selling
brand
Category’s top
brand
Significant level
It is a familiar
product
3.55
3.43
*
It is a reliable
product
3.43
3.25
**
It’s a product that
suits my tastes
3.00
2.81
*
It is a product with
many discounts
2.28
2.78
***
It’s a product that’s
often on the flyer
2.40
2.73
***
I like the company
(manufacturer) of
this product
2.89
2.75
I like the advertising 2.60
and promotion of
this product
2.44
I like the packaging 2.60
design of this
product
2.57
I like the quality of
this product
2.91
2.69
I like the ease of use 2.95
of this product
2.70
**
I am satisfied with
this product
3.00
2.74
**
I would purchase
even if it’s more
expensive than
other products
2.30
2.05
**
If this product is not 2.24
available at the
store, I want to
purchase it some
other time
2.03
**
If you don’t have
this item in store,
I’d like to buy
another one
2.43
2.49
I want to purchase
this product even if
it’s not discounted
2.13
1.87
***
(continued)
4.4 Analysis on Customers of Long Sellers
65
Table 4.1 (continued)
Long-selling
brand
Category’s top
brand
Significant level
3.17
2.87
***
I would recommend 2.54
this product to a
friend or
acquaintance
2.24
***
Average buying
0.898
price of the list price
0.823
***
Percentage of
money spent at the
time of end island
promotion
0.166
0.364
***
Percentage of
money spent at the
time of flyer
advertising
0.153
0.179
***
I would buy this
product if it was
discounted
Purchase
indicator
***p < 0.001, **p < 0.05, *p < 0.10
with sports drinks, the purchase indicator “percentage of money spent at the time of
end island promotion,” becomes significant.
The top-selling brands within a category are bought when promotions are on,
so they are more likely to be perceived cheaply priced among consumers than their
long-selling counterparts. In both categories, sports drinks and curry roux, the product
image for discounts is significantly higher for category top-selling brands than for
long-selling brands: “It is a product with many discounts” and “It’s a product that’s
often on the flyer” are significantly higher than that of long-selling brand.
Also, the variables that showed a significant difference for consumers who buy
long-selling brand compared to top-selling brand were affective commitment variables. These include “I would purchase even if it’s more expensive than other products,” “I want to purchase this product even if it’s not discounted,” “If this product
is not available at the store, I want to purchase it some other time,” and “I would
recommend this product to a friend or acquaintance”. For sports drinks, four of
these five variables, and for curry roux, all five variables were significantly higher
for long-selling products than for the top category-selling brands. In both categories,
purchasers of long-selling brands are more likely to appreciate the quality of their
products than purchasers of top-selling brands, through word of mouth. This result
is consistent with the results of the analysis in Chap. 2. Chapter 2 showed that
word-of-mouth is often purchased through awareness and interest. When purchasing
a product during a promotion, consumers purchase products following a stimulusresponse model. Therefore, they are more likely to purchase without awareness or
interest.
66
4 Studies on Commitment
Table 4.2 The difference in the mean value between the top and long-selling brands of curry roux
Survey data
Long-selling
brand
Category’s top
brand
Significant level
It is a familiar
product
3.42
3.31
It is a reliable
product
3.38
3.12
***
It’s a product that
suits my tastes
3.33
2.97
***
It is a product with
many discounts
2.22
3.07
***
It’s a product that’s
often on the flyer
2.17
2.98
***
I like the company
(manufacturer) of
this product
3.13
2.92
*
I like the advertising 2.57
and promotion of
this product
2.47
I like the packaging 2.54
design of this
product
2.42
I like the quality of
this product
3.32
2.87
***
I like the ease of use 3.19
of this product
2.85
***
I am satisfied with
this product
3.24
2.85
***
I would purchase
even if it’s more
expensive than
other products
2.76
2.1
***
If this product is not 2.67
available at the
store, I want to
purchase it some
other time
2.12
***
If you don’t have
this item in store,
I’d like to buy
another one
2.29
2.57
***
I want to purchase
this product even if
it’s not discounted
2.69
2.01
***
(continued)
4.4 Analysis on Customers of Long Sellers
67
Table 4.2 (continued)
Long-selling
brand
Category’s top
brand
Significant level
3.33
3.03
***
I would recommend 2.69
this product to a
friend or
acquaintance
2.22
***
Average buying
0.851
price of the list price
0.873
Percentage of
money spent at the
time of end island
promotion
0.202
0.530
Percentage of
money spent at the
time of flyer
advertising
0.127
0.178
I would buy this
product if it was
discounted
Purchase
indicator
***
***p < 0.001, **p < 0.05, *p < 0.10
In other words, while it is important to win customers who respond to promotion
in order to ensure sales, it is necessary to gain support from customers with a strong
affective commitment in order to become a long-selling brand. It is noticed that
getting caught in sales and putting efforts only in promotion will not grow the brand
into a long seller. Since consumers with a strong affective commitment are likely
to spread the word as shown in the aforementioned study by Inoue, a brand could
potentially grow through word-of-mouth to become a long seller if it can win the
segment with a strong affective commitment from its time of launch. It is evident
that the concept of information sharing argued in the AISAS® and AIDEES has high
significance in building a brand.
4.5 Chronological Change in Affective Commitment
Next, the change in consumer’s affective commitment is looked upon in order to
explore the effect of promotion on affective commitment. As described earlier,
the CABIE has conducted two surveys—with a gap of about a year and a half in
between—among consumers with purchase history. The samples that responded to
both the first and second surveys were extracted and comparison is made on how
their brand assessment changes for two brands (Brands A and B); the sample was
68
4 Studies on Commitment
It is a familiar product.
It is a reliable product.
It's a product that suits my tastes.
It is a product with many discounts
It's a product that's oen on the flyer
I like the company (manufacturer) of this product
I like the adversing and promoon of this product.
I like the packaging design of this product
I like the quality of this product.
I like the ease of use of this product.
I am sasfied with this product.
I would purchase even if it’s more expensive than other products
If this product is not available at the store, I want to purchase it some other
me
I would purchase even if it’s not discounted
I would recommend this product to a friend or acquaintance
I don’t want to fail by switching to another product
It is too much trouble to switch to another product
0.000
2005
0.500
1.000
1.500
2.000
2.500
3.000
3.500
4.000
2007
Fig. 4.1 Brand assesement change for Brand A
relatively large with no product renewal or discontinuation. There were 15 individuals who purchased Brand A and 19 individuals who purchased Brand B as of the
first and second survey dates. They are shown in Figs. 4.1 and 4.2, respectively.
First, while the level of satisfaction with the product has increased for both Brands
A and B, the factors that make up the level of satisfaction are different. In the case
of Brand A, the satisfaction score has improved from 2.8 to 3.0 on average. In
terms of the variables that cause satisfaction, it is noticed that: (i) the variables
related to affective commitment of the product (“I would purchase even if it’s more
expensive than other products,” “I would purchase at another store if this product
is not available at the store,” “I would purchase even if it’s not discounted,” “I
would recommend to my friends and acquaintances”) are almost similar to those in
the previous year; (ii) the variables related to familiarity (familiar, credible, it’s a
product that I like) have also improved when compared to the previous year), the
scores of the variables related to the images of bargain (“it’s often on sale,” “it’s
often on the flyer”) have improved more considerably than the variables related to
affective commitment and familiarity. Based on this, it can be concluded that the
level of satisfaction with Brand A has been primarily due to promotion.
Meanwhile, the average satisfaction level with Brand B has also improved from
3.0 to 3.4. Looking at the details, the scores of the variables related to affective
commitment (“I would purchase even if it’s more expensive than other products,”
“If this product is not available at the store, I want to purchase it some other time,”
and “I would purchase even if it’s not discounted”) and the variables related to
4.5 Chronological Change in Affective Commitment
69
It is a familiar product.
It is a reliable product.
It's a product that suits my tastes.
It is a product with many discounts
It's a product that's oen on the flyer
I like the company (manufacturer) of this product
I like the adversing and promoon of this product.
I like the packaging design of this product
I like the quality of this product.
I like the ease of use of this product.
I am sasfied with this product.
I would purchase even if it’s more expensive than other products
If this product is not available at the store, I want to purchase it some
other me
I would purchase even if it’s not discounted
I would recommend this product to a friend or acquaintance
I don’t want to fail by switching to another product
It is too much trouble to switch to another product
0
0.5
2005
2007
1
1.5
2
2.5
3
3.5
4
4.5
Fig. 4.2 Brand assesement change for Brand B
calculative commitment (“I don’t want to fail by switching to another product” and
“It is too much trouble to switch to another product”) has got better while the rating
on promotion (“it’s often on sale”) has declined. Based on this, it can be justified that
the level of satisfaction with Brand B is composed of increased affective commitment
and calculative commitment.
Table 4.3 shows the actual sales status of Brands A and B during this period,
compiled based on the POS data. Even though Brand A has better figures than Brand
B in terms of both sales share and the percentage of loyal customers, the average
discount rate and the percentage of sales at the time of bargain are much higher than
Brand B. This indicates that the current situation of sales is being earned through
Table 4.3 Actual sales status of brands A and B
Market Percentage Average
share
of sales at selling
(%)
the time of price
bargain
(JPY)
Highest
selling
price
(JPY)
Lowest Average Penetration
selling discount rate (%)
price
rate (%)
(JPY)
Percentage
of loyal
customers
(%)
Brand 10.9
A
92.8
177.7
338
138
24.6
22
41.7
Brand
B
86.5
210.7
338
153
18.1
11.2
22.8
7.4
70
4 Studies on Commitment
discounts. Although statistical verification is not possible due to the limited number
of samples at Figs. 4.1 and 4.2, it can be presumed from the above results that despite
Brand A possessing a higher market share than Brand B, the affective commitment
for Brand A has not improved as these market shares were achieved by running a
promotion more frequently than Brand B.
It has already been discussed in Sect. 4.1 that the internal reference price declines
when consumers purchase during promotion. Likewise, how will promotion affect
the development of commitment? According to Sato (2006), which explains the
relationship between consecutive purchases and commitment through a covariance
structure analysis on samples of the CABIE database for a single year, it has been
shown that consecutive purchases have a strong effect on both affective and calculative commitments.20 Sato’s study does not show the effect of promotion since
the analysis has been conducted by defining consecutive purchases as purchasing
consecutively regardless of the promotion situation. However, it is meaningful in
terms of showing that the purchasing situation affects commitment and the effect
of promotion. It is also consistent with the aforementioned results associated with
Brand B.
In this chapter, the same consumers’ survey data is used for investigation. A
structural equation modeling (SEM) was conducted on the data from the first period
(as of 2005) and the second period (as of 2007) along with the promotion touchpoint
of these subjects and purchase data during these periods. The analysis examines how
three factors—affective commitment, attachment to the product, and the image of
bargain—as of 2005 have changed the way the customers purchase during list pricing
and promotion by the time the 2007 survey was conducted. The DEIJ data, being a
panel data based on POS scan, also records the sales floor where the consumer made
a purchase as well as the status of flyer distribution. Therefore, it can be understood
under what kind of promotion consumers made purchases. Since it is impossible
to have a single brand with respondents at both points of time due to the nature of
the product category in which products are renewed and discontinued considerably,
brands that had samples for both points of time were added together and the data
was pooled. Ten product categories including vegetable juice, yogurt, and curry roux
were used for the analysis. The number of samples was 382. In order to eliminate the
effect of the product category each brand falls under, the average purchase indicator
score for each category is calculated, then it is subtracted from the brands under the
respective category, and the net score is used as the purchase indicator score of each
brand.
Figure 4.3 shows the results of this analysis. Here, a significant relationship at a
15% level is depicted by a solid line. As a note, since the notations get complicated, the
error terms corresponding to the measured variables of the model and the covariance
between these error terms are omitted. In addition, since some of the measured
variables used in each statement for affective commitment, product attachment, and
bargain image are similar, a covariance was assumed between the error terms of the
measured variables in order to improve the fitness of the model under the constraints.
The fitness of the model is checked first and then the parameters are interpreted.
The GFI, which indicate the fitness of the model, is above 0.9, and the RMSEA is
4.5 Chronological Change in Affective Commitment
2005 affective
commitment
0.671
F4
0.239
0.186
0.00
F2
0.06
2005 the image
of bargain
2006 affective
commitment
-0.228
F1
2005
attachment to
the product
71
-0.08
-0.086
2005 2006
Purchases during
promotion
0.08
0.611
-0.03
2005 2006
purchases during
list pricing
F8
F5
0.119
0.00
2006
attachment to
the product
F7
0.04
2006 the image
of bargain
0.529
F3
GFI
0.912 AGFI
F6
0.874 RMSEA 0.056 CFI 0.949
Fig. 4.3 Changes in shopping practices and commitment, attachment and image of bargain
within the acceptable range of 0.1 or less. So the fitness of the model is concluded
to be high.
Next, the figure is interpreted. The affective commitment, product attachment,
and bargain image formulated in 2005 strongly affects the affective commitment,
product attachment, and bargain image in 2007. It is confirmed that although the
consumer’s purchasing situation during the survey period affected the alreadyformulated consumer awareness and attitudes, the attitudes formulated up to the
previous year had a high impact.
In relation to the purchasing situation, the affective commitment has a negative
effect on purchases during promotion but a positive effect on purchases during list
pricing. Furthermore, when a purchase is made during promotion, it will have a
negative impact on the commitment in the following year. It implies that strong
affective commitment makes it easy to purchase at the list price and makes it difficult
to purchase under promotion, and that purchasing at the time of promotion will lower
affective commitment. In other words, it is creating a vicious cycle where those who
do not have a strong affective commitment are likely to purchase when promotion is
running and the act of purchasing at the time of promotion further reduces affective
commitment. This is important as it clearly reveals the negative effect of promotion.
It is consistent with the fact that the score of affective commitment for Brand A—the
one that actively runs discount sales of the beverages—has declined compared to
the previous year. While previous studies have shown that discounts reduce brand
value in relation to reference price (as I explained at Sect. 4.1), the results here also
revealed that promotion not only lowers the internal reference price but also weakens
the consumer’s affective commitment, leading to reduced brand value. This result is
consistent with the fact that long-selling products are favored by the segment with a
strong affective commitment.
Further, it is important to note that product attachment has a positive effect on
purchase at the time of promotion and the purchase at the time of promotion also
72
4 Studies on Commitment
positively affects the product attachment for the following year. This result indicates
that product attachment is strengthened by making purchases during a promotion.
It is seen from this result that even though product attachment has been considered
as a favorable scale, it is not necessarily the case. One needs to be careful when
interpreting the meaning of product attachment.
4.6 Conclusion and Future Implication
In this chapter, the concept of commitment—which affects consumer’s word-ofmouth activities after purchasing a product—along with studies on sales promotion
was discussed. Initially, it was shown by reviewing previous studies that while promotion has a large positive effect on sales, it has a negative effect on consumer’s emotion
with respect to the reference price. Specifically, it meant that promotion deteriorates
consumers’ sense of trust and pricing for the brand even though it helps to increase
sales temporarily. For this reason, it was argued that repeated purchasing at the time
of promotion can hardly qualify a consumer as a true loyal user and that it was necessary to measure consumers’ affective commitment at the time of making a purchase.
Then, while referring to previous studies, the relationship between promotion and
affective commitment was investigated by using the CABIE, a database of the DEIJ.
At first, comparison was made between the top-selling and long-selling brands
in two product categories, sports drinks and curry roux, to see how they are bought,
and the affective commitment and price sensitivity of consumers who buy them.
The results revealed that top-selling brands are more sold at the timing of promotion than that of long-selling brands. The top-selling brands were significantly more
likely to be sold at a discount than their long-selling counterparts, as they were more
likely to be bought during promotions. Also, the top-selling brands in terms of affective commitment scores were significantly lower than their long-selling brands. The
results show that sales that rely on promotions are effective for building top-selling
brands, but to build a long-selling brand, you have to raise affective commitment.
Next, a questionnaire survey on affective commitment was conducted twice among
the same sample population that had purchase history by spacing them by year and a
half to investigate the change in affective commitment and the way consumers make
purchases. First, two brands under the beverage category were examined to see the
changes in affective commitment and the sales situation of those two brands during
that period. It showed that even though Brand A, which ran promotions frequently
during the survey period, had a higher share/loyalty based on the POS data than
Brand B, which did not run many promotions, the second score of affective commitment did not change from the first one and the image of bargain had increased quite
considerably. The increase in the level of satisfaction was not very large, either. By
contrast, Brand B improved the score of affective commitment and had a very large
increase in the level of satisfaction. Though the outcome was not supported significantly, it was inferred that there is a relationship between promotion and affective
commitment.
4.6 Conclusion and Future Implication
73
In order to confirm this relationship between promotion and affective commitment,
brands that continuously existed for two years were extracted from a group of brands
that were continuously purchased by the survey respondents for two years. They were
then analyzed by using the structural equation modeling to understand the changes
in consumers’ purchasing situation and affective commitment. The results revealed
that consumers with a strong affective commitment tend to purchase the products
at the list price and not purchase when promotions are running. Furthermore, the
affective commitment declines among consumers who purchased when a promotion
was running, confirming that promotions weakened affective commitment. While
previous studies noted that promotions lower the value of a brand, this result showed
that weakened affective commitment among consumers who received promotions
was one of the reasons. This result shows the weaknesses of promotion-dependent
sales.
As described, besides showing that affective commitment was helpful in building
a long-selling brand, it was demonstrated that purchases under the condition where
a promotion is running lower affective commitment. As it was clear from previous
studies, since affective commitment has a large effect on consumers’ intention to
recommend, sales methods that rely too much on promotion are problematic when
the effect of information sharing typified by word-of-mouth was considered.
Notes
1.
2.
3.
4.
5.
6.
7.
Dentsu Inc. (2020), https://www.dentsu.co.jp/news/release/2020/0311-010
027.html.
Totten, John C.; Block, Martin P. (1994), Analyzing Sales Promotion, Dartnell,
p. 70.
Blattberg Robert C.; Neslin Scott A. (1990), ‘Sales Promotion, Concepts,
Methods, and Strategies’, Prentice Hall, Barbara E. Kahn; Leigh McAlister
(1997), ‘Grocery Revolution: The New Focus on the Consumer’, AddisonWesley Pub. Co., Naoto Onzo; Takeshi Moriguchi (1994), ‘Sales Promotion–
Theory, Analysis, and Strategy’. DOBUNKAN SHUPPAN. Co., Ltd. (written
in Japanese), Akira Shimizu (2004), ‘Retail Strategy Based on The Client
Perspective’, Chikura Publishing Co., Ltd., (written in Japanese).
Inman, Jeff J. (2012), ‘Shopper Insights: What We Know and Where We Are
Headed’, Keio University Global-COE Special Session.
Danaher, Peter J.; Smith Michael S.; Ranasinghe Kulan; Danaher Tracey S.
(2015), ‘Where, When, and How Long: Factors That Influence the Redemption
of Mobile Phone Coupons’, Journal of Marketing Research, Vol. 52, pp. 710–
725.
Akira Shimizu (2004), ‘Consumer Perception of Price’ in ‘Retail Strategy
Based on The Client Perspective, Chikura Publishing Co., Ltd. pp. 163–197
(written in Japanese).
Urbany, Joel E.; Bearden, William O.; Weilbaker, Dan C. (1988), ‘The Effect
of Plausible and Exaggerated Reference Prices on Consumer Perceptions and
Price Search’, Journal of Consumer Research, Vol. 15, pp. 95–110.
74
4 Studies on Commitment
8.
Kalwani, Manohar U.; Yim, Chi Kin. (1992), ‘Consumer Price and Promotion
Expectation: An Experimental Study’, Journal of Marketing Research, Vol.
29, pp. 90–100.
Urbany et al., op.cit.
Suri, Rajneesh; Manchanda, Rajesh V.; Kohli, Chiranjeev S. (2000), ‘Brand
Evaluations: A Comparison of Fixed Price and Discounted Price Offers,
Journal of Product & Brand Management, Vol. 9, pp. 193–206.
Gupta, Sunil (1988), ‘Impact of Sales Promotions on When, What, and How
Much to Buy’, Journal of Marketing Research, Vol. 25, pp. 342–355.
Atsuko Inoue (2012), ‘The effect and possibility of non-price reduction promotion’, Bulletin of Nikkei Advertising Research Institute, Vol. 260, pp. 63–77
(written in Japanese).
Woolf, Brian P. (1996), Customer Specific Marketing: The New Power in
Retailing, Teal Books.
Akira Shimizu (2006), ‘Relationship between Attitude and Behavior’ in
Strategic Consumer Behavior, Chikura Publishing Co., Ltd., pp. 195–214
(written in Japanese).
Yukihiro Aoki (2004), ‘Product involvement and brand commitment’, in Shuzo
Abe ‘New direction of Consumer behavior study’, Chikura Publishing Co.,
Ltd., pp. 95–117 (written in Japanese).
Yukihiko Kubota (2006) ‘A Multidimensional Commitment Model for Relationship Marketing’, Journal of Marketing & Distribution, Vol. 9, No. 1,
pp. 59–86 (written in Japanese).
Yukihiko Kubota, Atsuko Inoue (2004), ‘Multi dimension of consumer
commitment and the effect of them’, Chukyo Kigyou Kenkyu, Vol. 26,
pp. 11–27 (written in Japanese).
Atsuko Inoue (2008), ‘The relationship between brand commitment and brand
position’, Journal of Distribution & Marketing, Vol. 470, pp. 16–26 (written
in Japanese); Atsuko Inoue (2009), ‘The Relationship of Brand Commitment
and Consumer Behavior’, Journal of Marketing & Distribution, Vol. 12, No.
2, pp. 3–22 (written in Japanese).
Takashi Teramoto (2006), ‘Brand segmentation based on consumer attitude
and buying behavior’, Journal of Distribution & Marketing, Vol. 440, pp. 4–10
(written in Japanese).
Eisaku Sato (2006), ‘How do the repeat buying strengthen the brand value?’,
Journal of Distribution & Marketing, Vol. 440, pp. 18–26 (written in Japanese).
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Chapter 5
Mechanism of Attitude Formation
for Consumers Who Convey Information
Based on the discussions in Chaps. 1 to 4, we were able to confirm that individuals
who were highly satisfied with the product they purchased are likely to spread the
word after the purchase, just as suggested in previous studies. However, the discussions also showed that the level of satisfaction alone cannot explain the behavior. In
response, the decision-making process leading to spreading of the word was examined in detail. It was observed that consumers who searched information with recognition and interest are likely to spread the word after purchasing a product. Furthermore,
consumers share positive reviews when they have affective commitment toward the
product and such reviews actually help potential customers in making a decision.
In other words, the findings indicated that we cannot accurately understand wordof-mouth behavior unless the decision-making process leading to the post-purchase
review and the mechanism in which affective commitment is formed are explained.
Even though explaining the process leading to the post-purchase review is important as described, there is no comprehensive decision-making model for consumers
which takes the post-purchase review into consideration, as outlined in Chap. 2. This
is because the purpose of the existing models is to explain the process that leads to a
purchase rather than explaining the process leading to communication of information
to others. Meanwhile, from the perspective of reaching consumers via advertising
messages, models such as AISAS® (attention → interest → search → action →
share) and AIDEES (attention → interest → desire → experience → enthusiasm →
share) do mention about the flow that leads to consumers’ word-of-mouth behavior.
However, these models do not consider attitude formation—which is regarded important in consumer behavior research—since the intention is to show the flow of information. Thus, there is no comprehensive decision-making model that takes findings from studies on consumer behavior—or the concepts such as attitudes toward
products, purchasing, and satisfaction—and combines them with word-of-mouth
communication after a purchase, which is actually practical knowledge.
Of the past comprehensive decision-making models, this chapter will focus
on the elaboration likelihood model (ELM), which takes the flow of advertising
messages into consideration to explain how attitude toward advertisements is
© Springer Nature Singapore Pte Ltd. 2021
A. Shimizu, New Consumer Behavior Theories from Japan, Advances in Japanese
Business and Economics 27, https://doi.org/10.1007/978-981-16-1127-8_5
75
76
5 Mechanism of Attitude Formation for Consumers Who Convey Information
formed, and consider how the attitudes that lead to word-of-mouth communication
are formed. The ELM has not been applied to product purchase because it is a
model for understanding the attitude toward advertising messages rather than the
attitude toward products. However, this seems like a suitable foundation to use
for creating a new comprehensive decision-making model because it takes into
consideration the concept of attitude—which is regarded important in past theories
of consumer behavior—and the concept of advertising message flow—which is
regarded important in AISAS® and AIDEES. In what follows, the ELM concept
will be explained, followed by the actual analysis.
5.1 The ELM Concept
As outlined in Chap. 2, the comprehensive decision-making process models for
showing how consumers determine their attitudes toward a product and eventually
purchase it are divided roughly into the stimulus-response type and informationprocessing type. The elaboration likelihood model (ELM) likewise shows how
consumers determine their attitudes toward advertising messages. Presented by Petty
and Cacioppo (Petty and Cacioppo 1983, Petty and Cacioppo 1986),1 this is a comprehensive model created by focusing on how consumers render meaning to and interpret
advertising messages (see Fig. 5.1). This model divides the process of interpreting
messages and forming attitude into the logically-interpreting central route and the
emotionally-interpreting peripheral route, and shows the difference in the strength
of attitude by route. The process that leads to determining the attitude is described
below.
Information
Motivation to elaborate
NO
Peripheral route
Ability to elaborate
Central
(cognitive)
processing
Peripheral
(emotional)
processing
NO
Central route
Attitude
Fig. 5.1 The ELM concept
5.1 The ELM Concept
77
The ELM first questions whether the consumer has the motivation to elaborate on
the message when he/she receives a new one. “Motivation to elaborate” refers to the
purpose behind promoting information processing; in the absence of such motivation,
consumers try to decide their attitudes by using peripheral clues. “Peripheral clues”
are indirectly related factors such as the color of the product and the credibility of the
originator of the message, although they are not directly related to the message itself.
When there is motivation, the next question is whether the consumer has the ability
to consider the message. If he/she does not have the ability to properly judge even
though there is motivation, the attitude is formed via the peripheral route at that point.
Attitudes are formed via the central route only when there is a positive motivation to
elaborate the message and the ability to process it. Furthermore, it is assumed that
the attitude is formed in the Fishbein style in this case. In other words, while the
central route is a decision-making route when there is a motivation and the ability
and the attitude are actively determined, the peripheral route is a decision-making
route when the attitude is determined based on peripheral information due to low
motivation or limited ability.
Since the strength of the attitude thus formed differs between these two routes,
i.e., the attitude formation by centrally changing the attitude versus peripherally
changing the attitude, the subsequent change in the attitude when encountering an
additional message is said to differ. In other words, while the change in attitude is less
likely to occur by seeing an additional message when the attitude was determined
via the central route since it was formed by carefully considering a given message,
the change in attitude is likely to occur and the subsequent decision making often
changes when the attitude was determined via the peripheral route since the attitude
is weak. For this reason, just how you can make consumers to build attitudes via
the central route becomes important in ultimately building a truly loyal user base by
preventing them to switch brands.
Needless to say, it is not that consumers use only one of the decision-making
routes; they make decisions using both the central and peripheral routes in the
majority of cases. There may be a case in which consumers switch from the peripheral route to the central one once their knowledge increases over the course of the
decision-making process. In other words, it becomes important to know which one is
heavily weighted rather than which one of the routes—the central or peripheral—is
used when making a decision.
As described earlier, the ELM assumed that the route on which the ultimate
attitude is formed differs by whether the consumer is willing to proactively gather
information and by whether he/she has the ability to judge that gathered information. Because the way it separated the two routes is easy to understand, there have
been many studies in this regard. According to Kitchen, Kerr, Schultz, McColl, and
Heather (2014), who reviewed studies on the ELM, since 1981, there have been 125
papers on marketing and advertising related to the ELM.2 They state that the ELM is
useful for advertising researchers studying attitudinal change because the model is an
influential, valuable, and popular framework. Speaking in the context of this chapter,
what is more important is not which of the two routes determined the attitude but
what the subsequent behavior is. This is because we can assume when considering
78
5 Mechanism of Attitude Formation for Consumers Who Convey Information
the characteristics of the ELM along with the findings up to Chap. 4 that consumers
provide positive reviews if they make a decision based majorly on the central route,
make a purchase, and gain satisfaction.
5.2 Direction of the ELM Application
As is clear from the above discussion, the ELM is a useful concept to understand
consumers’ decision-making process in purchasing a product that leads to word-ofmouth communication. However, very few studies have been done in this regard,
or in the context of the flow of consumers’ product purchase. This is because the
ELM, which is primarily intended to demonstrate the formation of attitudes toward
advertising messages, is not applied to studies to understand the formation of attitude
that leads to product purchase. In fact, many word-of-mouth studies that used the
ELM regarded online reviews as advertising messages and explored the formation
of attitude toward it.
For example, Yi-Hsiu and Hui-Yi (2015) used reviews on restaurants in Taiwan,
divided these online reviews into the central routes and peripheral route, and studied
which is useful in attitude formation. It showed that the number of followers, which is
considered the peripheral route, is more useful in attitude formation than the contents
of review, which are considered the central route.3 With online reviews, Ketron
empirically examined the contents, which are the central route, and the grammatical
and structural quality, which are the peripheral route. Based on this, it was confirmed
that the credibility of online reviews is affected not only by the contents but also by
the grammatical and structural factors (Seth 2017).4
The studies by Tam and Ho (2005) and Colicev, Malshe, Pauwels, and O’Connor
(2018) investigated how the message interpreted through the ELM affects the subsequent process of making a decision on product purchase.5 In their studies, reviews
and advertising messages on a given product were classified into the central route
information and the peripheral route information based on the idea of the ELM
in order to show at which stage of the decision-making process when actually
purchasing a product those pieces of information were affected. The effect of the
central and peripheral route information on word of mouth is unknown since neither
study includes the word-of-mouth communication in the process of making a decision on product purchase. However, these are studies that provide hints in exploring
decision-making processes that affect word of mouth since they indicate the difference in the roles of the central and peripheral routes, such as how the information
route that is affected in product recognition and attitude formation differs.
The study by Tam et al. (2005) is a study that measured the effect of personalized
websites on each stage of decision making. Their study selected the level of product
suitability to consumer’s preference as a stimulus for the central route and the sorting
cues and the number of recommended products as stimuli for the peripheral route;
classified consumers’ decision-making process into three stages of attention, elaboration, and acceptance of the personalized offer; and conducted an experiment on the
5.2 Direction of the ELM Application
79
selection of ringtones for mobile phones. Based on this, it became clear that while
the level of product suitability—a stimulus for the central route—affects the stages
of elaboration and acceptance of the personalized offer and the number of recommended products—a stimulus for the peripheral route—it does not affect the stage of
accepting the personalized offer even though it affects the stage of consumer’s attention and elaboration. In other words, while keywords related to the peripheral route
do not affect the ultimate choice although they are effective in terms of attention to
the elaboration of information, keywords related to the central route are effective in
the stages of elaborating information and accepting the personalization even though
they are not effective in drawing attention. Therefore, the difference in the roles
played by the central route and the peripheral route was demonstrated.
The study by Colicev et al. (2018) defined corporate news posted on media such
as Facebook and Twitter as “owned media” and brand discussions on SNS among
consumers as “earned media.” Then, they explored 273 days of online conversations on 45 brands in order to see which stage of the customer’s journey for ultimately choosing a product is affected by the results of interpreting the media contents
centrally and peripherally. Based on this, it was shown that while simple peripheral
route information such as the number of posts is used at the stage of directing attention
to a product, the central route information is used for elaboration when purchasing
a product during the customer journey. It was further demonstrated that consumers
use information both on the central and peripheral routes, compare own experience
with others’ or the performance mentioned by the manufacturer, and determine the
level of their satisfaction during the stage of consumer satisfaction. In other words,
it indicates that both the central and peripheral information are important for satisfaction with the product, which affects word of mouth since the information on the
central route and the information on the peripheral route have different influences on
each stage of decision making.
Based on this study by Colicev et al., we can see that information on the central
route alone does not necessarily drive satisfaction. In fact, this is also mentioned
in studies on satisfaction. For example, according to the latest research by Sivadas
and Rupinder (2017), satisfaction with restaurants consists of two parts: the food
they serve (essential part) and their service (incidental part). While the food served
contributes more to satisfaction than the service, customers do not spread the word
after enjoying the food unless they are satisfied not only with the food but also with
the service.6 It turns out that the information on the incidental part affects word of
mouth in addition to the information on the essential part.
Emotional connection is also important for word of mouth after satisfaction.
For example, Ladhari (2007) showed the effect of emotions on satisfaction and the
subsequent word-of-mouth communication.7 He divided emotions into pleasure and
arousal and measured their effect by using films as subjects. Based on this, pleasure and arousal affected all—including satisfaction, positive reviews, and intention
to spread the word. This statement is also consistent with the study by Yamamoto
and Katahira (2008) presented in Chap. 2. They showed that when enthusiasm, a
type of emotion, is generated in the middle of decision making, it affects the intention to recommend the product.8 It is also consistent with the study by Inoue (2009),
80
5 Mechanism of Attitude Formation for Consumers Who Convey Information
mentioned in Chap. 4 on commitment. The study showed that the intention to recommend to others is affected by affective commitment as well as fascination commitment, which is a stronger emotion.9 In other words, some emotion obtained at the
time of purchase affects satisfaction and word of mouth.
Based on the discussions so far, the study will use real data and verify the relationship between attitude formation and word-of-mouth communication according
to the ELM, as well as the relationship between the word-of-mouth communication
and affective commitment.
5.3 Explanation of the Analyzed Data
As mentioned above, the purpose of the analysis here is to show that word-of-mouth
communication occurs when a product is evaluated via the central route of the ELM
and that affective commitment is also involved in word-of-mouth communication.
Here, I have decided to address these questions by dividing the sample population
into those who judge product information by using the central route and those who
judge by using the peripheral route, and exploring the difference in factors such
as affective commitment and the intention to spread the word by those two sample
groups. In what follows, I will briefly describe the data and the flow of the analysis.
The analysis looked at three health food categories. In addition to the usage
of the given product category, the survey questions asked about the usual attitude
toward health—knowledge, information collection, and information communication regarding the given product category as well as the media exposure at each
decision-making process. The sample size is 764 for Product Category X, 744 for
Product Category Y, and 859 for Product Category Z. Due to the nature of the product
categories, 60% of all respondents are in their 40 and 50 s for all products.
First, I will explain the motivation and ability for distinguishing the central route
from the peripheral route of the ELM.
Variables for motivation to elaborate were created by modifying questions on the
level of usual interest in health since the subject products are health foods. The level
of interest in health, which is an item all respondents were asked regardless of the
product category, consists of seven dummy variables. Specifically, they are “I try to
maintain regular hours,” “I try to eat well-balanced meals,” “I try to get enough sleep
for my health,” “I exercise moderately for my health,” “I try to take health foods and
supplements on a daily basis,” “I make sure to manage my health,” and “I check my
state of health such as weight, body fat, and blood pressure on a daily basis.” In this
study, I prepared “health interest score” by summing these seven statements (ranging
from 1 to 7 points: m = 2.11, S.D. = 1.941). Table 5.1 shows the distribution of the
score. Based on how scores are distributed, I classified those who scored a 0 point
as “low level of health interest” (26.7% of the total), those who scored 1 to 3 points
as “middle level of health interest” (50%), and those who scored 4 points or higher
as “high level of health interest” (23.3%).
5.3 Explanation of the Analyzed Data
Table 5.1 Score of health
interest
Score
81
Frequency
%
Cummulative %
0
472
26.67
26.67
1
360
20.34
47.01
2
273
15.42
62.43
3
252
14.24
76.67
4
181
10.23
86.89
5
100
5.65
92.54
6
82
4.63
97.18
7
50
2.82
100.00
1770
100.00
Total
Next, the ability to elaborate was determined based on whether the respondent
had the ability and the habit to gather information on the product category. In
the case of health foods studied in this survey, whether the respondent regularly
collected information on the product category was used as an alternate parameter
for the ability since those who maintained their knowledge daily were extremely
rare. There are five questions related to the ability, comprising “it is fine to put time
and effort to collect product information,” “I check on new or hot products,” “I use
various sources to collect information,” “word-of-mouth information is useful in
considering a product,” and “I collect opinions and information from experts,” which
are collected based on a five-point scale rating. Here, I summed the five statements
to assign a score ranging from 5 to 25 and determined the ability to elaborate—from
strong (16 points or higher) to moderate (11 to 15 points) to weak (5 to 10 points)
based on the distribution of the score and opinions of product managers at the
pharmaceutical manufacturers. The mean and standard deviation for each product
category are: m = 18.05, S.D. = 4.88 for Category X, m = 8.69, S.D. = 5.14 for
Category Y, and m = 16.93, S.D. = 5.49 for Category Z, respectively.
After determining the level of health interest and the ability to elaborate, I took
a general view of the combination of three levels of motivations and abilities using
crosstab and determined whether the decision-making process for each sample was
the central route, peripheral route, or neither by product category. Specifically, those
deemed to use the central route are the ones who fall under the three cells of “strong
motivation × strong ability,” “strong motivation × moderate ability,” and “moderate
motivation × strong ability”; those deemed to use the peripheral route are the ones
who fall under the three cells of “weak motivation × weak ability,” “moderate motivation × weak ability,” and “weak motivation × moderate ability”; and those deemed
neither are the ones who fall under the remaining three cells. As a result, the number
of individuals on the central route and the number of individuals on the peripheral
route were 351 and 297 in Product Category X, 289 and 213 in Product Category
Y, and 231 and 266 in Product Category Z, respectively, and the remaining samples
were deemed neither.
82
5 Mechanism of Attitude Formation for Consumers Who Convey Information
Next, I will explain the statements on communicating information and affective
commitment.
As for communicating information, five out of ten statements on collecting and
communicating information, namely “I have friends to talk with,” “I provide information to online communities,” “I provide stories to online communities,” “sometimes
I give advice when I’m asked to,” “when there is a product I think is good, I recommend it to others,” and “I would like to exchange information” were rated on the
five-point scale and simply summed with the perfect score being 25. Then, using that
score, the level of information communication was classified into three groups: “don’t
communicate at all” (5 to 10 points), “somewhat communicate” (11 to 15 points),
and “actively communicate” (16 points or higher). The mean and standard deviation
for each product category are: m = 13.17, S.D., = 5.34 for Category X, m = 14.00,
S.D. = 5.70 for Category Y, and m = 13.27, S.D. = 5.68 for Category Z, respectively.
The items for affective commitment were measured by each brand. Considering
the literature review and the analysis results in Chap. 4, five statements were set
up. Specifically, they are “I am attached to this product,” “this product meets my
preferences,” “I want to purchase this product even if it’s expensive,” “this product
will not disappoint me,” and “this is an indispensable product for me”; each statement
was measured on a five-point scale rating.
Furthermore, in order to investigate what kind of media the consumers of the
central route are exposed to and what message they focused on at that time, items
related to those were studied as well. With regard to media, I looked at search engines
and mobile sites in addition to mass media such as TV and newspaper. As for message,
I looked at items related to the central and peripheral routes such as the actual product,
price, and sales & marketing campaign. I instructed the respondents to answer each
dummy variable with 0 or 1.
5.4 Analysis Results
First, I prepared a crosstab between the difference in decision-making route and the
level of information communication, and examined whether the consumers on the
central route communicate information at a different level compared to those on the
peripheral route. Table 5.2 shows the results for health food Category Z.
Here, the Chi-square test showed that consumers who make decisions based on
the central route are more actively communicating information than those who make
decisions via the peripheral route (χ2 (4) = 105.502, p < 0.001). In other words, it
was demonstrated that consumers who are motivated to purchase health foods, are
capable of discerning the difference, and make up their mind via the central route are
more likely to spread the word after purchasing a product compared to the consumers
on the peripheral route who are not capable of judging health foods, indicating the
merit for regarding the ELM as the base of a comprehensive decision-making process
model for generating word of mouth.
5.4 Analysis Results
83
Table 5.2 Crosstab between decision-making routes and the level of information communication
Level of information communications
Decision
making route
Don’t
communicate
information
Rarely
communicate
information
Communicate
information
Total
Neither
101
158
103
362
Peripheral route
147
85
34
266
35
93
103
231
283
336
240
859
Central route
Total
Next, I measured whether the level of affective commitment differs based on the
difference in the central and peripheral decision-making routes. First, I selected those
who stated they purchase the top brand in each of the three health food categories:
X, Y, and Z (denoted as x1, y1, and z1, respectively). I did so because affective
commitment can only be tabulated per brand and the number of top brand purchasers
was large enough to allow tabulating by separating the central route from the peripheral route. The number of survey subjects and the breakdown of the purchasers by
the central route versus peripheral route are x1 = 235 (central route = 108, peripheral route = 127), y1 = 226 (central route = 112, peripheral route = 114), and
z1 = 114 (central route = 66, peripheral route = 48). The mean scores of affective commitment factors among these subjects were tabulated by brand as well as
decision-making route, compared by product category, and tested for the difference
in means in order to determine whether there was indeed a difference in means. The
mean scores of affective commitment factors of these 3 brands are shown in Figs. 5.2,
5.3 and 5.4.
It became evident from here that under Brand x1, those who made a decision
by following the central route scored significantly higher on all items composing
affective commitment compared to those who made a decision by following the
peripheral route: I am attached to this product,” t(232.598) = 3.289, p < 0.01, “this
product meets my preferences,” t(233) = 3.972, p < 0.01, “I want to purchase this
product even if it’s expensive,” t(233) = 3.432, p < 0.01, “this product will not
disappoint me,” t(233) = 3.682 p < 0.01 and “this is an indispensable product for
me” t(233) = 3.062, p < 0.001.
With Brand y1, although there was no significant difference between the individuals on the central route and the individuals on the peripheral route in terms of “I
am attached to this product”(t (224) = 1.665, p < n.s), and “this product will not
disappoint me,” (t(213.566) = 0.892, p < n.s.), there was a difference in other three
statements: “this product meets my preferences,” t(224) = 3.745, p < 0.001, “I want
to purchase this product even if it’s expensive,” t(224) = 2.813,p < 0.01,“this is an
indispensable product for me” t(224) = 2.413, p < 0.05).
84
5 Mechanism of Attitude Formation for Consumers Who Convey Information
Fig. 5.2 Affective committment and Central/Peripheral route (brand x1)
Fig. 5.3 Affective committment and Central/Peripheral route (brand y1)
5.4 Analysis Results
85
Fig. 5.4 Affective committment and Central/Peripheral route (brand z1)
With Brand z1, while the item “I am attached to this product” was 10% level
of significant difference (t(112) = 1.96, p < 0.10), there was a 5% or 1% level of
significant difference in other four factors:” this product meets my preferences,”
t(112) = 3.04, p < 0.01, “I want to purchase this product even if it’s expensive,”
t(112) = 2.278, p < 0.05, “this product will not disappoint me,” t(90.417) = 2.946
p < 0.01 and “this is an indispensable product for me” t(112) = 3.614, p < 0.01.
We can conclude that affective commitment after a purchase is higher when
the decision is made via the central route since there is a difference between the
consumers who made a decision via the central route and the consumers who made
a decision via the peripheral route under the three statements “this product meets
my preferences,” “this product will not disappoint me,” and “this is an indispensable
product for me” for all brands, even though there are slight differences by product
category.
These analyses demonstrated the relationship between the mechanism of how
word-of-mouth is generated, the decision-making process, and affective commitment. While previous studies have shown that consumers with strong affective
commitment with a given product are likely to spread the word, the analyses showed
that it was due to two phenomena: the fact that those who made a decision via the
central route had a stronger affective commitment than those who made a decision
via the peripheral route and the fact that those who made a decision via the central
route spread the word more actively than those who made a decision via the peripheral route. In other words, we can conclude that the relationship between the level of
affective commitment and word of mouth is created through the mediation of making
a decision via the central route.
As described, we found that decision making on the central route not only generates word of mouth but also helps in developing affective commitment. In addition,
as is evident in Chap. 4, it is extremely useful for a company to have consumers
make decisions on the central route because brands supported by consumers with
86
5 Mechanism of Attitude Formation for Consumers Who Convey Information
strong affective commitment becomes long-selling brands. Getting many people to
spread positive information would also contribute to the sale of the product since
positive word-of-mouth information helps in increasing the intention to purchase
among those who have not tried before, as shown in the case of cosmetics reviews
in Chap. 3.
Next, I analyzed whether there is a difference between the consumers who make a
decision on the central route and the consumers who make a decision on the peripheral
route in terms of the information source used and the information required at the time
of decision making by using Brand x1 which had a sample population large enough
for the analysis and showed a difference in all of five affective commitment factors.
The decision-making process was explored by dividing the process into five stages:
normal time, at the time of comparison, at the time of purchase consideration, at the
time of purchase decision, and after the purchase.
First, the percentage of individuals who use at least one information source at
each decision-making stage was tabulated by separating the central route from the
peripheral route. Figure 5.5 shows the results. The percentage of information source
users is naturally higher among the consumers who follow the central route than
the consumers on the peripheral route under the normal exposure to media because
“whether the respondent has the ability and the habit to collect information on the
given product category” was used as a proxy variable for the ability to elaborate by
definition in the first place. However, we can see that the percentage of information
source users is lower among those on the peripheral route than those on the central
route even in other stages. What is especially notable is the post-purchase behavior;
while 70% of those on the central route are still searching for some information after
the purchase, the percentage is only 30% among those on the peripheral route. It
100.0
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
Central route
10.0
0.0
Normally
At the me of At the me of At the me of Aer purchase
comparison
purchase
decision
Fig. 5.5 % of those who use at least one information source
Peripheral route
5.4 Analysis Results
87
seems that consumers on the peripheral route are not very interested in the product
after the purchase, including reviews.
Next, I tabulated the information sources that were actually used in order to
see what percentage of the consumers on the central route as well as the ones on the
peripheral route use them at each decision-making stage, which is shown in Table 5.3.
Regardless of using the central route or the peripheral route, consumers are usually
highly exposed to information on TV and at the storefront and the importance of
information at the storefront increases as they proceed to compare, consider purchase,
and actually purchase. The rate of information source usage is higher among the
consumers on the central route than the consumers on the peripheral route at almost
all stages; the difference is particularly notable with the information sources such as
magazines, newspapers, manufacturer’s website, social media, information obtained
from friends and acquaintances, and information obtained from family, which are
sources one proactively uses to collect information. This is consistent with the notion
“individuals who gathered information with recognition and interest are more likely
to become influencers” mentioned in Chap. 2. We can see that while consumers on
the central route are proactively gathering information, consumers on the peripheral
route are usually not so interested in information but focus on gathering information
at the time of purchase. It was found that people who actively gathered information
through various media on a regular basis were more likely to make decisions through
a central route than those who gathered information through a purchasing venue.
Figures 5.6 and 5.7 show the contents of the information consumers actually
require by route and stage. Comparing the two figures, we can see that consumers on
the central route are exposed to various contents throughout the entire decisionmaking process than consumers on the peripheral route. Looking closely at the
contents they are exposed to, consumers on the central route are seeing information on product advertisements and product details on a regular basis; the exposure
to product advertisements declines and it is replaced by the product price and store
clerk’s recommendations as they move to the comparison stage and purchase consideration stage. Approximately 20% of them are still looking at information on product
usage even after the purchase, indicating that their interest in the product remains
strong even after the purchase. In contrast, with consumers on the peripheral route,
the exposure to the product price increases as they enter the comparison stage and
purchase consideration stage; although this is not so different from the consumers
on the central route, the difference is that they do not look at product information on
a regular basis and they do not get as many recommendations from store clerks as
the consumers on the central route do. Only 10% or less are looking at information
after the purchase. As described, the simple tabulations showed that although the
consumers on the central route and the consumers on the peripheral route are similar
in terms of the media they use at each stage, the level of exposure is different—in
particular, there is a difference in the level of exposure to the media they proactively
collect information from—and the information required for decision making at each
stage is different. It was demonstrated that the consumers on the central route are
different from those on the peripheral route in terms of the content of information in
addition to the total amount of information.
37.23
36.80
21.65
17.75
19.91
51.08
32.03
30.74
28.14
18.18
29.00
Newspaper inserts
Manufacturer’s website
Websites other than manufacturer’s
Social media
Actual products seen at the store
Promotional materials seen at the store
Information obtained from store clerk
at the store
Information obtained from friends and
acquaintances
Information obtained from experts
Information obtained from family
36.36
Magazines
Newspapers
70.13
29.44
Radio
56.02
12.78
5.26
9.77
10.90
12.03
24.44
7.89
6.02
7.89
14.29
11.65
12.03
13.16
23.81
14.29
18.18
40.26
23.81
52.38
15.58
16.45
28.57
23.38
12.99
12.99
9.52
27.71
12.03
5.64
10.53
20.68
13.16
35.34
3.76
4.51
10.53
12.41
5.26
5.26
4.89
16.92
Peripheral
Comparison
Peripheral Central
Usually
Central
TV
Media
16.88
12.99
13.42
48.48
24.24
59.31
10.39
8.23
12.99
15.58
6.06
4.76
3.46
16.88
Central
7.14
5.26
4.51
32.33
19.55
49.62
1.88
0.75
3.76
5.64
2.26
1.88
1.50
15.04
Peripheral
Considering
purchase
Table 5.3 Rate of media usage by decision-making stage (central route versus peripheral route)
16.02
10.39
9.52
44.59
15.15
48.92
5.19
4.33
6.49
6.06
3.03
2.16
1.30
9.52
Central
6.77
4.51
3.01
27.44
13.16
39.10
1.13
1.13
2.63
2.26
1.13
0.75
1.50
6.77
Peripheral
Deciding purchase
15.58
5.63
8.66
12.99
6.06
21.21
6.93
6.93
14.29
5.63
3.46
3.90
2.60
10.39
Central
5.64
3.38
3.01
2.26
3.01
8.65
1.13
0.38
4.51
1.13
1.50
1.50
1.13
4.14
Peripheral
After purchase
88
5 Mechanism of Attitude Formation for Consumers Who Convey Information
5.4 Analysis Results
89
Informaon on the product itself
Informaon on product usage
Informaon on product adversement
Recommendaon by store clerk
Product price
Product (brand) image
60
50
40
30
20
10
0
Normally
At the me of
comparison
Consideraon at the
me of purchase
At the me of
decision
Aer purchase
Fig. 5.6 Information required for the central route by stage
In order to statistically explain the difference between consumers on the central
route and those on the peripheral route, which became evident from these simple
tabulations, I used CHAID (CHi-squared Automatic Interaction Detector) analysis
and explored the difference between consumers on these two routes in terms of
the media and information they used. Often used when the independent variables
are categorical variables, the CHAID analysis is a method that extracts variables in
the order to being most influential in classifying the dependent variable until there
is ultimately no more variable that affects the classification (Shimizu 2010).10 It is
suitable for the hypothesis-discovery type of analysis to identify variables with higher
priority. Here, I will look at the consumers on the central route and the consumers
on the peripheral route and explore (1) media with a large difference in the usage
rate and (2) information with a large difference in the usage rate. The results of the
analysis that used product category X are shown in Figs. 5.8 and 5.9, respectively.
First, based on Fig. 5.8, “whether they usually read magazine articles” becomes the
first aspect that separates the central route and the peripheral route in terms of media
usage; the subsequent point of separation is “whether they usually look at products at
the storefront.” In the case of consumers who usually read magazine articles and look
at products at the storefront, 82.8% of them are on the central route. Conversely, with
90
5 Mechanism of Attitude Formation for Consumers Who Convey Information
60
Informaon on the product itself
Informaon on product usage
Informaon on product adversement
Recommendaon by store clerk
Product price
Product (brand) image
50
40
30
20
10
0
Normally
At the me of
comparison
Consideraon at the
me of purchase
At the me of decision
Aer purchase
Fig. 5.7 Information required for the peripheral route by stage
consumers who usually do not read magazine articles, usually do not look at products
at the storefront, and usually do not watch TV programs, 76.3% are on the peripheral
route. Even if consumers do not usually read magazine articles, if they usually look
at products at the storefront and obtain information from store clerks at the time
of comparative consideration, 70% of them are on the central route. As described,
while there are several patterns of consumers on the central route, consumers who
normally pay attention at the storefront and proactively collect information (read
magazine articles or obtain information from store clerks) get on the central route.
The discriminatory power of the central route versus peripheral route is 68.6% based
on the figure.
Next, Fig. 5.9 shows the analysis on the types of information that differed during
the normal time and at the time of comparison. Around 75.3% of consumers who
routinely pay attention to the prices of products get on the central route. If consumers
refer to reviews and user ratings on the product at the time of comparison even if they
do not normally pay attention to the price, 70.2% end up on the central route. On the
other hand, 64.1% of those who normally do not care about the prices of products and
do not refer to reviews or user ratings get on the peripheral route. The figure further
increases to 69.5% when they are not normally exposed to information on the product
description. We can say that the consumers on the central route are characterized
by routinely comparing product prices and using reviews instead of information
Fig. 5.8 Media with a large difference in the usage rate
5.4 Analysis Results
91
5 Mechanism of Attitude Formation for Consumers Who Convey Information
Fig. 5.9 Information with a large difference in the usage rate
92
5.4 Analysis Results
93
communicated by the manufacturer when they compare. The discriminatory power
of the central route versus the peripheral route is 67.4%.
To summarize the above analysis results, two things became evident: those who
made the decision on the central route had higher affective commitment than those
who made the decision on the peripheral route, making them proactively spread the
word. Furthermore, it became clear that the consumers on the central route and those
on the peripheral route differ not only in the amount of information they are exposed
to but also the source and content of information they refer and that the consumers
on the central route require the kind of information and contents they cannot obtain
unless proactively collected (Fig. 5.9).
5.5 Conclusion and Future Implication
While the conventional comprehensive decision-making models for consumers have
aimed at primarily understanding the process up to purchasing and mentioned very
little about spreading the word after the purchase, studies on advertising communication have shown the effectiveness of word-of-mouth communication after the
purchase. Deeming it insufficient to understand only the process up to the purchase
and considering it necessary to include the process up to word-of-mouth communication in consumers’ comprehensive decision-making model in this SNS era, this study
picked the ELM, which focused on the elaboration of advertising messages among
the past decision-making models, and applied it to the product purchase behavior.
After conducting a survey on the health food category and analyzing it, it became
evident that consumers who form their attitude by following the central route are
highly likely to spread the word, and the attitude formed by following the central
route strengthens affective commitment. While previous studies have suggested that
individuals with strong affective commitment spread the word, it was shown here
that such a relationship exists because both word-of-mouth and the state of affective
commitment are phenomena that occur when consumers form their attitude via the
central route.
In response to these analysis results, I used Brand x1, which had a large sample size
and showed a large difference in affective commitment between the central route and
peripheral route, and conducted simple tabulations and CHAID analysis to confirm
the difference between the consumers on the central route and those on the peripheral
route in terms of media usage and required information. It was found that: consumers
on the central route get more information than those on the peripheral route; the usage
rate of the media that requires proactive collection of information was also higher
compared to the consumers on the peripheral route; consumers also routinely look
at information on product description; and consumers look at information even after
the purchase.
94
5 Mechanism of Attitude Formation for Consumers Who Convey Information
Although the importance of word of mouth has been suggested in many studies,
there has been no study that has indicated the mechanism of generating word of
mouth in relation with consumers’ decision-making process. Linking it with the
ELM this time, the mechanism in which consumers with interest and concern are
likely to spread the word as a result of making the decision on the central route was
demonstrated.
Notes
1.
2.
3.
4.
5.
6.
7.
8.
9.
Petty, Richard E.; Cacioppo, John T. (1983), ‘Central and peripheral routes to
Advertising Effectiveness: The Moderating Role of Involvement’, Journal of
Consumer Research, Vol. 10, pp 135–146. and Petty, Richard E.; Cacioppo,
John T. (1986), ‘Communication and Persuasion: Central and Peripheral
Routes to Attitude Change”, Springer.
Kitchen, Phillip J.; Kerr, Gayle; Schultz, Don E.; McColl, Rod; Pals, Heather
(2014) ‘The elaboration likelihood model: review, critique and research
agenda’, European Journal of Marketing, Vol. 48 No. 11/12, pp. 2033–2050.
Yi-Hsiu Cheng; Hui-Yi Ho (2015), ‘Social Influence’s Impact on Reader
Perceptions of Online Reviews’, Journal of Business Research, Vol. 68,
pp. 883–887.
Seth Ketron (2017), ‘Investigating the Effect of Quality of Grammar and
Mechanics (QGAM) in Online Reviews: The Mediating Role of Reviewer
Credibility’, Journal of Business Research, vol. 81, pp 51–59.
Tam, Kar Yan; Ho Shuk Ying (2005), ‘Web Personalization as a Persuasion Strategy: An Elaboration Likelihood Model Perspective’, Information
Systems Research, Vol. 16, No. 3, September, pp. 271–291 and Colicef,
Anatoli; Malshe. Ashwin; Pauwels, Koen; O’Connor, Peter (2018), ‘Improving
Consumer Mindset Metrics and Shareholder Value Through Social Media: The
Different Roles of Owned and Earned Media’, Journal of Marketing, Vol. 85,
pp. 37–56.
Sivadas, Eugene; Rupinder Paul Findal (2017), ‘Alternative Measures of
Satisfaction and Word of Mouth’, Journal of Service Marketing, Vol. 31/2,
pp. 119–130.
Riadh Ladhari (2007), ‘The Effect of Consumption Emotions on Satisfaction and Word-of-Mouth Communication’, Psychology & Marketing, Vol. 24,
pp. 1085–1108.
Yamamoto, Hikara; Katahira, Hotaka (2008), ‘To find the Influencer and the
Effect of WOM- Approach of AIDEES model’, Japan Marketing Journal, Vol.
109, pp. 4–18 (written in Japanese).
Inoue, Atsuko. (2009), ‘The Relationship of Brand Commitment and
Consumer Behavior’, Journal of Marketing & Distribution, Vol. 12, No. 2.
pp. 3–22 (written in Japanese).
5.5 Conclusion and Future Implication
10.
95
Shimizu, Akira (2010), ‘Econometric Analysis of Consumer Buying
Behavior’, Chiohiko Minotani and Atsushi Maki, ed. Handbook of Applied
Econometrics, Asakura Publishing, Co., Ltd. (written in Japanese).
Chapter 6
Emergence of Communication-Oriented
Consumers
The digital divide among Japanese consumers is growing as shown in Chap. 1.
Chapter 2 logically and empirically confirmed that the most up-to-date consumers
spread the word about their experiences with a product or service. Chapter 3 demonstrated that they are especially likely to spread the word when they recognized
and took interest in a product or service, collected information, and then became
satisfied with what they purchased. All the information they spread, in turn, affects
further purchases. In addition to satisfaction, commitment to the brand is important
in information sharing. Chapter 4 demonstrated that the percentage of very affectively committed consumers is particularly high among those who favor long-selling
brands. Chapter 5 showed how an attitude is formed among consumers who spread
the word on a product or service, showing that those who search for information with
recognition and interest are likely to make decisions using the central route, and such
consumers are the ones who are very affectively committed and are most likely to
spread the word.
These research findings tell us that companies must understand good customers
from an angle which is different from what they’ve used in the past if they are
to manage customers by taking into consideration their own brand power, and the
effect of word-of-mouth publicity initiated by their customers. In other words, in
addition to managing customers by focusing on those who spend more, it will become
important for companies to segment customers based on how savvy they are with
information, which improves the effect of word-of-mouth publicity, and the level
of affective commitment which is related to the brand power. Doing so will help
companies properly manage customers who scored high on these items. The idea
of managing customers as if they are assets has been used widely since the study
Customer Equity by Blattberg et al. However, past customer management studies
have focused on maximizing the monetary lifetime value generated per customer.
There are not many studies that discuss the effect of information sharing by customers
with depth. Yet, considering the research findings presented in all the earlier chapters,
those who share information are extremely promising customers in terms of building
a long-selling brand and recommending the brand, and are worth managing. In what
© Springer Nature Singapore Pte Ltd. 2021
A. Shimizu, New Consumer Behavior Theories from Japan, Advances in Japanese
Business and Economics 27, https://doi.org/10.1007/978-981-16-1127-8_6
97
98
6 Emergence of Communication-Oriented Consumers
follows, I will first review the theories on customer management. Then, naming
consumers who have strong affective commitment and share information proactively
as “communication-oriented consumers,” I will make a case for the effectiveness of
managing these consumers based on real data.
6.1 The Idea of CRM and Its Development
There are studies on store loyalty, which is a theory of retailing in the old days
as a method of considering the relationship between a company and customers,
and managing customers. However, customer relationship management (CRM) has
fast evolved as a concept that is attracting attention in recent years. CRM aims to
maximize revenues generated from a long-term, continuous business relationship
with customers, rather than merely aiming at developing good relationships with
customers to satisfy them. It became popular in the late 1990s (Minami 2004).1 In
theory, CRM is a concept that encompasses traditional research areas related to retail
and customer, such as customer retention, database marketing, and loyalty marketing,
covering even customer satisfaction, relationships, and service marketing in a broader
sense (Minami 2008, Yamamoto 2008).2
There are several reasons why the concept of CRM became widely popular.
One is because maintaining and developing existing customers is far more cost
effective than acquiring new customers is. We know from the data that good customers
particularly make significant contributions to the bottom-line. According to Reichheld (1996), for example, the cost of acquiring a new customer is five to ten times
the cost of satisfying and maintaining existing customers. Therefore, he states that
it is efficient to strengthen the relationship with existing customers.3 Woolf (1996)
showed that the good customers in the top 20% spend 15 times what the customers in
the bottom 20% spend in the case of food retailers. He also showed that the attrition
rate after one year is 4% among the good customers in the top 20%, while it reaches
64% among the customers in the bottom 20%, and argued that it is not efficient to
treat customers equally in the food retail industry, where the profit margin is not
large.4 Based on a series of studies on the distribution industry in Japan, Nakamura
showed that since the good customers in the top 10% account for more than 40%
of sales in any product category, and up to about 85% of sales can be accounted
for when the range of good customers is expanded to include the top 30%, sending
customers in the bottom 30% to other stores would increase profitability (Nakamura
1998, Nakamura 2006). He also indicated that good customers are highly profitable
because they purchase many related items.5
The second reason is the fact that the environment in which hardware for storing
customer’s purchase data became available. The data on purchase history of good
customers is enormous. For example, the frequent shoppers program data used by
supermarkets in order to acquire good customers leaves one record of history per
purchased item. Since good customers in Japan frequent a supermarket at least thrice
a week and purchase 15 items per shopping expedition, it will amount to 45 records
6.1 The Idea of CRM and Its Development
99
a week, exceeding 2,000 records a year. If there are 10,000 good customers, the
annual purchase history data reach 20 million records, which are further appended
to information such as price, quantity, sales promotion, and competitors at the time
of purchase. For this reason, computer resources capable of storing a large amount
of data have become necessary. However, the fact the prices of hard disk for data
storage dropped dramatically, right around the time when the concept of CRM was
introduced, boosted the implementation of CRM.
The third reason is that the analytical technique called data mining was introduced to analyze large data, and the CRM procedure to incorporate data mining was
established. Data mining refers to a group of analytical methods, including cluster
analysis and CHAID (CHi-squared Automatic Interaction Detector analysis) that
identify meaningful relationships based on a large set of data. It is a group of analytical methods that you employ to discover hypotheses rather than to verify them. Since
data mining methods were introduced, it became possible to extract a good customer
segment ranked high in terms of spending, to identify the difference between good
customers and the remaining segments or to show that there are several patterns
within the segment of good customers, for example, in order to establish efficient
CRM execution procedures (Shimizu 2004).6
CRM uses concepts such as customer equity and lifetime value as success
variables, in order to represent the long-term relationship with customers.
First, customer equity, which is a concept advocated by Blattberg, Getz, and
Jacquelyn (2001), regards relationships with customers as an asset, and tries to
manage the relationships to optimize sales.7 Specifically, customer equity refers to
the management of customer lifecycle, quantification of customer value, acquisition
and retention of customers, and the optimization of the product mix for cross-selling
in order to retain customers and optimally balance locking customers in. Supposedly,
it would ensure a stable and long-term customer base, which will become an asset
to the company.
In contrast, lifetime value is a more concrete goal for success. The idea is to understand the value (lifetime value) contributed by one customer over his/her lifetime, and
adjust the level of management based on that. Specifically, the method that calculates
the current gross profit per customer is often used. For this reason, the lifetime value
of bargain-oriented customers would be lower, even if their spending is large. New
methods of lifetime value calculation are also being introduced in recent years. For
example, Borle, Siddharth, and Jain (2008) proposed a method to apply customer
attrition to the extended negative binominal distribution model to improve accuracy.8 In any event, the aim is to increase efficiency by giving preferential treatment
to customers who have a higher lifetime value than others.
As described, existing good customers could become corporate assets and
contribute in terms of cost and profit. Furthermore, the environment to practice
this and the methods to evaluate the results have been established. In response,
many companies have created loyalty program cards, introduced CRM to organize
customers, and produced results in Japan, as well.9 However, it is not that all companies are successful in CRM. According to Kotler, about 30% of the total number of
100
6 Emergence of Communication-Oriented Consumers
companies, have succeeded in CRM (Kotler 2003).10 In fact, several studies point
out cautions while practicing CRM.
For example, Liu and Yan (2009) confirmed that the effect of loyalty programs
differ by industry growth and company size, based on a study of the financial statements of airlines and customer surveys.11 According to this study, while the difference
in the details of loyalty programs does not become a big issue when the entire market
in the industry is growing, it makes a difference when the market growth is in stasis. It
also showed that the effects of loyalty programs are higher among larger companies
than smaller companies. In other words, the study indicates that the effects vary by
company size and competitive situation.
Sakai (2006) used the data on hair salon customers and verified that some
customers voluntarily establish relationships with the company, while others do
not.12 It showed that while the overall satisfaction is not high among customers
who have no other option but to use a given salon because their loyalty is based
on constraints, even though they have been using that salon, the loyalty based on
constraints decreases, and the intention to return to the salon increases when the
overall satisfaction increases. This study suggests that it is risky to calculate lifetime
value only on the basis of a customer using a particular salon because his/her intention to return is not strong if the customer is not satisfied even if he/she continues
to return. In other words, this study indicates that there is a limitation on measuring
customer value based on purchase history alone, and that CRM will not work unless
the customer’s feelings are taken into consideration.
Even though earlier CRM literature might make you think you can succeed by
simply processing the purchase history data and ranking consumers, in reality, it is
clear that you wouldn’t succeed unless you understand factors such as the position of
your company in the industry and the state of true loyalty of customers. That being
said, since it is definitely a very effective means to prevail in fierce competition, there
are numerous discussions regarding the direction and development of CRM research
in the future.
For example, Bouldin, Staelin, Ehret, and Johnston (2005) summarized the areas
of CRM research that should be worked on in the future into six groups, and
argued particularly for the importance in exploring the causal relationships during the
process of consumers building true loyalties for the company, as well as for the need
for broader success matrices.13 Rust and Chung (2006) summarized the potentials for
future development into 14 issues from the perspective of relationships and discussed
subjects, such as the security problems that would arise from the relationships, easier
access to customers due to technological innovations, and the potential in customer
management that considers the relationships among themselves.14 Baohong (2006)
showed seven directions in which CRM should evolve, from the perspective of technological progress. He also specifically discussed subjects such as understanding
customers more accurately based on new data, addressing new channels, and the
potential of automation.15
What is particularly notable among these notions are the facts that the emergence
of big data, such as information on web browsing history and comments on SNS
linked to each customer, has enabled us to explore deeper about customers. The
6.1 The Idea of CRM and Its Development
101
development of networks among individuals in addition to the company’s network
with individuals, has also led to the possibility of managing things even to that extent.
The emergence of big data is also drawing attention in the field of marketing in
general. For example, deeming the emergence of big data as the second innovative
factor after the emergence of POS data in performing data analysis on consumers,
Chintagunta, Hanssens, and Hauser (2016) presented various possibilities in data
analysis.16 As examples, they listed certain facts such as (1) one can understand the
profile of a customer, such as customer attributes, more accurately by looking at the
customers’ web browsing history, (2) one can understand related purchases from a
consumer’s point of view now, (3) one can match purchase history and recommendations, and (4) one can understand how a company is rated based on the remarks
left by consumers on SNS. It is clear that they are all closely related to CRM and
will affect the development of CRM.
With regard to the network between individuals, it hasn’t quite been realized
yet, even though the need has been mentioned. For example, having demonstrated
that lifetime value is composed of four factors, namely, increased consumption per
consumer increases revenue, managing existing customers allows the reduction of the
cost of acquiring new customers, the added value can be shown by premium price, and
favorable reviews by existing customers increase the number of new customers, Lee,
Lee, and Feick (2006) makes a case that the last point, that is the effect of reviews,
is rarely discussed.17 Ascarza, Ebbes, Netzer, and Matthew(2017) also states that
there is almost no literature or practice that studies CRM by taking social effects
into consideration, even though mutual social relationships have been considered in
explaining many marketing phenomena.18 However, the study by Lee et al. indicated
that consumers with a strong propensity to spread the word are useful in acquiring
new customers because their attrition rate is low, at around 4%, regardless of the
size of their spending. Meanwhile, the study by Ascarza et al. showed that the usage
rate increased by 35% among target customers in response to a campaign, and also
increased by 10% among customers who are connected via a network to those target
customers. These days, where good customers can easily connect with each other,
thanks to the popularization of the smartphone and SNS, CRM that takes even this
kind of personal relationship into account will become important now, more than
ever.
As described, the way CRM is implemented is changing, with the evolution of
the data environment and SNS. In this chapter, I will discuss the effects of networks
among customers, specifically, the effects of word-of-mouth by good customers,
among other things. In all the chapters so far, we have seen that consumers spread
the good word when they are satisfied with the product or service, or when they have
formed an emotional connection with the product or service, like while forming an
affective commitment. In what follows, I will name consumers who are good wordof-mouth “communication-oriented consumers” and show the potential of CRM by
focusing on that point.
102
6 Emergence of Communication-Oriented Consumers
6.2 What Is a “Communication-Oriented Consumer”?
Based on the discussions in the earlier chapters, it was made clear that consumers
must have an emotional connection with the product they choose, before spreading
the word on it. That emotional connection is measured as affective commitment.
If it is strong, information is highly likely to be spread. In other words, affective
commitment is a key to positive information sharing. In addition, the review on CRMrelated literatures above has showed that the influence capitalizing on the relationship
among customers, or the strong ability of the customer to communicate with other
potential customers to spread word on the product or service, has become important
in CRM in the recent years. Therefore, in this chapter, I will attempt to establish a new
angle for customer management by setting the aggressiveness toward information,
that is, “the propensity to collect and spread information” and “the level of affective
commitment” as two axes, and creating four consumer segments based on the level of
value on each axis. The name of each segment is as shown in Fig. 6.1: those who are
more affectively committed and aggressive in collecting and spreading information
are “communication-oriented consumers,” those who are more affectively committed
but don’t collect or spread information are “habitual consumers,” those who are less
affectively committed but collect and spread information are “comparison-oriented
consumers,” and those who are less affectively committed and don’t collect or spread
information are “unplanned consumers.”
After establishing the conceptual definition this way, products that are favored
more by communication-oriented consumers than by others were identified and
the relationship between communication-oriented consumers and brand rating was
explained.19 Specifically, the affective commitment axis was prepared by using scores
on five statements including, “I am attached to this brand,” “this brand meets my
preferences,” “I want to purchase this brand even if it’s more expensive than other
brands,” “this brand will not disappoint me,” and “this is an indispensable brand for
me.” For the information axis, a synthetic variable created based on five statements
on information collection (“it’s OK to spend time and effort to collect information
in order to purchase a product you can be satisfied with,” “I check products that are
Fig. 6.1 Consumer
segmentation based on the
affective commitment and
information
High
habitual
consumers
communicationoriented customers
Unplanned
consumers
comparison-oriented
consumers
The level of
Affective
Commitment
Low
Weak
the propensity to collect
and spread information
Strong
6.2 What Is a “Communication-Oriented Consumer”?
103
new or are talked about a lot,” “I use various sources while collecting information,”
“word-of-mouth information is useful while examining a product for purchase,” and
“I try to test the product before I purchase it”) and five statements on information
sharing, (“I have someone or a group of people to talk to about this product category, either in everyday conversation or on in an online community,” “I bring up this
product category in the conversations I have, or in online communities,” “sometimes
I give advice on this product category to my acquaintances when they ask me for it,”
“when there is a product that I think is good, I recommend it to others,” and “I enjoy
exchanging information on products”). The number of samples is 900.
First, I looked at the relationship between these consumers and the product categories. Figure 6.2 plots 15 products based on the percentage of user composition.20
As shown earlier, the first quadrant is the communication-oriented type, the second
is the habitual type, the third is the unplanned type, and the fourth is the comparisonoriented type. Taking a general view of where products are segmented makes it clear
how consumers are currently seeing information on those products, and with what
kind of attitude.
We can see from the figure that the products with a high percentage of
communication-oriented consumers are cosmetics, alcoholic beverages, and home
appliances. As was shown in Fig. 2.1 in Chap. 2, cosmetics is a product category
in which consumers willingly sought ambiguous information even before the emergence of the internet, so that word-of-mouth information has been essential. We can
say this remains the same after the internet has spread wide. People tend to habitually purchase alcoholic beverages without collecting information because these
are non-essential grocery items. However, the fact that wine, among other alcoholic beverages, is widely sold and reviewed on the internet, might be a reason
why the percentage of communication-oriented consumers is high for wine. In the
case of home appliances, they are typical products for which consumers make an
information-processing type of decision because the purchase interval is long and the
price is high. They are also increasingly sold on the internet in recent years. It seems
we got this result because price comparison sites and review sites on the internet are
used often in recent years for this reason.
Fig. 6.2 Products
segmentation based on the
affective commitment and
information
High
The level of
Affecve
Commitment
Low
PCs
automobiles
clothing
cosmecs
alcoholic beverages
home appliances
video soware
travel
books
snacks
processed foods
so drinks
health foods
interior goods
gardening products
Weak
the propensity to collect
and spread informaon
Strong
104
6 Emergence of Communication-Oriented Consumers
Products strongly favored by habitual consumers included PCs, automobiles, and
clothing. They can be regarded as a group of products for which the percentage of
people who purchase, based on their level of commitment to the manufacturer, than
aggressiveness in acquiring information, is high. Specifically, it is a product category
that people are extremely likely to purchase more of, or, a repeatedly of.
Product categories with a large percentage of unplanned consumers include video
software, travel, and books, where content is important. It seems that they settled into
this position because of the nature of the content business (numerous manufacturers,
wide variety of products, product appeal via various media, etc.). Then again, as
seen in Chap. 4, since young people account for a large percentage of sales and
there were many study cases in which SNS played a large role with products that
can be purchased or downloaded online, there is a possibility that the percentage
of communication-oriented consumers will increase as sales channels shift to the
internet.
Product categories such as snacks, processed foods, soft drinks, health foods,
interior goods, and horticulture and gardening products are positioned under products
with a large percentage of comparison-oriented consumers. It is characterized by the
fact that products with different natures, such as health foods and interior goods with
strong product involvement, as well as frequently-discounted snacks, process foods,
and soft drinks, are included. The former is considered a group of products requiring
information in order to compare factors like ingredients and effects, because there is a
large risk if it doesn’t work. The latter is considered a group of products that requires
information in order to compare prices because there is little difference among the
products themselves.
The above categories of surveyed products were classified based on the percentages of consumers who favored them. These classifications reflect how products
are sold and purchased; therefore, we can conclude that the classification based
on the axes of aggressiveness toward information and affective commitment has
explanatory power. In what follows, I will use real data and consider CRM that uses
communication-oriented consumers.
6.3 Characteristics of Communication-Oriented
Consumers
Here, I will show an analysis of beer and beer-like beverages21 and demonstrate
that communication-oriented consumers are useful while considering CRM. The
samples include 2,769 males and females aged between 20 and 60. They were classified into four segments according to the above definition. As a result, 416 fell
under the communication-oriented type, 905 under the habitual type, 1,288 under
the unplanned type, and 106 under the comparison-oriented type.
First, consumers with high affective commitment have the same characteristics
as those shown in Chap. 4. In Chap. 4, it was shown that consumers with high
6.3 Characteristics of Communication-Oriented Consumers
105
Table 6.1 Routinely purchase long-selling brands
Brand A
Affective commitment
Total
High
Low
Frequency
181
94
Expected frequency
133.9
141.1
Not Buy Brand A
Frequency
501
625
1126
Expected frequency
548.1
577.9
1126
Total
Frequency
682
719
1401
Expected frequency
682
719
1401
Buy Brand A
Brand B
Affective commitment
275
275
Total
High
Low
Frequency
183
138
Expected frequency
156.3
164.7
Not Buy Brand B
Frequency
499
581
1080
Expected frequency
525.7
554.3
1080
Total
Frequency
682
719
1401
Expected frequency
682
719
1401
Buy Brand B
321
321
affective commitment favor long-selling products and buy products regardless of
price. Table 6.1 shows whether respondents routinely purchase the two long-selling
brands (Brand A and Brand B, other than the top-selling brands) by dividing them
into two types: those with higher affective commitment scores (communicationoriented type and habitual type) and those with lower scores (unplanned type and
comparison-oriented type). As a result, for brand A, χ2 (1) = 40.230, p < 0.0001,
and for brand B χ2 (1) = 6.758, p < 0.001, and there was a significant difference in
responses. From this, it is evident that the group with higher affective commitment
routinely purchased more long-selling brands than the group with lower emotional
commitment.
Table 6.2 shows the extent to which people usually cite price-related items as
reasons for purchasing beer. While the low affective commitment group was significantly more likely than the high affective commitment group to cite reasonable price
(χ2 (1) = 52.658, p < 0.0001) and discounted prices (χ2 (1) = 68.798, p < 0.0001)
as reasons for purchasing, the high affective commitment group was significantly
more likely than the low affective commitment group to cite buying a product they
liked even if the price was slightly higher (χ2 (1) = 138.204, p < 0.0001). These
two results confirm that, similar to what was shown in Chap. 4, the group with high
affective commitment preferred long-selling products and chose products independent of price compared to the group with low affective commitment, even for beer
products.
Next, I checked the difference in attributes among four segments of consumers by
performing analysis of variance (ANOVA). First, Table 6.3 shows the differences by
age and gender. We can see from the size of Pearson chi-square value that there are
106
6 Emergence of Communication-Oriented Consumers
Table 6.2 Reasons for purchasing beer
Affective commitment
High
Low
125
278
192.3
210.7
1196
1170
1128.7
1237.3
1321
1448
1321
1448
Discounted prices
Yes
Frequency
Expected frequency
Frequency
Expected frequency
Frequency
Expected frequency
No
Total
Affective commitment
High
Low
366
620
470.4
515.6
955
828
850.6
932.4
1321
1448
1321
1448
Reasonable price
Yes
No
Total
Frequency
Expected frequency
Frequency
Expected frequency
Frequency
Expected frequency
Buying a product they liked even
if the price was slightly higher
Yes
Frequency
Expected frequency
No
Frequency
Expected frequency
Total
Frequency
Expected frequency
Affective commitment
High
Low
537
292
395.5
433.5
784
1156
925.5
1014.5
1321
1448
1321
1448
Total
403
403
2366
2366
2769
2769
Total
986
986
1783
1783
2769
2769
Total
829
829
1940
1940
2769
2769
Table 6.3 The difference in attributes among four segments of consumers
㻯㼛㼙㼙㼡㼚㼕㼏㼍㼠㼕㼛㼚㻙㼛㼞㼕㼑㼚㼠㼑㼐㻌㼠㼥㼜㼑
㻴㼍㼎㼕㼠㼡㼍㼘㻌㼠㼥㼜㼑
㼁㼚㼜㼘㼍㼚㼚㼑㼐㻌㼠㼥㼜㼑
㻯㼛㼙㼜㼍㼞㼕㼟㼛㼚㻙㼛㼞㼕㼑㼚㼠㼑㼐㻌㼠㼥㼜㼑
㼀㼛㼠㼍㼘
㻲㼞㼑㼝㼡㼑㼚㼏㼥
㻱㼤㼜㼑㼏㼠㼑㼐㻌㼒㼞㼑㼝㼡㼑㼚㼏㼥
㻲㼞㼑㼝㼡㼑㼚㼏㼥
㻱㼤㼜㼑㼏㼠㼑㼐㻌㼒㼞㼑㼝㼡㼑㼚㼏㼥
㻲㼞㼑㼝㼡㼑㼚㼏㼥
㻱㼤㼜㼑㼏㼠㼑㼐㻌㼒㼞㼑㼝㼡㼑㼚㼏㼥
㻲㼞㼑㼝㼡㼑㼚㼏㼥
㻱㼤㼜㼑㼏㼠㼑㼐㻌㼒㼞㼑㼝㼡㼑㼚㼏㼥
㻲㼞㼑㼝㼡㼑㼚㼏㼥
㻱㼤㼜㼑㼏㼠㼑㼐㻌㼒㼞㼑㼝㼡㼑㼚㼏㼥
㻹㼍㼘㼑
㻞㻜㻙㻞㻥
㻡㻥
㻠㻝㻚㻢
㻤㻥
㻥㻜㻚㻡
㻝㻜㻡
㻝㻞㻤㻚㻤
㻞㻠
㻝㻢㻚㻜
㻞㻣㻣
㻞㻣㻣
㻹㼍㼘㼑
㻟㻜㻙㻟㻥
㻤㻠
㻡㻟㻚㻟
㻥㻡
㻝㻝㻢㻚㻜
㻝㻡㻣
㻝㻢
㻝㻥
㻞㻜㻚㻡
㻟㻡㻡
㻟㻡㻡
㻹㼍㼘㼑
㻠㻜㻙㻠㻥
㻠㻤
㻢㻥㻚㻝
㻝㻣㻥
㻝㻡㻜㻚㻟
㻞㻝㻝
㻞㻝㻠㻚㻜
㻞㻞
㻞㻢㻚㻢
㻠㻢㻜
㻠㻢㻜
㻹㼍㼘㼑
㻡㻜㻙㻡㻥
㻟㻞
㻢㻡㻚㻣
㻝㻡㻟
㻝㻠㻞㻚㻤
㻞㻠㻞
㻞㻜㻟㻚㻟
㻝㻜
㻞㻡㻚㻟
㻠㻟㻣
㻠㻟㻣
㻲㼑㼙㼍㼘㼑
㻞㻜㻙㻞㻥
㻠㻢
㻟㻝㻚㻠
㻢㻠
㻢㻤㻚㻟
㻣㻢
㻥㻣㻚㻞
㻞㻟
㻝㻞㻚㻝
㻞㻜㻥
㻞㻜㻥
㻲㼑㼙㼍㼘㼑
㻟㻜㻙㻟㻥
㻢㻜
㻠㻣㻚㻜
㻥㻤
㻝㻜㻞㻚㻟
㻝㻟㻞
㻝㻠㻡㻚㻢
㻞㻟
㻝㻤㻚㻝
㻟㻝㻟
㻟㻝㻟
㻲㼑㼙㼍㼘㼑
㻠㻜㻙㻠㻥
㻠㻣
㻡㻟㻚㻥
㻝㻝㻞
㻝㻝㻣㻚㻟
㻝㻣㻢
㻝㻢㻣㻚㻜
㻞㻠
㻞㻜㻚㻣
㻟㻡㻥
㻟㻡㻥
㻲㼑㼙㼍㼘㼑
㻡㻜㻙㻡㻥
㻠㻜
㻡㻟㻚㻥
㻝㻝㻡
㻝㻝㻣㻚㻟
㻝㻤㻥
㻝㻢㻣㻚㻜
㻝㻡
㻞㻜㻚㻣
㻟㻡㻥
㻟㻡㻥
㼀㼛㼠㼍㼘
㻠㻝㻢
㻠㻝㻢
㻥㻜㻡
㻥㻜㻡
㻝㻞㻤㻤
㻝㻞㻤㻤
㻝㻢㻜
㻝㻢㻜
㻞㻣㻢㻥
㻞㻣㻢㻥
6.3 Characteristics of Communication-Oriented Consumers
107
Table 6.4 The difference in drinking habits among four segments of consumers
㻯㼛㼙㼙㼡㼚㼕㼏㼍㼠㼕㼛㼚㻙㼛㼞㼕㼑㼚㼠㼑㼐㻌㼠㼥㼜㼑
㻴㼍㼎㼕㼠㼡㼍㼘㻌㼠㼥㼜㼑
㼁㼚㼜㼘㼍㼚㼚㼑㼐㻌㼠㼥㼜㼑
㻯㼛㼙㼜㼍㼞㼕㼟㼛㼚㻙㼛㼞㼕㼑㼚㼠㼑㼐㻌㼠㼥㼜㼑
㼀㼛㼠㼍㼘
㻲㼞㼑㼝㼡㼑㼚㼏㼥
㻱㼤㼜㼑㼏㼠㼑㼐㻌㼒㼞㼑㼝㼡㼑㼚㼏㼥
㻲㼞㼑㼝㼡㼑㼚㼏㼥
㻱㼤㼜㼑㼏㼠㼑㼐㻌㼒㼞㼑㼝㼡㼑㼚㼏㼥
㻲㼞㼑㼝㼡㼑㼚㼏㼥
㻱㼤㼜㼑㼏㼠㼑㼐㻌㼒㼞㼑㼝㼡㼑㼚㼏㼥
㻲㼞㼑㼝㼡㼑㼚㼏㼥
㻱㼤㼜㼑㼏㼠㼑㼐㻌㼒㼞㼑㼝㼡㼑㼚㼏㼥
㻲㼞㼑㼝㼡㼑㼚㼏㼥
㻱㼤㼜㼑㼏㼠㼑㼐㻌㼒㼞㼑㼝㼡㼑㼚㼏㼥
㻭㼘㼙㼛㼟㼠
㼑㼢㼑㼞㼥㼐㼍㼥
㻝㻤㻟
㻝㻟㻢㻚㻥
㻟㻝㻟
㻞㻥㻟㻚㻣
㻟㻟㻣
㻠㻝㻡㻚㻡
㻢㻟
㻠㻥㻚㻥
㻤㻥㻢
㻤㻥㻢
㻠㻌㼛㼞㻌㻡㻌㼐㼍㼥㼟
㼜㼑㼞㻌㼣㼑㼑㼗
㻢㻣
㻡㻤㻚㻡
㻝㻟㻥
㻝㻞㻡㻚㻡
㻝㻢㻟
㻝㻣㻣㻚㻢
㻝㻠
㻞㻝㻚㻟
㻟㻤㻟
㻟㻤㻟
㻞㻌㼛㼞㻌㻟㻌㼐㼍㼥㼟 㻻㼚㼏㼑㻌㼍
㼜㼑㼞㻌㼣㼑㼑㼗
㼣㼑㼑㼗
㻥㻠
㻞㻞
㻝㻜㻝㻚㻡
㻠㻡㻚㻝
㻞㻝㻜
㻣㻤
㻞㻝㻣㻚㻢
㻥㻢㻚㻣
㻟㻞㻝
㻝㻤㻝
㻟㻜㻣㻚㻥
㻝㻟㻢㻚㻤
㻟㻥
㻝㻠
㻟㻣㻚㻜
㻝㻢㻚㻠
㻢㻢㻠
㻞㻥㻡
㻢㻢㻠
㻞㻥㻡
㻞㻌㼛㼞㻌㻟㻌㼜㼑㼞
㼙㼛㼚㼠㼔
㻣
㻟㻝㻚㻜
㻢㻜
㻢㻢㻚㻡
㻝㻟㻜
㻥㻠㻚㻝
㻢
㻝㻝㻚㻟
㻞㻜㻟
㻞㻜㻟
㼀㼛㼠㼍㼘
㻟㻣㻟
㻟㻣㻟㻚㻜
㻤㻜㻜
㻤㻜㻜㻚㻜
㻝㻝㻟㻞
㻝㻝㻟㻞㻚㻜
㻝㻟㻢
㻝㻟㻢㻚㻜
㻞㻠㻠㻝
㻞㻠㻠㻝
significant differences among these four segments (χ2 (21) = 123.162, p < 0.0001).
First of all, the communication-oriented type has a large percentage of those aged
under 40 among both, males and females. On the other hand, the habitual type had
a higher percentage of those aged above 40. It is assumed that this result is due to
the fact that younger people are more active in disseminating information. As for the
comparison-oriented type, the percentage peaks for those in their 20 s and declines
for those in their 50 s. There is no particular characteristic for the unplanned type.
Table 6.4 looks at the relationship between drinking frequency and four segments.
We also see from the size of Pearson chi-square here that there are significant differences in drinking frequency among the four segments (χ2 (12) = 108.387, p <
0.0001). Specifically the drinking frequency is very high among communicationoriented types and habitual types. So the significance of targeting these two groups
of consumers for CRM was demonstrated.
Existing CRM theories reveal that both types, communication-oriented type and
habitual type, are appropriate targets, but we next identified how communicationoriented type and habitual type people differ in their information sharing. As shown
in Chap. 3 while addressing the impact of blogs, the kind of blog content that had
the highest impact on search was the blogger’s actual usage experience. Therefore,
I checked to see if the subjects talked about their impressions or ratings during in
person or online conversations with friends or online conversations with someone
other than their friends, and also whether they submitted comments to media after
they actually drank the product or not. The results are shown in Table 6.5. With
regarding to information sharing, habitual type people are significantly less likely to
disseminate information than communication-oriented type people (“I don’t send out
information.” χ2 (1) = 153.841, p < 0.0001). Communication-oriented type people
participated significantly more than habitual type people in all of the following ways:
sharing information in real life with friends (“I bring it up in conversation with
friends.” χ2 (1) = 65.910, p < 0.0001), sharing information online (“I talk about it
in an online community with friends.” χ2 (1) = 40.863, p < 0.0001), and sharing
information with non-friends online (“I talk about it in an online community with
non-friends.” χ2 (1) = 27.471, p < 0.0001). This reveals that affective commitment
as well as the propensity to collect and spread information will be an important axis
in considering CRM in the future.
108
6 Emergence of Communication-Oriented Consumers
Table 6.5 Difference between communication-oriented type and Habitual type in their information
sharing
I don’t send out information
Yes
Communication-oriented type
Habitual type
Total
Frequency
82
509
591
Expected frequency
186.1
404.9
591
No
Frequency
334
396
730
Expected frequency
229.9
500.1
Total
Frequency
416
905
1321
Expected frequency
416
905
1321
730
I bring it up in conversation with Communication-oriented type Habitual type Total
friends
Yes
No
Frequency
259
345
604
Expected frequency
190.2
413.8
604
Frequency
157
560
717
Expected frequency
225.8
491.2
717
416
905
Total Frequency
Expected frequency
416
905
1321
1321
I talk about it in an online
community with friends
Communication-oriented type
Habitual type
Yes
Frequency
50
28
Expected frequency
24.6
53.4
78
No
Frequency
366
877
1243
Expected frequency
391.4
851.6
1243
Total
Total
78
Frequency
416
905
1321
Expected frequency
416
905
1321
I talk about it in an online
community with non-friends
Communication-oriented type
Habitual type
Yes
Frequency
57
48
Expected frequency
33.1
71.9
105
No
Frequency
359
857
1216
Expected frequency
382.9
833.1
1216
Total
Total
105
Frequency
416
905
1321
Expected frequency
416
905
1321
Table 6.5 Difference between communication-oriented type and Habitual type in
their information sharing
So what are the brands that communication-oriented type consumers, who also
have high emotional commitment and scores on the propensity to collect and spread
information, prefer in beer products? Figure 6.3 shows the difference in beers and
beer-like alcoholic beverages favored by each segment based on the percentage
favored by the type of consumers, just as Fig. 6.2. Looking at the segmented brands,
6.3 Characteristics of Communication-Oriented Consumers
Fig. 6.3 Segmentation of
Beer brand
High
The level of
Affective
Commitment
Low
109
Beer A
Beer H
LMB B
LMB P
LMB S
BLA J
BLA K
BLA N
Weak
LMB E
LMB R
BLA D
BLA Q
BLA L
Premium beer C
Premium beer F
Beer B Beer I
Beer M
LMB O
LMB G
LMB T
the propensity to collect
and spread information
Strong
we can see that premium beers, beers are at upper right, low-malt beverages (LMB)
are at lower right and lower left, and the beer-like alcoholic beverages (BLA) are
positioned at lower left, respectively. It turns out to be linked with the way the price
range of beers and beer-like beverages falls in that order.
Among these, premium beers are very popular among communication-oriented
types. Beers A and H with a longer history and higher price than other beers, are
placed where the percentage of habitual consumers is large. Among low-malt beers,
Brand T scored high on the information sharing axis. Since this brand is touted
to cut down calories, it was placed where it is, probably because this point was
favored by consumers who are highly likely to share information. Of all beers and
beer-like alcoholic beverages, beer-like alcoholic beverages are popular among many
unplanned types. In particular, when I went to check the sales, Brand Q which was
rated low, had actually been discontinued shortly after the study. Based on this, we
can see that there are differences in the segment favoring beers, low-malt beers, or
the beer-like alcoholic beverages, even though they all fall under the beer category,
and that those differences can be explained by two axes of affective commitment and
information collection/sharing.
As described above, communication-oriented consumers are highly capable of
spreading the word. They make decisions by rating items related to the central route,
which turned out to be consistent with the conclusion on the profile of consumers who
spread the word as shown in Chap. 5. It was also shown that habitual consumers tend
to favor long-selling brands, and unplanned types were price-oriented. Therefore,
while practicing CRM, it becomes important to target communication-oriented types
if you want to maintain the freshness of the brand and target habitual types if you
want to create a long-selling brand.
110
6 Emergence of Communication-Oriented Consumers
6.4 Purpose of Managing Communication-Oriented
Consumers
Next, focusing on beers among beers and beer-like alcoholic beverages, I employed
the theory of consideration set and examined which decision-making route each
segment, that is, communication-oriented types, habitual types, and unplanned types,
follows to select one brand. “Consideration set” refers to a group of brands that
consumers considered at the time of purchase. It is thought that consumers form a
consideration set while collecting information based on their objective and ultimately
choose one brand from it. Understanding how the number of brands and the group
of brands’ consumers consider change in phases over the course of decision-making
will allow us to measure the brand power, and evaluate the role of media affecting
the process (Shimizu 2006).22 While there are various opinions on how the process
of consideration set should be viewed, it is defined as “recognize → understand the
detail → purchase within three months → form a consideration set → drink one
product most frequently → purchase only that product” to understand the flow to
purchase and for loyalty. The surveyed subjects are 1,401 respondents who answered
all these steps out of the entire sample population. A total of seven brands of beer
were included in the survey, two premium beer brands and five regular beer brands.
Of these, one premium beer (brand F) was newly launched, while brand C of the
premium beer and brand B of the beer were long-selling brands. Note that the number
of beer purchases by consumers belonging to the comparison-oriented type was small,
so the comparison was made for the three types, excluding the comparison-oriented
type.
First, Fig. 6.4 shows the shifts for all beers and beer-like alcoholic beverages
(including the comparison-oriented type). We can see from here that Beer A is overwhelmingly strong as “purchase within three months” while Beer B, Premium C,
Premium F and Beer H form the second group and Beers I, and M form the third
group. The results show that Brand A is the strongest brand (Brand A is no. 1 share
brand of beer market) and Brand I and Brand M are the weaker brands, but the
strength of the remaining four brands, which are in the second group, cannot be
determined.
Figure 6.5 shows these shifts with communication-oriented types. It is characterized by the fact that they are roughly divided into two groups, one consisting of
four brands of Beers A and B and Premiums C and F, and the other consisting of
the remaining, at two stages of “purchase within three months” and “form a consideration set.” Furthermore, the beers they favor are the ones with a highest market
share brand, 2 long-selling brands, and one new premium beer. It is characteristic
that they choose not only the beers with a higher market share and long-selling but
also the beer that is gaining market share; we can say they are capable of discerning
the brands being talked about.
Meanwhile, Fig. 6.6 shows the shifts among habitual types. Here, what is characteristic is that Beer A is stronger than other beers at all stages; in particular, it is
largely ahead of other beers at the stages of “drink most frequently” and “purchase
6.4 Purpose of Managing Communication-Oriented Consumers
111
1600
1400
1200
1000
800
600
400
200
0
recognize
understand the purchase within
form a
detail
three months consideraon
set
Premium F
Premium C
Beer A
Beer I
Beer M
Beer H
drink most
frequently
purchase only
that product
Beer B
Fig. 6.4 Decision making route of beer (all sample)
200
180
160
140
120
100
80
60
40
20
0
recognize
understand the purchase
form a
detail
within three consideraon
months
set
Premium F
Premium C
Beer A
Beer I
Beer M
Beer H
drink most
frequently
Beer B
Fig. 6.5 Decision making route of beer (communication-oriented type)
purchase only
that product
112
6 Emergence of Communication-Oriented Consumers
600
500
400
300
200
100
0
recognize
understand the purchase within
form a
detail
three months consideraon
set
Premium F
Premium C
Beer A
Beer I
Beer M
Beer H
drink most
frequently
purchase only
that product
Beer B
Fig. 6.6 Decision making route of beer (habitual type)
only that product” Fig. 6.7 shows the shifts among the unplanned types. While the
overall shifts are similar to that of habitual types, they rate Premiums C and F lower
compared to communication-oriented types and habitual types. As shown in Fig. 6.3,
consumers in unplanned type tend to prefer lower-priced low-malt beverages (LMB)
and the beer-like alcoholic beverages, which may have led to lower ratings of higherpriced premium beers. We can surmise that unplanned consumers don’t value these
two expensive brands because they are very price-conscious.
Based on these results, it became clear that whereas communication-oriented
consumers favor trendy beers with potential and don’t favor other beers although
it’s not like they drink the most popular beer disproportionally more, habitual types
and unplanned types don’t have a clear line to separate beers they like and the beers
they don’t like, and tend to ultimately drink the most popular beer, most frequently
including comparison-oriented type. In other words, it is extremely important to
observe the behavior of communication-oriented types, because the market trend that
cannot be understood just by market share can be understood by carefully observing
the brands favored by them. It became clear that to continue being favored by them
is essential in terms of CRM, as well, since they are highly capable of spreading the
word and very influential among people they are connected to.
6.5 Conclusion and Future Implication
113
700
600
500
400
300
200
100
0
recognize
understand the purchase within
form a
detail
three months consideraon
set
Premium F
Premium C
Beer A
Beer I
Beer M
Beer H
drink most
frequently
purchase only
that product
Beer B
Fig. 6.7 Decision making route of beer (unplaned type)
6.5 Conclusion and Future Implication
Through Chap. 5, I have talked about consumers who share information mainly in
terms of information processing. In this chapter, however, I discussed consumers
who share information as well as marketing strategies in relation to CRM strategy
by focusing particularly on the consumers’ levels of commitment and ability to
collect information and spread the word, and by dividing them into four segments.
While conventional CRM strategies stressed on strengthening the relationship with
consumers who spend more, given how recent studies discuss the importance of
relationships with consumers who spread the word, I created the above segments
accordingly, and, by using real data on beer, confirmed the effectiveness of this
segmentation. Naming the consumers who are particularly capable of collecting and
sharing information and have a strong affective commitment as “communicationoriented consumers,” I mainly explored their behavior.
First, we confirmed that consumers with high affective commitment, as shown
in Chap. 4, buy long-selling products and that they buy products without being
influenced by price. The results indicated that it is significant to capture consumers
with high affective commitment as a target for CRM. Next, we compared the
differences between the two types of consumers with high affective commitment
114
6 Emergence of Communication-Oriented Consumers
(communication-oriented type and c habitual type) in terms of their ability to disseminate information after using the target product. The purpose of the study was to
test the effectiveness of the actual experience of using the product as word of
mouth. The results revealed that communication-oriented type consumers were more
active in disseminating information than habitual type consumers, revealing that
communication-oriented type consumers were more suitable for CRM.
Finally, in order to further explore the beers, and these communication-oriented
consumers value, I took specific beer brands as examples and explored how they
determined which beer to drink at what stage of decision-making. There, whereas
looking at the brands purchased within three months by habitual types and unplanned
types showed they were favored at about the same level as the overall share,
communication-oriented types favored not only the brands with a large market share,
but the brands with momentum. There was a large gap between a group of very
popular brands and a group of unpopular brands. Their judgment of good and bad is
characterized by being extremely clear compared to the judgments by habitual types
and unplanned types, indicating that communication-oriented consumers are very
useful in measuring a brand’s potential that cannot be measured by market share.
Conventional CRM strategies have produced results by focusing on customers
with large spending and assessing product selection based on their purchase characteristics. However, since all retailers incorporated similar strategies as a result,
retailers ended up having similar product selection and not being able to maintain competitive advantage. As the internet developed, some CRM studies began
showing the effectiveness of considering customer’s ability to communicate information. However, it has not been shown exactly what kind of effect it has or what
kind of customers should be targeted. In this analysis, it was demonstrated that
communication-oriented consumers who score high on both of the two axes, the
ability to collect and share information and affective commitment, properly rate
upcoming beers that cannot be measured by market shares, showing that considering
their ability to spread word, they are the ideal targets while practicing CRM. We
can say that it showed the direction of the new CRM that considers the effects of
word-of-mouth publicity.
Notes
1.
2.
3.
Minami, Chieko. (2004), “An approach to CRM based on strategic marketing
perspective: Is CRM a magic wand?” The Journal of Marketing and Distribution, December Vol. 426, pp. 18–25. (written in Japanese).
Minami, Chieko. (2008), “The theoretical underpinnings of relationship
marketing: a service marketing approach,” Journal of Economics and Business
Administration, May Vol. 197, pp. 33–50, (written in Japanese) and Yamamoto,
Shoji. (2008), “Expanding the concept of customer value: From the perspective
of service marketing,” Japan Marketing Journal, Vol. 107, pp. 4–15. (written
in Japanese).
Reichheld, Frederick F. (1996), ‘The Loyalty Effect: The Hidden Force Behind
Growth, Profits, and Lasting Value’, Harvard Business School Press.
6.5 Conclusion and Future Implication
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
115
Brian P. Woolf (1996), Customer Specific Marketing, TEAL BOOKS.
Nakamura, Hiroshi. (1998), “Loyalty Marketing (1),” The Journal of
Marketing and Distribution, April issue, pp. 10–18 (written in Japanese) and
Nakamura, Hiroshi. (2006), “Collaborative CRM between Manufacturers and
Retailers Using ID/POS Data,” Senshu Business Review, Vol. 1, pp. 31–39.
(written in Japanese).
Shimizu, Akira. (2004), Retail Strategy Based on The Client Perspective,
Chikura Publishing Company. (written in Japanese).
Blattberg, Robert C; Getz,Gary; Jacquelyn, S. Thomas (2001), Customer
Equity Building and Managing Relationships as Valuable Assets, Harvard
Business School Publishing Corporation.
Borle,Sharad; Siddharth, Singh S.; Jain, Dipak C. (2008), Customer Lifetime
Value Measurement, Management Science, Vol. 54, No. 1, pp. 100–112.
For example, it is known that Rakuten in Japan increased their e-mail click rate
and conversion rate (the rate which the individual who clicked also purchase)
by introducing the idea of CRM.
Philip Kotler (2003), Marketing Insights from A to Z, John Wiley & Sons.
Liu, Yuping; Yan Rong, (2009), ‘Competing Loyalty Programs: Impact of
Market Saturation, Market Share, and Category Expandability’, Journal of
Marketing, Vol. 73, pp. 93–108.
Sakai, Maiko. (2006), “Customer Oriented Service Relationship Model: The
roles of Switching Barrier and Customer Loyalty in Building Relationship,”
Advances in Consumer Studies, Vol. 13, No. 1, pp. 29–56. (written in Japanese).
Bouldin,William; Staelin, Richard; Ehret, Michael; Johnston,Wesley J. (2005),
‘A Customer Relationship Management Roadmap: What Is Known, Potential
Pitfalls, and Where to Go’, Journal of Marketing, Vol. 69, pp. 155–166.
Rust, Roland T; Chung,Tuck Siong (2006), ‘Marketing Models of Service and
Relationships’, Marketing Science, Vol. 25, pp. 560–580.
Baohong Sun (2006), ‘Technology Innovation and Implications for Customer
Relationship Management’, Marketing Science, Vol. 25, pp. 594–597.
Chintagunta, Pradeep; Hanssens, Dominique M.; Hauser, John R. (2016),
Editorial: Marketing Science and Big Data, Marketing Science, Vol. 35, pp
341–342.
Lee, Johathan; Lee Janghyuk,;Feick Lawrence (2006), ‘Incorporating wordof-mouth effects in estimating customer lifetime value’, Database Marketing
& Customer Strategy Management, Vol. 14, pp. 29–39.
Ascarza, Eva; Ebbes, Peter; Netzer, Oded; Matthew,Danielson (2017),
‘Beyond the Target Customer: Social Effects of Customer Relationship
Management Campaigns’, Journal of Marketing Research, Vol. 54, pp. 347–
363.
The idea of “communication-oriented consumers” is a concept created in a
joint study with Dai Nippon Printing Co., Ltd. The outcome was published as
‘Identify communication-oriented consumers’ (Nikkei Business Publications,
2007).
116
6 Emergence of Communication-Oriented Consumers
20.
15 product categories were plotted based on the popularity among consumers
in which of those quadrants. Since the subjects are those who want to personally
purchase the product, the sample size varies by product category.
There are three types of beer and beer-like alcoholic beverages in Japan. The
one containing 67% or more malt is called “beer,” the one with less than 67%
malt content is called “low-malt beer,” and the one made of ingredients other
than malt is called “the third beer.” Each has a different alcohol tax rate; with
beer taxed at the highest rate and the third beer taxed at the lowest rate, the
prices vary accordingly.
For more information, see Shimizu, Akira. (2006), Strategic Consumer
Behavior, Chikura Publishing Company. (written in Japanese).
21.
22.
Chapter 7
Research on Uncertain Listeners
In Chap. 6, I used two scales to show that the brand segmentation is significant: (1)
consumers’ information sensitivity and (2) consumers’ affective commitment to the
relevant product. The data analysis of beers showed that “freshness” too possesses
vital brand power, particularly for the brands supported by “communication-oriented
consumers” who have a high level of affective commitment toward the relevant
product group and a high degree of information sensitivity. In other words, if you
know the purchaser, you can measure the brand’s strength. However, an extremely
large survey is required to measure consumers’ information sensitivity and affective
commitment to a product. When the number of products to be surveyed increases,
the survey volume increases as well because it then becomes necessary to measure
the affective commitment toward each product.
In this chapter, I address the above problem considering the notion “If you know
the purchasers of this brand, you can measure its brand power.” I aim to construct
general-purpose consumer segments that can keep down this burden of conducting a
survey and measuring the brand power for every product category. Several researchers
have attempted to create general-purpose segments in lifestyle research. In this
chapter, while referring to these research studies, I try to create segments from
easy-to-extract consumer lifestyle characteristics and degrees of preference for the
so-called “brand products.”
“Brand products” are created for targeting certain consumer segments and are
marked by high symbolism. It is assumed that consumers who use the same brand
products basically have the same values and lifestyle. This has been proved in past
research. For example, Escalas and Bettman (2005) stated that “consumer research
on reference groups has demonstrated congruency between segment membership
and brand usage”.1 They showed that it is possible to create differentiated segments
from the usage conditions of consumer brand products, as their values and lifestyles
are different.
In addition, there are different types of brands among the brand products released
into the market: the so-called “seasonal” brands that are a topic of conversation; niche
brands that are supported only by a specific segment or, conversely, brands that are
© Springer Nature Singapore Pte Ltd. 2021
A. Shimizu, New Consumer Behavior Theories from Japan, Advances in Japanese
Business and Economics 27, https://doi.org/10.1007/978-981-16-1127-8_7
117
118
7 Research on Uncertain Listeners
well received by everyone universally without any issue; and brands whose popularity
is already on the decline. People with high information sensitivity prefer seasonal
brands, while people whose information sensitivity is not that high prefer the brands
whose popularity is on the decline. Thus, to a certain extent, it is possible to ascertain a person’s information sensitivity from his/her brand preferences. Consumers’
information sensitivity, which was presented as an important scale in the previous
chapter, can also be measured from their brand preferences.
In this chapter, I will first simply review the past segmentation research that used
the lifestyle theory. I will then create consumer segments using data on consumers’
actual preferences for brand products. Next, I will use the created segments to evaluate
the brands of various product categories. The consumer segment with a high brand
power for support is named “active listeners,” while the consumer segment that does
not have it is named “uncertain listeners.” I will analyze the relationship between the
measurements of brand power and each of these segments.
7.1 Lifestyle Research Is Flourishing
In the history of the consumer behavior theory and in the context of marketing
strategy, one area of consumer behavior research is segmentation, targeting, and positioning (STP) marketing proposed by Kotler. He developed it to explore consumer
behavior on the axis of segmentation, which is the “S” in STP. Within this development process, lifestyle research has played a major role. There are two methods
of segmentation: a priori segmentation, which involves dividing the market into
multiple markets using variables estimated in advance, and cluster segmentation,
which involves mathematically processing the variables associated with the market
divisions and creating segments that cannot be discovered in advance (Wind 1978).2
In lifestyle research, creating segments through the latter method is particularly
regarded as an effective theory. Lazer introduced it in the second half of the 1960s
in combination with motivation research and personality research. To support this
research, a method was established for creating segments by preparing a scale for
subjective awareness using a Likert scale and, subsequently, by conducting a factor
analysis and a cluster analysis (Lazer 1963).3
In lifestyle research, it is easier to discover unique new segments when compared
to a priori segmentation using demographic factors. Moreover, their explanatory
power is high, and they have been used in a huge number of cases as the axis on which
the segmentation is conducted. Wells (1975), in his review, stated that, in the event of
conducting segmentation for a specific product, compared to other variables, lifestyle
variables elucidate by aggregating many scales; as a result, their power of discernment
is extremely high and they can be used to create new segments.4 Wind (1978) noted
that within the papers he reviewed on the problems with and development of segmentation, in the case of conducting cluster segmentation, lifestyle factors have become
a powerful weapon, and they are particularly important when determining an advertising strategy.5 Beane and Ennis (1987), who reviewed the segmentation research
7.1 Lifestyle Research Is Flourishing
119
conducted up to the start of the 1980 s, figured out five variables that should be emphasized when conducting segmentation: 1. geographic factors, 2. demographic factors,
3. lifestyle, 4. purchase history, and 5. image-combining lifestyle and behavior. They
found that using lifestyle factors is highly valuable in the cases where it is not
possible to successfully conduct segmentation using demographic factors.6 Wedel
and Kamakura (2000) summarized the four reasons that make lifestyle factors popular
for conducting segmentation: (i) The quality of the motivational insights obtained is
useful in understanding the underlying reasons for observed consumer behavior; (ii)
The generality of the consumer profiles obtained through psychographic segmentation makes it applicable to a wide range of products and services; (iii) The conceptual
framework is flexible enough such that measurement instruments can be tailored
to specific domains of application; (iv) The consumer profiles obtained are implementable as they provide guidance for new product development and the execution
of advertising messages.7
In this way, segmentation through lifestyle factors has long been used as an effective method for creating unique segments. However, since its introduction, there
have been various discussions on the reliability of this scale. As can be judged from
the fact that the phrase ‘tailor-made’ appeared in the previously mentioned paper
by Wind, in general, when determining lifestyle factors for each product, the question items are often created subjectively by the analyzer. Consequently, when the
analyzer or the era changes, the items become different even when surveying the
same lifestyles, which renders a comparison impossible. It also has a disadvantage
when attempting to secure the accuracy of answers and increase the consistency with
previous research. New questions will be added in response to the change of era so
that the number of question items tends to become extremely large.8
Doubts have also been raised about segmentation through lifestyle factors from
the viewpoint of the segmentation-creation method. There have been various discussions on the criteria for determining the pros and cons of a segmentation method. For
example, Kotler (1967) cites six criteria: 1. the homogeneity of the people belonging
to the segment; 2. segment’s size and interests; 3. stability of the segment; 4. accessibility of the segment; 5. integration of the segment and corporate objectives; and
6. possibility of action by companies for the segment.9 DeSarbo and Grisaffe (1988)
propose five segmentation criteria: 1. clear differences between the segments; 2. the
robustness of the created segments; 3. the accessibility of the segment; 4. the ease
of creating the segment; and 5. the feasibility of implementing measures that use the
segment.10 On checking segmentation through lifestyle factors against these segmentation criteria, I noticed it is difficult to say that lifestyle factors are appropriate to use
in comparison to segmentation that uses objective indicators such as demographic
factors and purchase histories. There are several reasons behind this, such as the
analyzer’s tendency to subjectively create question items; the stability, robustness,
and accessibility of the extracted segments; and the feasibility of implementing them.
To compensate for these weak points, another method is frequently used to secure
validity: creation of the question items in accordance with the activities, interests,
opinions (AIO) criteria (Wells and Tigert 1971).11 This technique increases their
versatility for general application on the questions belonging to the three AIO areas
120
7 Research on Uncertain Listeners
when surveying lifestyle factors. In-depth interviews were carried out on the members
belonging to the segments created in this way, which led Lastovicka and others to
develop the DD method (Lastovicka 1982, Lastovica, Murry, Joachimsthaler 1990)12
and other methods to confirm the effectiveness of the segments. In addition, when
requiring segments with uniqueness, the question items on lifestyle factors are based
on the AIO criteria number in the hundreds, which places a large burden on the
respondent. Therefore, Kamakura and Wedel (1995) used the potential-class method
to create a system to derive stable segments even when the number of questions
is reduced.13 However, even when such techniques are used, unique segments are
still discovered because there are many subjective question items. It means that when
making the criteria more severe and reducing the number of questions, the uniqueness
of lifestyles disappears and the value of the method also decreases.
Integrated lifestyle factors were developed for deriving unique segments while
removing the subjectivity from the scale. In the discussions so far, lifestyle research
dealt with individual products, such as cars or clothes. However, integrated lifestyle
factors show the consumers’ life as a whole. Typical examples of this approach
are Values & Lifestyle (VALS), which was developed mainly in Stanford University, VALS2, and List of Values (LOV), which was developed at the University of
Michigan. Lifestyle research on individual products is called individual lifestyle
research.
VALS was introduced in 1980. It was developed based on Maslow’s hierarchy
of five needs. The survey contains a total of 800 questions on values, lifestyle, and
consumption behavior. The survey was conducted among 1,648 Americans to create
a total of nine lifestyle segments. VALS2 was proposed in 1989 for reducing the large
number of questions in VALS, which was considered a problem. VALS2 comprises
42 questions andeight lifestyle segments. LOV was created based on the values of the
main roles in life and Maslow’s theory. Specifically, a list of nine phrases (two items
on the individual’s inner orientation, three items on interest in the outside world, and
four items on the individual’s values) are evaluated in each of the nine stages. The
segments are determined from the resulting scores.
A feature of these integrated lifestyle approaches is that they use integrated question items rather than individual lifestyle items decided subjectively by researchers
according to the individual product. They also seek to explain not a specific product
segment, but lifestyle in general. The questions are about lifestyle in general, which
reduces the fluctuations in scores from question items about differences between
product segments, and the questions asked in the survey are fixed. This has the
advantage of securing the survey’s robustness. The survey results can be compared
too, regardless of who takes the survey and when it is taken. VALS, VALS2, and
LOV are widely used in practice throughout the world, especially in Japan where
a Japanese version of VALS has been created. The prediction accuracies of the
respective methods differ depending on the conditions. Past research has shown that
the prediction accuracy of VALS is higher if demographic factors are not considered. As LOV offers more questions relating to image, it is superior for measuring
consumption trends.14
7.1 Lifestyle Research Is Flourishing
121
However, it has been pointed out that there are problems with the integrated
methods, as the questions are integrated. A main problem is that when compared to
the individual lifestyle method tailored according to product, there are many abstract
questions. Another problem is that because there are many question items in which
the influence of the home appears, inevitably the impact of demographic factors and
the sense of social hierarchy become stronger. Then, there is the aspect of it having
nuances different from the conventional individual lifestyle method.
In fact, Steenkamp and Hofstede (2002) who reviewed the literature on international comparisons, showed that even if there are question segments that are
considered to secure robustness, these question items reflect the conditions in the
United States, which is where they were created; so they are hardly useful at all
for international comparisons.15 Wedel and Kamakura (2000), who applied actual
data to compare several integrated lifestyle scales, including VALS and LOV, also
concluded that although they have been shown to be effective, “two critical issues
must be considered in applying those measurement instruments to the identification of lifestyle segments: stability of the lifestyle typology and its generalizability.”
They also noted that “future research in this area should further investigate ways
to establish a theoretical link between values and behavior.” While confirming that
an integrated-type lifestyle analysis has potential as a segmentation method, they
concluded that it needs to theoretically show a relationship with real consumer
behavior as a requirement.16
To summarize, segmentation through an individual lifestyle method is effective for
extracting unique segments, but its versatility for general applicability is extremely
low because it has a large number of questions and includes many that ask for
the respondent’s subjective opinion. Therefore, for increasing versatility, integrated
lifestyle methods have been researched. However, in this method, the uniqueness of
the individual lifestyle method is lost. Therefore, there is the need for a method that
has both uniqueness and versatility and that theoretically shows the relationship with
consumer behavior.
7.2 Relationship Between Lifestyles and Brands
A majority of existing research suggests a relationship between the so-called “brand
products” and consumer lifestyles. The reason is that a brand is often used as a
method of self-expression. Past research has shown that consumers tend to like and
possess these brands, as the brand has a symbolic image and one’s self-concept is
enhanced by having that brand (Puzakova, Kwak, and Rocereto 2009).17 According
to Chernev, Hamilton and Gal (2011), who reviewed the relationship between brands
and lifestyles, the main research into brands has focused on the following four categories: 1. research into brands that identify the self, 2. research into reference groups
as one method of expressing the meaning of a brand, 3. research that explores the
nature of the relationship between consumers and brands, 4. research into brands
122
7 Research on Uncertain Listeners
that show the value of prestige, and 5. research on consumers’ emotional ties with
brands.18
Among these research studies, the ones that relate to segmentation through brands
are those in the second category: that is, research exploring the relationship between
reference groups and brands. Reference groups refer to the groups that people belong
to and affect their individual behavior. There are various theories on the roles of
reference groups, with the study by Park and Lessing (1977) being the most cited
one in consumer behavior research.19 They proposed that reference groups have 1.
a role as sources of information, 2. a role for utilitarian judgments, and 3. a role for
value expression. As brands are a method of self-expression, people who belong to
the same reference group will prefer the same brand, as a method of value expression.
Research on purchases of brand products has demonstrated that reference groups
have influence. Bearden and Etzel (1982) examined product and brand decisions
within a panel of 645 consumers and found that reference group effects are stronger
for publicly consumed brands.20 Childers and Rao (1992) replicated the earlier study
with a sample of 345 American and Thai consumers and obtained similar results.21
Studies have also shown that brands that are preferred by a reference group that
a person does not belong to are not preferred by that person. Escalas and Bettman
(2005) conducted an experiment among students and found that consumers report
stronger self-brand connections for brands with images that are consistent with the
image of an in-segment than for brands not in this category. Brands consistent with
out-segment are less likely to show a self-brand connection than brands that do not
belong to this category. Brands not consistent with out-segment have a positive effect
on self-brand connections too.22 Furthermore, researchers have noted that even in a
reference group that a person does not belong to, the degree of preference for a brand
will be different, depending on the psychological distance of the person in question.
White and Dahl (2007) carried out an experiment among 55 students and found that
consumers showed a greater tendency to avoid products associated with dissociative
reference groups than with out-segments.23 In other words, as the brands preferred
by each reference group are different, when investigating the degree of preference
for a brand, several non-exclusive reference groups can be created.
From these research studies, we see that brands can be used to communicate
membership in particular social or professional segments, through both the use of
brands that signal membership in desirable segments and the avoidance of brands
that signal membership in undesirable segments. We can construct several segments
by using consumers’ degree of preference for a brand and expect that these segments
will reflect the consumers’ lifestyles.
7.3 Possibilities for Constructing Segments Using Brands
First, before creating the segments, I will simply explain the actual attitudes of the
Japanese toward brand products. Until the year 2000 or so, the Japanese liked many
of the world’s leading brands.24 However, of late, due to the simple use of social
7.3 Possibilities for Constructing Segments Using Brands
123
networking service (SNS), the trend of self-expression through SNS has progressed
throughout the world, particularly among young people. Moreover, they can be used
for self-expression outside of a wide range of real reference groups too. Chernev
et al. (2011) stated the following about the rise of SNS as a trend for self-expression
other than by brands25 : “The rapid growth of social media and peer-to-peer communications present another opportunity for self-expression. Facebook, YouTube, and
Twitter provide customers with an environment in which they can voice their opinions and find other people that share the same interests, thus enriching their social
identity.”
In Japan, a high percentage of young people use SNS. According to a survey
among young people conducted in 2019 by the Research Institute for High Life,
25.2% of males and 46.3% of females use Instagram four or more times a week as a
method of self-expression, while 55.6% of males and 73.9% of females use LINE, an
information-exchange-tool, within their reference group four or more times a week.
In fact, the percentage of the Japanese who own a luxury brand is not high,
particularly among young people. According to a survey on luxury brand ownership
conducted by MyVoice Communications in 2019 and 2016, among 10,000 people
aged in their teens to their sixties (2019: 10,611 people—5,711 males, 4,900 females;
2016: 10,890 people—5,508 males, 5,382 females), the percentages of people who
answered that they did not own even 1 of the 27 luxury brands surveyed was 48.1% in
the 2019 survey and 46.1% in the 2016 survey.26 In particular, 80% of the respondents
aged in their teens and twenties did not own even one of the brands. As the young
Japanese do not own luxury brands, the conventional idea that reference groups affect
the ownership of luxury brands cannot be used for young people.
However, the ownership rate among young people of the so-called fast fashion
brands, which are inexpensive but still highly fashionable, is high. TesTee Lab, which
disseminates results of research into young people, conducted a survey of fast fashion
brands in March 2019 among 1,808 people in their teens and twenties (428 males
and 419 females in their teens, and 481 males and 480 females in their twenties).27
According to this survey, the usage rate of Uniqlo, which is a leading fast fashion
brand in Japan, was more than 90%, while more than 70% of males and 90% of
females had the experience of using the GU brand. For overseas brands, the usage
rate of H&M was slightly less than 30% among males and slightly more than 50%
among females, and 15% of males and 30% of females used ZARA, indicating major
differences between users and non-users. The usage rates of Uniqlo and GU among
young people are high. While segments cannot be created from the usage and nonusage of these two brands alone, there are differences in the usage rates of overseas
brands for the same fast fashion brands, which allows for the creation of segments
using these brands. This shows that when using brands to create segments that include
young people, it is necessary to include not only luxury brands but also fast fashion
brands.
The use of luxury brands is also growing rapidly among young people through
rentals, or the so-called subscription method. According to estimates by the Yano
Research Institute, the scale of the subscriptions market in 2018 was 560 billion yen,
and it is projected to reach 860 billion yen by 2023. Laxus, which is under the umbrella
124
7 Research on Uncertain Listeners
of a major apparel manufacturer, is a typical example of a service for brand product
subscriptions. About 250,000 people use its app, which has a continuous usage rate
of 95%, with a majority of users being women in their twenties and thirties.28 This
shows that instead of creating segments from the perspective of “brand ownership,”
it is suitable to create them using “brand preferences.”
Among the conventional luxury brands, according to the previously mentioned
survey in 2019 by MyVoice Communications, the top three ranked brands by
ownership rate were Coach (19.8%), Burberry (17.5%), and Louis Vuitton (17.3%),
followed by Gucci (14.9%), Chanel (10.2%), and Tiffany (10.1%), which shows that
the rates are widely distributed. In the 2016 survey as well, the top three ranked brands
by ownership rate had been the same, namely Coach (20.8%), Burberry (18.7%), and
Louis Vuitton (17.8%), followed by Gucci (15.5%), Tiffany (11.0%), and Hermes
(10.8%). As can be seen, the rates were widely distributed in both the surveys,
revealing that consumer preferences are divided. In short, the suggestion of past
research that reference groups affect purchases of luxury brands seems to apply to
groups other than young people.
Interestingly, even though nearly 50% of the survey subjects own no brands, the
ownership rates of the top three brands show numbers of a certain level. This leads
to the inference that consumers who own luxury brands must be owning multiple
brands. In fact, a survey of luxury brand ownership conducted among the Japanese
in 2000 showed that people with a high intention to purchase “Max Mara” also have
a high intention to buy “Bulgari,” “Armani,” “Hermes,” and “Jil Sander,” and that
the people who support these brands are chic oriented. The study also revealed that
people with a high intention to purchase “Tiffany” also have a high intention to
buy “Cartier,” “Louis Vuitton,” “Prada,” and “Feragamo,” and that the people who
support these brands are strongly prestige oriented (Sasaki 2000).29 In other words,
when creating reference groups, it is necessary to combine multiple brands.
The above-mentioned actual conditions offer the following conclusions:
• When carrying out segmentation using brands in Japan, the researchers must also
consider young people and include not only luxury brands but fast fashion brands
as well.
• Based on the current situation of subscriptions, rather than brand ownership, they
need to use brand preferences as the scale.
• Instead of determining the reference groups from a single brand, they should
consider combining multiple brands.
Now, I will attempt to create segments using actual data.
7.4 Data Used to Create the Segments
I used data from the CANVASS consumer awareness survey carried out every year by
Yomiko Advertising to create the segments. In CANVASS, the placement method is
used to survey more than 1,000 items with 1,700 general consumers. Currently, nearly
7.4 Data Used to Create the Segments
125
all marketing-related surveys are transitioning to being online surveys. However, this
survey uses the placement method, which has an advantage that it also targets respondents who do not participate in online surveys and whose information sensitivity is
comparatively low. The awareness items include lifestyle-related items, such as food,
clothing, and shelter, and also information, society, and shopping. The data provided
for use in this analysis were from three points in time: 2002, 2004, and 2006.
I created lifestyle segments from the consumer brand preferences at each of these
respective points in time and measured the stability of the segments from whether
or not the same segment was extracted at each point in time. After this, I clarified
the characteristics of the consumers who belong to the created segments from the
awareness survey. I used 70 brands, including both luxury and fast fashion brands
that were continuously surveyed at the three points in time to create the segments.
I also used the actual sales data obtained from the POS of the Distribution
Economics Institute of Japan (DEIJ) at the three points in time to search for the relationships with the food brands supported by each of the above-mentioned segments.
DEIJ has the data on purchase-histories of processed foods and daily commodities in supermarkets. In CANVASS, for more than 20 food categories (for example,
beers, snacks, instant noodles), the subjects’ purchase histories, the brands they have
purchased the most in the food category during the last one year, and their affective
and calculative commitment toward these brands they had purchased the most were
surveyed. The results of the survey showed which consumer segments support which
brands from within the food categories. By combining this data and the actual sales
data according to brands provided by the Distribution Economics Institute of Japan,
we can measure the relationships with the support segments at the level of brands.
Furthermore, for confirming the robustness of each segment created using
CANVASS, I conducted a survey of preferences for the same 70 brands among 2,068
monitors of the Distribution Economics Institute of Japan and created the segments
based on this.
7.5 Segments and the Profiles
First, I used the data on preferences for the 70 brands surveyed in CANVASS and
created segments separately for the three points in time for which the data were
provided. As previously mentioned, in light of the conditions in Japan, when using
consumer preferences to create reference groups, one must consider the degree of
preference not for a single brand but for multiple brands. In particular, as there are
brands such as Uniqlo and GU that are preferred by many people, a decision has to
be made whether to combine these brands and other brands.
The previous studies on an area named consumer brand association indicate a
network of similarities between how consumers feel about multiple brands. For
example, Henderson, Iacobucchi, and Calder (1998) introduced two methods to
graphically represent various sport car brands30 : Repertory Grid and Pairwise Similarities Judgment. In the Repertory Grid method, the subjects are asked about their
126
7 Research on Uncertain Listeners
image of each brand’s characteristics—for example, classy, low price. From the
similarities of these characteristics, the distances of the similarities of each brand are
calculated and aggregated. Pairwise similarities judgment is a method in which each
of the subjects is asked to judge only the similarities between the brands. According
to Henderson et al., the latter method can be used for market segmentation. This is
because “the similarities between the associative networks of a sample of consumers
can be used to sort them into empirically derived segments based on their perceptions
of features thought to be descriptive of the brands in the competitive marketplace.”
Therefore, in this chapter too, I employ this method and use brands to carry out the
segmentation.
From the degrees of preference for the 70 brands, we observe that at each of
the three-year points in time, the degrees of preference for eight brands (a1 to a8)
are extremely high. Due to confidentiality obligations, I cannot reveal the specific
names of these eight brands, but they include a well-established sports brand, a
well-established luxury brand, a well-established casual brand, and a brand whose
seasonal popularity had passed as a brand.
Next, I created the consumer segments from consumers’ degrees of preference
(prefer/not prefer dummy) for these 70 brands. One of the segmentation criteria,
the ease of creation criterion, was met because it was created by asking about the
preferences of 70 brands. Analysis of each of the three years of data resulted in four,
distinctive segments that were stably derived: “early listeners,” “active listeners,”
“uncertain listeners,” and “disinterested listeners.” Since these four segments can
be extracted at all three points in time, robustness, which is one of the criteria for
segment creation, can be taken as being met. The remaining samples, in which
no stable segments are formed, were summarized as “passive listeners” segment.
Detailed descriptions are provided later, but brief explanations of the relationships
between the segments and the degrees of preference for the brands are given here.
“Early listeners” is a segment that prefers brands that have only very recently been
launched in the market and whose information sensitivity is extremely high. Like
“early listeners,” “active listeners” is a segment that prefers new brands, as well as the
above-mentioned eight brands. In contrast to this, “uncertain listeners” is a segment
that prefers some of the brands among the above-mentioned eight brands only, and
these people do not prefer currently fashionable brands. “Disinterested listeners” is
a segment that has no interest in brands, and these people do not prefer any of the 70
brands. “Passive listeners” is everyone other than the people in the above-mentioned
segments.
The first step was to see if the segments created by the preferred brand group met
the criteria for segmentation, as well as whether the clusters created by the preferred
brand group had a good grasp of lifestyle. For this reason, we used the 2006 data to
see if there was a difference between the segments created for the demographic and
lifestyle factors.
For the demographic factors, we used life stage variables and looked at the differences between segments. This is shown in Table 7.1. We can see from the size
of Pearson chi-square value that there are significant differences among these five
segments (χ2 (28) = 137.128, p < .0001). It shows that there are differences in the
7.5 Segments and the Profiles
127
Table 7.1 Life stage variables between segments
Uncertain
listeners
Disinterested
listeners
Active
listeners
Early
listeners
Passive
listener
Total
Teen ager
(generation
Y) (13–19)
Frequency
21
30
2
3
101
157
Expected
frequency
15.01
22.79
13.09
6.59
99.51
157
Post-baby
boomer
juniors
(20–23)
Frequency
11
23
20
20
113
187
Expected
frequency
17.88
27.15
15.59
7.85
118.52
187
Baby
boomer
juniors
(24–31)
Frequency
20
18
19
22
154
233
Expected
frequency
22.28
33.83
19.43
9.78
147.68
233
Post-new
humans
(32–39)
Frequency
30
28
18
19
184
279
Expected
frequency
26.68
40.51
23.26
11.71
176.84
279
New
humans
(40–48)
Frequency
28
31
19
5
169
252
Expected
frequency
24.10
36.59
21.01
10.58
159.72
252
Post-baby
boomer
(49–53)
Frequency
16
19
14
1
98
148
Expected
frequency
14.15
21.49
12.34
6.21
93.81
148
Baby
boomer
generation
(54–59)
Frequency
14
32
26
2
133
207
Expected
frequency
19.79
30.05
17.26
8.69
131.20
207
Senior
(60–69)
Frequency
24
68
25
0
135
252
Expected
frequency
24.10
36.59
21.01
10.58
159.72
252
Total
Frequency
164
249
143
72
1087
1715
Expected
frequency
164
249
143
72
1087
1715
life stages of consumers in each segment. From this Table 7.1, we can shows that
“early listener” segment is a high percentage of this segment comprises the young
generation, specifically those in the age range from post-new humans to post-baby
boomer juniors. On the other hand, a majority of “active listeners” are post-baby
boomer juniors and the baby boomer generation. For “disinterested listeners”, a high
percentage of them are teens and senior citizens. For both “passive listeners” segment
and “uncertain listeners” segment, there were no differences due to demographic
factors.
With respect to lifestyle, we looked at differences between segments on four
scales: clothing, food, information sensitivity, and high sensitivity. There is no
128
7 Research on Uncertain Listeners
uniform standard for the variables used in lifestyle research, as discussed in 7–1. For
this reason, here we have chosen to explore the differences between the segments
created above, based on the questions about clothing, information sensitivity, and
sensitivity to information, which were investigated in CANVASS. The differences
in the ideas about clothing, information sensitivity and high sensitivity are shown
in Figs. 7.1, 7.2 and 7.3, respectively. Since the number of people comprising each
segment is different, these figures show the average response rates for each segment.
I choose what I wear.
Change the color paern to suit your mood.
I have luxury brand shoes and bags.
Fashion is a form of self-expression.
Be mindful of TPO.
I'm also obsessed with shoes, bags and accessories.
I'm very parcular about what I wear.
Buy something that is expensive but will last a long me.
Enjoying Fashion
I tend to buy them at bargain prices.
There's a place for familiarity.
0.000
passive listener
early listeners
0.100
0.200
0.300
acve listeners
0.400
0.500
0.600
disinterested listeners
0.700
0.800
0.900
1.000
uncertain listeners
Fig. 7.1 Life style score for clothing
My source of informaon is television.
I want to learn more about many things.
I'll find out what I don't know.
If you don't know what to do, look it up online.
People can teach me a lot.
My informaon sources are online.
I value real word of mouth over the media.
I can idenfy the informaon that's right for me.
Read the DMs oen.
I'm a quick learner of informaon.
The media rather than word of mouth.
0.000
passive listener
early listeners
Fig. 7.2 Lifestyle score for infomation
0.100
0.200
acve listeners
0.300
0.400
disinterested listeners
0.500
0.600
uncertain listeners
0.700
7.5 Segments and the Profiles
129
Even when I don't buy, I check out the store.
I'll try trial campaigns and sampling.
Focus on the atmosphere of the store.
I like to think of new ways to use it.
They say I have good taste in fashion and music.
Buy with the help of mulple media and expert opinions.
Compare products with informaon from the Internet
Check out the hoest spots and products as soon as possible.
I can see the difference in quality between products.
Hit products are purchased relavely early.
I know roughly what's next in vogue.
0.00
passive listener
early listeners
0.10
acve listeners
0.20
0.30
disinterested listeners
0.40
0.50
0.60
0.70
uncertain listeners
Fig. 7.3 Lifestyle score for information sensitivity
The first “early listener” segment has absolutely no interest in the so-called wellestablished brands. They prefer only new brands and tend to choose clothes that
are currently fashionable. Their interest in clothes is high and they enjoy fashionable items, but their ownership rate of luxury brands is not high (Fig. 7.1). A large
percentage of them collect information from the Internet (Fig. 7.2). A majority of
them are the so-called people with good sense who actively compare products on the
Internet for fashion and music (Fig. 7.3). Overall, 72 people, or 4.2%, belonged to
this segment.
The second “active listeners” segment is highly flexible. On one hand, they did
prefer some so-called well-established brands. On the other hand, their degree of
preference was high for cutting-edge brands as well. They already own luxury brands,
but they also have the ability to discern TPO (Fig. 7.1). While some of them support
comparatively old-fashioned ideas and traditions, they also have many friends and
strong information-collection capabilities. In particular, many of them use real wordof-mouth information (Fig. 7.2). What we should focus on here is the fact that for
the high sensitivity indicators shown in Fig. 7.3. It can be said they are a segment of
people with extremely high sensitivity. As they support not only new brands but also
well-established brands, this segment is called a progressive-conservative segment.
As they act while listening to companies and the evaluations of people around them,
they are a segment with “active listening” type elements. Overall, 143 people, or
8.3%, belonged to this segment.
The segment named “disinterested listeners” showed no interest for even 1 of
the 70 brands surveyed. For their reaction values to clothes, they were 30–50 points
lower than the “active listeners” for every item (Fig. 7.1). The same trend was seen not
only for clothes but also for all the other lifestyle items. In particular, the difference
130
7 Research on Uncertain Listeners
with active listeners was large for the high sensitivity scores (Fig. 7.3). Overall, 249
people, or 14.5%, belonged to this segment.
The fourth segment prefers, from among the above-mentioned eight brands, only
well-established brands and brands whose seasonal popularity has passed. Its reactions to the clothes’ lifestyle items are similar to those of the “disinterested listeners”
segment. In some cases, their scores were even lower than those of the “disinterested
listeners” segment, depending on the item. While they answered that they were interested in specific, well-established brands, it is doubtful whether they truly understand
the value of a brand, as their high-sensitivity scores were extremely low. It is thought
that people in this segment select well-established brands, as they have no confidence
in evaluating the things they see with their own eyes. As shown below, the future is
uncertain for products that are preferred by this segment; that is why it is named the
“uncertain listeners” segment. Overall, 164 people, or 9.6%, belong to it.
The people who do not belong to any of these four segments were grouped in
the “passive listeners” segment. This is because although there were other segments,
such as segments that prefer sports brand and segments that prefer luxury brands,
these segments were not stably extracted every year. Overall, 1,087 people, or 63.4%,
belonged to this segment.
In this way, I created the segments using CANVASS. To confirm the stability of
these segments, I asked the same questions to the purchase panel of the Distribution
Economics Institute of Japan (DEIJ) and created the segments from the results.
Figure 7.5 shows the composition ratios of each segment when using CANVASS
and when using the panel of the DEIJ. We can see in the figure that compared to that
in CANVASS, the ratio of the “active listeners” segment is higher in DEIJ’s panel;
however, the ratios of the “disinterested listeners,” “early listeners,” and “passive
listeners” segments are lower. DEIJ’s panel uses a store panel method. It means this
panel contains people who are extremely loyal to specific stores and who cooperate
with various types of experiment surveys. So, inherently it comprises people with
high information sensitivity, and this is considered to be the reason behind the above
results (Table 7.2).
Figure 7.4 shows their high sensitivity reaction rates from these two surveys, based
on “active listeners” and “uncertain listeners, which have stable member numbers.
The items for which the reaction values of “active listeners” in the CANVASS are
higher than the reaction values of the panel of DEIJ were “They say I have good
Table 7.2 Composition ratios of CANVASS and DEIJ
CANVASS (frequency)
CANVASS %
DEIJ (freguency)
DEIJ %
Uncertain listeners
164
9.6
216
Disinterested listeners
249
14.5
181
8.8
Active listeners
143
8.3
553
26.7
Early listeners
10.4
72
4.2
34
1.6
Passive listener
1087
63.4
1084
52.4
Total
1715
100.0
2068
100.0
7.5 Segments and the Profiles
131
Even when I don't buy, I check out the store.
I'll try trial campaigns and sampling.
Focus on the atmosphere of the store.
I like to think of new ways to use it.
They say I have good taste in fashion and music.
Buy with the help of mulple media and expert opinions.
Compare products with informaon from the Internet
Check out the hoest spots and products as soon as possible.
I can see the difference in quality between products.
Hit products are purchased relavely early.
I know roughly what's next in vogue.
0.000
acve listeners (DEI)
uncertain listeners (DEI)
0.100
0.200
0.300
acve listeners (CANVASS)
0.400
0.500
0.600
0.700
uncertain listeners (Canvass)
Fig. 7.4 Difference of information sensitive score between CANVASS and DEIJ
taste in fashion and music.” and “I compare products with information from the
Internet.” Conversely, the items for which the reaction rates were higher in DEIJ’s
panel were “‘ll try trial campaigns and sampling.” For the other items, the differences
were not that great. In the “uncertain listeners” segment in DEIJ panel, their reaction
values were high for “even when I don’t buy, I check out the store.” and “I’ll try
trial campaigns and sampling.” For these items, the differences were great with the
“uncertain listeners” segment in CANVASS. As previously mentioned, DEIJ’s panel
is constructed to survey various types of stores. Many people who register on its
panel are asked to cooperate in surveys, such as on store experiments and campaign
effects, and it is believed that this is the reason behind such results.
7.6 Segments and Brand Evaluations
Now it is clear that consumers can be segmented from their degrees of preference
toward brands and that these segments are robust. Also differences in life stage and
lifestyle ratings between the segments created were evident. Next, I will use these
created segments to evaluate the sales of a processed food brand and confirm their
effectiveness.31
Table 7.3 shows the aggregated results obtained from POS data and card membership data (Frequent Shoppers Program data: FSP data) on variables relating to sale
methods, such as on processed foods brand shares and discount conditions in 2004
by DEIJ. Normally, from these numbers, aspects such as the following year’s sales
and product replacements can be determined.
From these numbers, the candidates for brands that may be taken off the shelves
the following year or later are considered to include J, K, and E, which have low
monetary shares; J, K, and C, which do not have high loyalty rates; and J, K, D, and
132
7 Research on Uncertain Listeners
Table 7.3 Sales index based on POS and FSP data
Market share
(monetary base)
Average selling
price (JPyen)
Average number
of face exposure
Loyalty rate
Special purchase
rate
A
5.3
163.7
2.4
21.5
92.3
B
3.3
206
3
26.1
64.3
C
4.6
157.6
1.5
15.4
93.5
D
2.9
175.4
1.4
15.9
81.7
E
0.4
229.6
2.3
17
29.2
F
8.1
160.1
1.7
22.6
90.7
G
2.5
248
2.6
28.4
59.4
H
2.8
229.1
1.4
17
51.6
I
9.9
173.2
1.9
25.4
92
J
0.1
257.9
1
10.3
0.3
K
0.2
189
1
13.2
36.7
H, whose average number of face exposures, which show the number of exposures
in stores, are low. When judged comprehensively, the first candidate for the brand
to be taken off the shelves is J or K. Of these 11 brands, the brand that was actually
taken off the shelves in 2006 was brand E. According to the numbers, E is a brand
that sold even when not in a bargain sale, and this result could not be expected from
the above information on discounts and sales methods. However, I will estimate the
reason why brand E was taken off the shelves from the segment supporting it.
Among the four previously shown segments, Fig. 7.5 shows the evaluations of
each of the brands from the rates of purchases in 2004 from the “active listeners”
70
Brand C
Brand G
deviation values of
active listener
60
Brand B
30
Brand K
40
Brand D
50
Brand F
Brand H
Brand A
60
Brand I
deviation values of
uncertain listener
Brand J
40
Brand E
30
Fig. 7.5 Positioning of brands based on deviation values of active listener and uncertain listener
7.6 Segments and Brand Evaluations
133
segment and the “uncertain listeners” segment. In this figure, the vertical axis is the
rate of the active listener segment and the horizontal axis is the rate of the uncertain
listeners segment. For making the differences in the rates clearer, each segment’s
respective rates are compared, and their relative strengths are shown by expressing
the deviation values.32 From this figure, we see that brand E has a high uncertain
listeners’ deviation value of 67, but its active listener segment’s deviation value is
less than 30. From this, we can understand that at the time of the 2004 survey, it was
not supported by the so-called high sensitivity people, but rather by low sensitivity
people. Similarly, we can also see that brand J had a high deviation value of 65,
showing it was supported by the uncertain listeners segment, while its active listener
segment’s deviation value was around 46, which is thought to be a factor why it was
not taken off the shelves. In contrast to this, brand K has a deviation value of 50 for
the active listener segment and 40, which is below average, for the uncertain listeners
segment. It can be said that its uncertain listeners rate is low, and therefore its risk
of being taken off the shelves is less than that of brand J.
Looking at the evaluations for a single fiscal year in this way, we can know the
respective rates of the uncertain listener and the active listener segments. Note that
we cannot know the market indicators created from the POS data and the FSP data.
When looking at the changes over the years, I will attempt the same for products
with strong sales, among the new products.
Figure 7.6 shows new products that first appeared in 2006, and the trends in the
brands that had already been on the market but not been performing well in recent
marketing indicators. First, in the case of brand x, which in this year secured a
share of 2.9% of the brands that first appeared in 2006 and a brand loyalty rate of
15.6%, its active listener segment’s deviation value was 63 and its uncertain listeners’
deviation value was less than 40. From this, we understand that from the time it
70
deviaon values of
acve listener
Brand X
60
30
40
50
60
Brand F
2002
Brand F
2004
40
Brand F
2006
30
Fig. 7.6 Positioning of powerful new Brand X and weaker Brand F
deviaon values of
uncertain listener
134
7 Research on Uncertain Listeners
was first launched, it was extremely popular and supported by customer segments.
Among those products whose sales increased immediately after being launched, there
were some brands that secured market share by luring away many switch consumers
through promotions. However, in the case of brand X, it is estimated that it had true
brand power.
In contrast to this, for brand F, the active listener segment’s share fell relatively in
2002, 2004, and 2006, while the uncertain listener rate was also close to a deviation
value of 50. It means that F’s brand power was falling over these three points in time.
The figures extract and express the actual marketing indicators for F. Although these
figures show that the loyalty rate was basically unchanged, its average sales price
and share both trended downward, and its brand power was falling.
From Figs. 7.7, 7.8 and 7.9 show the rates of the listening segments based on the
AISAS® flow for three hit products from two years ago. As described in Chap. 2,
AISAS® shows the decision-making flow, from the consumer paying attention to the
product, through the information search and the purchase, and finally the word-ofmouth communication. The processes are reflected in the full form of AISAS: Attention, Interest, Search, Action, and Share. As it is difficult to create segments distinguishing Attention and Interest, here, these two are combined into “Has awareness”
and are aggregated.
From these figures, we understand that for each of the products, the rate of active
listeners rises alongside the progress of the decision-making process, while the rates
of uncertain listeners and disinterested listeners fall. In whichever product category,
the active listeners’ information-sharing rate is extremely high. It means they are
considered to be market mavens, and in SIPS, they have an existence close to that of
100
90
3.4
4.2
12.3
3.7
4.3
4.1
25.4
26.7
29.7
59.3
58.5
15.2
80
70
60
57.6
59.6
50
40
57.7
30
20
10
8.5
7.8
18.2
13.3
5
6.7
0
Total
aenon & interest
disinterested listeners
uncertain listeners
Fig. 7.7 AISAS flow of beer brand A
search
passive listener
4.7
5.8
acon
acve listeners
3.7
4.7
share
early listeners
7.6 Segments and Brand Evaluations
135
120
100
3.4
4.2
12.3
2.3
1.4
0.9
16.6
80
36.3
45.9
51.3
60
57.6
59.6
40
52.3
20
41.4
35.9
8.5
7.1
18.2
12.8
4.4
4.4
5
6.3
6.8
5.1
aenon & interest
search
acon
share
0
Total
disinterested listeners
uncertain listeners
passive listener
acve listeners
early listeners
Fig. 7.8 AISAS flow of process food brand B
120
100
3.4
4.7
5.3
18.8
20.3
1.1
0
12.3
80
39.1
50.6
60
57.6
59
58.9
40
50.6
42.6
20
8.5
6.4
5.1
11.1
10.5
4.6
4.6
3.3
3.3
aenon & interest
search
acon
share
18.2
0
Total
disinterested listeners
uncertain listeners
Fig. 7.9 AISAS flow of durable brand C
passive listener
acve listeners
early listeners
136
7 Research on Uncertain Listeners
evangelists. In contrast to this flow of good product information from active listeners
to potential customers, the rates of uncertain listeners and disinterested listeners who
make a purchase are low. Even if they do make a purchase, they do not share information. Furthermore, in the early listener segment, which is an advanced information
segment the same as the active listener segment, they hardly share any information
even if they conduct information search activities and make a purchase. As the name
indicates, early listeners catch information at an early stage, such as on new products. However, as they do not share information, we can assume them to be a concept
similar to innovators based on past research.
In this way, many of the people who share information on hit products are active
listeners. From this, we can understand the importance of ascertaining the purchase
segments for a product to become a hit, in addition to sales. It is also vital to ascertain
the people with high information sensitivity and who make a product a topic of
conversation after the purchase. In other words, it is important to ensure that a high
percentage of the total listeners are active listeners.
7.7 Conclusion and Future Implication
In this chapter, with reference to lifestyle research that is often used for marketing
strategy as a segmentation axis, I constructed five segments from consumer brand
preferences. As brands are tools of self-expression, past research has relied on
showing that they can be created from these reference groups and that segments
can be created from multiple brand preferences. In particular, as shown in Chap. 6,
we know that differences in consumers’ information sensitivity significantly affect
their product evaluations. In this chapter, the focus was on finding whether a brand
becomes popular or unpopular, and this can be used to determine the differences in
consumers’ information sensitivity.
First, I created five segments from the preferences for an actual clothing brand
through a questionnaire survey. These five segments are: “early listener” segment,
which only likes advanced brands; the “active listener” segment, which likes leading
brands and well-established brands; the “uncertain listener” segment, which only
likes well-established brands; the “disinterested listener” segment, which shows
absolutely no interest in any of the brands; and the “passive listener” segment,
consisting of everybody else. As each of these segments was extracted at all the
three points in time surveyed, we can conclude that they were comparatively robust.
Each segment’s degree of interest in the clothing, their information sensitivity, and
their high sensitivity indicators were compared. As there were significant differences
in information sensitivity and the high sensitivity indicators, it was found that the
“early listener” segment, which likes cutting-edge brands, and the “active listener”
segment scored extremely highly for sensitivity. Conversely, the “uncertain listener”
and “disinterested listener” segments, which do not like cutting-edge brands, scored
lowly not only for the high sensitivity indicators but also for information sensitivity.
7.7 Conclusion and Future Implication
137
I then created the same segments using the store panel of DEIJ. Although there
were differences between its results and the results of the questionnaire survey for
the composition ratios of people, basically the same trend was seen in this panel for
the high sensitivity indicators, further increasing the robustness of these segments.
Next, for further clarifying the segments’ effectiveness, I combined actual sales
data on processed food. Among the people who support these food brands, I focused
on the composition rates of the active listener segment and the uncertain listeners
segment. As a result, even for brands that were not considered to have problems from
the marketing indicators created with the so-called POS data and FSP data, I showed
that brands are taken off the shelves when the rate of the active listener segment is
low and the rate of the uncertain listeners segment is high. I also showed that among
new products, strong brands have many active listeners and few uncertain listeners
from their launches. Another result of the study was that in the case of brands for
which the rate of active listeners decreases and the rate of uncertain listeners increases
over the passage of time, their numbers deteriorate over the years even in the actual
marketing indicators. Moreover, on taking an overview of the trends of the segments
that purchase hit products according to the flow of AISAS®, at the stage of sharing
information after a purchase, it was found that the rate of active listeners is extremely
high and the rate of uncertain listeners is extremely low. From these facts, it can be
confirmed that by measuring, in particular, the purchaser profiles and the respective
sizes of the rates of these active listeners and uncertain listeners segments, it becomes
possible to evaluate brands more effectively than by using marketing indicators.
In this analysis, for the convenience of obtaining data, the results were obtained
mainly from the processed food category. It was clarified that rather than looking at the
numbers sold, this research looks at who are making purchases. Promotions, which
until now have been centered on discounts, do have a major impact on sales. However,
they have been frequently criticized too, such as for reducing the reference price and
lowering the brand value. With regard to these criticisms, from the commitment
research in Chap. 4, we saw that a discounted purchase is regarded as one of the
reasons for a fall in emotional commitment. While from the analysis in this chapter,
it can be stated that the lowering of brand value is from the decline in emotional
commitment and also from the changes to the purchase segments. On considering
this together with the effects of blogs and other elements discussed in Chap. 3, we
can conclude that whether or not a brand is supported by the active listener segment
will have an effect on the evaluation of the brand not only at the present time but in
the future as well.
In terms of the method of segmenting consumers from their preferences for these
brands, in retail stores such as department stores, which hold data on customers’
purchase histories of luxury-brand clothing, it has been shown that it is possible to
identify customers’ lifestyles from their purchase-history data. The leading retail
stores utilize data mining technologies to manage various customers. Retail stores
handling brands that have symbol characteristics can also use this purchase data
to estimate customers’ lifestyle without conducting a questionnaire survey. This is
considered to be an extremely effective method.
138
7 Research on Uncertain Listeners
Overall, the findings show the importance of constantly managing those who are
the purchasers of products rather than managing the number of products sold.
Notes
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Escalas, Jennifer Edson; Bettman, James R. (2005), ‘Self-Construal, Reference
Groups, and Brand Meaning’, Journal of Consumer Research, Vol. 32, pp. 378–
389.
Wind, Yoram (1978), ‘Issues and Advances in Segmentation Research’,
Journal of Marketing Research, Vol. 15, No. 3, pp. 317–337.
Lazer, William (1963), ‘Life Style Concepts and Marketing’, AMA Proceedings on Scientific Marketing, AMA, pp. 140–151.
Wells, William (1975), ‘Psychographics; A Critical Review’, Journal of
Marketing Research, Vol. 12, pp. 196–213.
Wind, Yoram (1978), ibid.
Beane, T.P.; Ennis, D.M. (1987), ‘Market Segmentation: A Review’, European
Journal of Marketing, Vol. 21, pp. 20–42.
Wedel, Michel; Wagner, Kamakura A. (2000), Market Segmentation—
Conceptual and Methodological Foundations Second Edition’, International
Series in Quantitative Marketing, KAP.
For example, Plummer classified the users and non-users of a bank card by
combining the AIO scale and demographic factors, but the items on the AIO
scale increased to 300. For details, refer to Plummer, Joseph (1971), ‘Life Style
Patterns and Commercial Bank Credit Card Usage’, Journal of Advertising
Research, Vol. 35, pp. 35–41.
Kotler, P. (1967), Marketing Management, Prentice-Hall.
DeSarbo, Wayne; Grisaffe, Douglas (1998), ‘Combinatorial Optimization
Approaches to Constrained Market Segmentation: An Application to Industrial
Market Segmentation’, Marketing Letters, Vol. 9, pp. 115–13.
Wells, William; Tigert, Douglas (1971), ‘Activities, Interests, and Opinions’,
Journal of Advertising Research, Vol. 11, pp. 27–35.
Lastovicka, John (1982) ‘On the Validation of Lifestyle Traits: A Review and
Illustration’, Journal of Marketing Research, Vol. 19, No. 2, pp. 126–138, and
Lastovicka, John, Murry, Jr., John P., Joachimsthaler, Erich (1990), ‘Evaluating
the Measurement Validity of Lifestyle Typologies With Qualitative Measures
and Multiplicative Factoring’, Journal of Marketing Research, Vol. 27, No. 1,
pp. 11–23.
Wagner, Kamakura A.; Wedel, Michel (1995), ‘Life-Style Segmentation with
Tailored Interviewing’, Journal of Marketing Research, Vol. 32, No. 3, pp. 308–
317.
Kahle, Lynn; Beatty, Sharon E.; Pamela, Hommer (1986), ‘Alternative
Measurement Approaches to Consumer Value: The List of Values (LOV)
and Values and Life Style (VALS), Journal of Consumer Research, Vol. 13,
pp. 405–409, and Novak, Thomas P.; MacEvoy, Bruce (1990), ‘On Comparing
7.7 Conclusion and Future Implication
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
139
Alternative Segmentation Schemes: The List of Value (LOV) and Values and
Life Style (VALS)’, Journal of Consumer Research, Vol. 17, pp. 105–109.
Steenkamp, Jan-Benedict; Hofstede, Frenkel Ter (2002), ‘International Market
Segmentation: Issues and Perspectives’, International Journal of Research
in Marketing, Vol. 19, pp. 185–213.
Wedel, Michel; Wagner, Kamakura A. (2000), ibid.
Puzakova, Marina; Kwak, Hyokjin; Rocereto, Joseph (2009), ‘Pushing the
Envelope of Brand and Personality: Antecedents and Moderators of Anthropomorphized Brands’, Advances in Consumer Research, Vol. 36, pp. 413–420.
Chernev, Alexander; Hamilton, Ryan; Gal, David (2011), ‘Competing for
Consumer Identity: Limit to Self-expression and Perils of Lifestyle Branding’,
Journal of Marketing, Vol. 75, pp. 66–82.
Park, C. Warn; Lessing, V. Parker (1977), ‘Students and Housewives: Differences in Susceptibility to Reference Group Influence’, Journal of Consumer
Research, Vol. 4, pp. 102–110.
Bearden, William; Etzel, Michael (1982), ‘Reference Group Influence on
Product and Brand Purchase Decisions’, Journal of Consumer Research, Vol.
9, pp. 183–194.
Childer, Terry; Rao, Akshay (1992), ‘The Influence of Familial and Peer-based
Reference Groups on Consumer Decisions’, Journal of Consumer Research,
Vol. 19, pp. 198–211.
Escalas, Jennifer; Bettman, James (2005), ibd.
White, Katherin; Dahl, Darren (2007), ‘Are All Out-Groups Created
Equal? Consumer Identity and Dissociative Influence’, Journal of Consumer
Research, Vol. 34, pp. 525–535.
Looking at the percentages of total sales of Gucci around 1996, 22% were
from Japan, but on looking at the nationalities of purchases around the world,
it was noted that 70% of sales were to Japanese. For the details, please refer
to Katahira, Hotaka (1998), The Essence of Power Brands, DIAMOND, Inc.
(written in Japanese).
Chernev, Alexander, Hamilton, Ryan, Gal, David (2011), ibid.
https://myel.myvoice.jp/products/detail.php?product_id=24810.
https://lab.testee.co/fastfashion#i-9.
https://corp.laxus.co/.
Sasaki, Reiko (2000), Two Consumer Groups Who Support the Popularity
of Overseas Brands—‘The Prestige Faction’ and ‘The Chic Faction’, Nikkei
Research Institute of Industry and Markets Monthly Forum Newsletter, pp. 8–
13. (written in Japanese).
Henderson, Geraldine; Iacobucchi, Dawn; Calder, Bobby (1998), ‘Brand Diagnostics: Mapping Branding Effects Using Consumer Associative Networks’,
European Journal of Operational Research, Vol. 111, pp. 306–327.
Although the targeted processed foods have differences in terms of size and
taste, as the targeted stores in this survey were supermarkets, the top selling
sizes and tastes of the brands were extracted and aggregated.
140
7 Research on Uncertain Listeners
32.
‘Deviation’ Value’ is a statistical indicator that has been used for many years
in Japanese admissions tests as a measure of a school’s admission difficulty. It
is calculated as (your score—average score)/standard deviation × 10 + 50.
Chapter 8
Studies on Connoisseurs
As outlined in Chaps 6 and 7, it has become clear that a brand can be assessed based
on the purchase activities of information-savvy consumers. In this chapter, I will go
a step further and consider forecasting the demand for new products by surveying
such information-savvy consumers. Given that it was possible to assess a brand based
on the purchase behaviors of information-savvy consumers, the ultimate aim is to
develop a system that provides feedback for product development by surveying them
in detail, and even understanding the reasons why they recognize particular brands.
In today’s Japan, new products are revamped or withdrawn rapidly; in particular,
products that are mainly sold through the channel of convenience stores face an
extremely tough situation, wherein they are removed from the shelf if they do not meet
a certain sales standard four weeks after release. Figure 8.1 compares the percentage
of newly-introduced products (by product category) that survived more than one
year at 1999, 2006, and 2019. While it differs by product category, the percentages
are extremely low at 75% for daily goods with longer consumer purchase intervals
in 1999 and around 50% for processed foods and snacks with shorter purchasing
intervals, indicating that the situation has become more competitive than before.
Therefore, manufacturers are required to constantly develop new quality products
and, at the same time, measure the product performance as early as possible in
order to prevent excess inventory. In the case of the products that must be purchased
repeatedly, the product appeal has been often measured based on consumers’ trial and
repeat purchases, which had actually produced decent results in the past. However,
those analyses required a certain time frame. There is a limit to the measurement based
on trial and repeat, since products with a longer purchase interval are discontinued
before repeat purchases occur in Japan’s situation, where products are removed from
the shelf in a mere four weeks.
In this chapter, I will propose a new demand forecasting model that surveys
information-savvy consumers to determine the success or failure of a given product
at an earlier point in time. It is based on the Study on Connoisseurs I have
been conducting jointly with Asahi Breweries’ customer lifestyle research center
HAPIKEN since 2003. In this study, we have researched and developed a method
© Springer Nature Singapore Pte Ltd. 2021
A. Shimizu, New Consumer Behavior Theories from Japan, Advances in Japanese
Business and Economics 27, https://doi.org/10.1007/978-981-16-1127-8_8
141
142
8 Studies on Connoisseurs
80
73.4
70.2
70
59.7
60
75.9
68.1
58.9
54.7 53.7
48.8
50
42.1
38.4
40
38.3
30
20
10
0
processed food
confeconery
1999
Beverages and alcoholic
beverages
2006
daily goods
2019
Fig. 8.1 Percentage surviving more than 1 year
to forecast the demand for new beer products before release. The basic framework
of the survey involves a method in which a panel of “connoisseurs” is set up on the
Internet to judge the success or failure by looking at pre-sales information (specifically, the press release). HAPIKEN had already been forecasting the demand before
the release by using methods such as market testing and group interviews. However,
the fundamental difference among these methods is that it is highly objective because
it uses a large consumer panel (about 1,000 individuals). Furthermore, it can forecast
not only the sales of own products but also the sales of products from competitors
since it uses published press releases. The method allows the panel to forecast without
seeing the actual products.
I will now review studies on conventional demand forecast and trendsetting
consumers and then introduce our forecasting attempt.
8.1 Demand Forecast for New Products
Various methods have been tried for forecasting the demand for new products. For
example, the classic work by Wind, Mahajan, and Cardozo (1981) divided these
methods into four major categories: (1) models based on concept testing, (2) models
that use the results of pre-test and forecast, (3) forecasting models based on market
testing, and (4) models that forecast based on the initial state after the product release.1
Of these, the forecasting models based on market testing under (3) and the models
that forecast based on the initial state after the product release under (4) are the ones
8.1 Demand Forecast for New Products
143
that measure how well the product will sell after product development is completed
and the product is actually manufactured.
The categories (3) and (4) are further divided into categories based on the type of
product in question. Deeming suggested that the effect of trial and repeat purchases
differs by durable versus non-durable goods, while Katahira proposed a classification
of models according to the nature of goods. This is because, while we only need to
understand trial purchases in the case of durable goods, since the market share is
established once it is purchased, subsequent repeat purchases have a greater impact
on market share than the trial purchase in the case of non-durable goods. Katahira
(1983) classified models by calling forecasting models for durable goods as “trial
models” and models for non-durable goods that need to consider repeat purchases as a
“trial-repeat models.”2 Nakamura (2001) further developed Katahira’s classification
and divided demand forecast models by adding the frequency of product purchase
to the trial purchase and repeat purchase axes.3
The consumer packaged goods I will analyze are for whom repeat purchases are
more important than that of trial purchases. Furthermore, non-durable goods are a
frequently-purchased product category. Therefore, to follow Nakamura’s classification, models that are deemed suitable include TRACKER4 by Blattberg and Golanty
(1978), which forecasts based on the results of market testing, and the Perfit & Collins
model, which uses panel data and forecasts the ultimate penetration rate based on trial
and repeat purchases immediately after the product release5 (Perfit, Collins 1968).
TRACKER is a demand forecasting model consisting of three models: a brand
name recognition model, a trial purchase model, and a forecasting model. It forecasts
demand by testing the market for three months, surveying 500–1,000 individuals in
the test market, and using the data, such as the value of the brand, spending on
media, and distribution status. Specifically, the name recognition model, which is
calculated based on the brand perception and advertising spending weight, assumes
that the number of people who recognize the product declines when the weight of
advertising spending declines. The trial purchase model measures the number of
people who make a trial purchase of the product during the period t. It is created
by defining two kinds of people—those who learned about the product for the first
time during the period t and those who had known but had not tried the product—
as potential customers and adding the product price to it. The last forecast model
improves the accuracy of forecast by dividing people into those who make a trial
purchase and those who make repeat purchases. According to Blattberg et al., it was
shown that sales can be forecast for many product categories.
The Perfitt & Collins model forecasts the final market share by calculating the
weights of trial purchase rate, repeat purchase rate, and purchase amount level (heavy,
average, light) based on the panel data and multiplying them. While the trial purchase
rate is calculated based on the percentage of trial purchasers among new purchasers
in the product category that a given product falls under, the repeat purchase rate
of each individual is calculated based on the number of purchases made with the
brand out of the total number of purchases made after trying a product within the
product category. The purchase amount level is calculated by dividing the number
of brand products purchased by the number of people who purchased that brand; it
144
8 Studies on Connoisseurs
is a numerical value that can also be regarded as the average number of items per
purchase. Previous studies deem that the model can forecast the market share after a
year or year and a half with great accuracy if data is available over a period of about
24 weeks after the release.
While these two models use consumers’ trial and repeat purchase data, the data is
aggregated and not considered at the individual level; this is the so-called aggregated
analysis. Therefore, the heterogeneity among individuals is not taken into consideration, entailing caution when using these models for the categories of products that
are consumed more among larger families. In addition, long-term data is required in
the case of aggregated data analysis. Therefore, these models will not suitable when
you want to forecast demand at an early stage. Under the aforementioned current
circumstances in Japan, it is quite difficult to obtain panel data over a period of
24 weeks, even if the accuracy would be good, because a product that is still sold
24 weeks after the release is a successful product with no particular need to forecast
the demand. In addition, in the case of TRACKER, there is a need to conduct a costly
market testing which is also a problem.
For this reason, there have been attempts to forecast demand in a short time by
applying the idea of trial and repeat purchases to personal purchase history data and
increasing the amount of data by using it without aggregation. Now that personal
purchase history data, such as frequent shopper program data (FSP data), is readily
available compared to the past, this method is an effective solution. An example of
this is a model that incorporates the hazard function primarily used in predicting the
survival rate of living organisms. Sugita, Nakamura, Tajima (1993) developed a split
hazard model that incorporates household heterogeneity, extracted six brands from
different product categories, and compared the predictive accuracy with the existing
Perfitt & Collins model.6 Their model uses the 13 weeks of data after release for
estimating the model and examines the model’s predictive accuracy against data
for the next 17 weeks. While the Perfitt & Collins model demonstrated a higher
predictive accuracy for chocolate confectionery and malt beverages out of the six
product categories surveyed, the split hazard model had a higher predictive accuracy
for the remaining four categories, i.e., ice cream, cereals, grilled rice balls, and
chocolate snacks, demonstrating that highly accurate prediction was possible even
if the time period is short.
As described, it is clear that incorporating the idea of trial and repeat purchases into
consumer packaged goods is important for improving predictive accuracy. However,
even the model by Sugita et al. requires 13 weeks of data to forecast demand; it
cannot be used for products such as beverages, which are mainly sold through the
convenience store channel and removed from the shelf in four weeks. From the
perspective of data analysis, securing a certain amount of data is a must and the time
frame has to be set long. To reduce the number of data points, it becomes necessary
to compensate by assuming some kind of a special function system like the hazard
function. We need ideas from different perspectives in order to forecast demand while
accommodating the current situation in Japan.
8.2 Studies on Trendsetters
145
8.2 Studies on Trendsetters
The idea of using trendsetting consumers in marketing strategies has been applied
for a long time. The pioneering research consists of studies on innovators. Originating from the diffusion of innovations by Everett Rogers, the idea is to think
about marketing that uses the segment of consumers who purchase the product at an
early stage of introduction. Rogers deemed that the difference in the speed at which
consumers adopt a product would be a normal distribution and named the portion
that exceeds 2σ at the tip—that is, 2.5% of the entire population—as “innovators,”
the 13.5% that falls between 2σ and 1σ as “early adopters,” the 34% between 1σ and
the mean as “early majority,” the 34% of people falling between the mean and-1σ as
“late majority,” and the 16% of people who adopt after −1σ as “laggards.”
The first individual who applied this innovator theory of Rogers was Robertson
who surveyed such consumers and used it for forecasting new product demand
(Robertson 1967).7 Advocating to use innovators in order to improve the accuracy
of market testing, he measured innovators by (1) whether they can take risks, (2)
whether they possess socialness, (3) whether they interact with other communities,
(4) whether they have ability to deal with liquidity in society, and (5) whether their
income is high, and surveyed if they had adopted the touch-tone telephone. It showed
that the innovators scored high in all five criteria, and the number of other new household products they had adopted was also large. Based on this result, he deemed it
meaningful to use innovators when selecting a new product.
It was Bass (1969) who turned the innovator theory into a mathematical equation
and used it for forecasting the demand for new products.8 He attempted to calculate
how durable goods are adopted at home by hypothesizing that innovators influence
others and by creating a formula. His model is characterized by the fact that it
incorporated, from the onset, how the role of innovators, which is important at the
beginning, will decrease as time goes by. When the data on 11 durable goods was
actually input and examined, the coefficient of determination exceeded 0.8 in case
of nine durable goods, excluding black-and-white TV and household freezers, thus
indicating a high predictive accuracy.
As for the profile of innovators, many studies argue that it is difficult to distinguish them from general consumers based on demographic factors and lifestyles.
For example, Pizam (1972), who reviewed 37 papers on innovators through 1972,
concluded that it is not possible to identify them since no relationship was found
in 17 papers, which is nearly half of the papers.9 McDonald and Corkindale (2003)
reviewed a massive number of past studies on innovators and showed that while there
are papers mentioning demographic and psychological factors as the characteristics
of innovators, those relationships are not evident in the majority of papers.10 On that
premise, they conducted a telephone survey among 1,000 housewives on adopting a
compact fluorescent light and attempted to clarify the characteristics of innovators.
The results showed that individuals with a significant interest in the product are more
likely to become innovators than young people, or those who view mass media often.
Similar results are also noted in Taylor’s study (1977).11 He looked at 14 consumer
146
8 Studies on Connoisseurs
packaged goods released in the past three years and studied the characteristics of
innovators. Based on his study, it became clear that those who made a purchase in
the first three months after the release were not innovators but heavy users in the
product category. He concluded that the probability of these people purchasing a
new product inevitably increases because the amount of purchases they make in the
product category is large.
As described, it has been thought that identifying innovators based on profile is
difficult and that the level of interest and the amount of purchase in a given product
category are the determinants. However, it is considered possible to identify innovators if we are to regard them as individuals who have purchased or intend to purchase
innovative new products rather than merely regarding them as those who adopt at
an early stage. For example, Venkatraman (1991) categorized innovators into cognitive innovators who solve problems with logical thinking and sensory innovators
who have a taste for new, mentally-stimulating experience and sometimes take pleasure in taking risks, and surveyed them about purchasing new, innovative PCs and
VCRs.12 Based on this, those who adopted PCs were highly-educated young men, and
those who adopted VCRs were young non-white collar workers with high income. It
became clear that while cognitive innovators are attracted to novelty, sensory innovators are attracted to risks; in terms of products, cognitive innovators are attracted
to the innovative PC and sensory innovators are attracted to the VCR. Im, Bayus, and
Mason (2003) attempted to explain the ownership rates for ten innovative electric
appliances by using the concept of innate consumer innovativeness.13 According to
this, it was found that the behavior of adopting new products is affected by high
income, young age, and innate consumer innovativeness. However, since the coefficients showed that income and young age had larger effects than innate consumer
innovativeness, it indicated that there are added restrictions to the actual behavior
even if the attitudinal aspect of innate consumer innovativeness is strong. Cowart,
Fox, and Wilson (2008) conducted an empirical study on the structural relationship of
innate consumer innovativeness, perceptual risk, and self-congruence by using home
entertainment equipment, music, and mobile devices as examples.14 It presented a
structure in which innovation—which affects purchase intention by itself—has a
positive effect on the level of self-congruence, which in turn leads to satisfaction with
the product and results in action intention. This study suggests that psychological
factors are more important than demographic factors.
As described, research on innovators has evolved from studying innovators with
regard to the term’s original meaning (i.e., those who adopt at an early stage) to
studying intentions to purchase innovative products. Meanwhile, research on opinion
leaders also ranks with research on innovators and is used for studying trendsetters.
The idea of opinion leaders was first discussed in the study by Lazarsfeld et al.
They demonstrated, based on people’s voting behavior in the presidential election
in 1940, that the mass media does not directly affect people. They showed there are
individuals who play the role of receiving, editing, and disseminating the influence
originally spread by the mass media and named them “opinion leaders.” We can say
the big difference here is that while the idea of innovators focuses on whether they
adopted at an early stage or not, the idea of opinion leaders thinks of the effect on
8.2 Studies on Trendsetters
147
others. This study has been widely used in the field of behavioral science; according
to the study by Myers and Robertson (1972), there were more than 1,000 studies
on opinion leaders by 1972, not limited to the academic fields.15 In the field of
marketing, it was frequently used for popularizing fashion and the like because of its
nature. For example, Summers (1970) verified the characteristics of opinion leaders
in women’s fashion clothing with actual data.16 Based on this, it became clear that
factors such as social connection and personality, in addition to demographic factors
like being young and having an income, have significant effects. We can regard this
as having shown that social connection is an important factor in terms of spreading
information.
While there are several reasons why the research on opinion leaders has spread,
the biggest factor is that the measurement scale is already established. King and
Summers (1970) formulated seven questions regarding whether the person talks
to friends about the product, whether he/she becomes the center of conversation
at the time, whether he/she is trusted as a source of information, etc. as a scale
to measure the level of the opinion leader.17 They also confirmed the existence
of comprehensive opinion leaders by interviewing 1,000 sample population about
their level of opinion leadership in six product categories—packaged food, women’s
clothing, household cleaners, cosmetics, large home appliances, and small home
appliances—and collecting basic data such as demographic factors at the same time.
Based on this, each product category had 23% to 30% opinion leaders, and 46%
of the respondents were opinion leaders in two or more product categories while
2.2% were opinion leaders in all six product categories. Conversely, there were
31% of the respondents who did not become opinion leaders regardless of product
category. Looking at the relationship of overlapped products, the combinations of
large appliances and small household appliances, women’s clothing and cosmetics,
and packaged food and household cleaners were most overlapped, while the least
overlapped combination was cosmetics and large home appliances. In other words,
it indicated that there are overlaps between product groups that can be viewed with
the same interest.
This scale, which was created by King et al., has been widely used in subsequent
studies and is known to be very highly valid. For example, Yavas and Riecken (1982)
has shown the robustness of the papers that utilized the scale by King et al. by using
KR-20, which is one of the reliability coefficients.18 In addition, Riecken (1983)
checked if the scale, created by King et al., works when the same questions are asked
to different groups and demonstrated that it is applicable once the statements are
improved.19 Furthermore, Childers (1986) showed that the explanatory power would
increase when the questions are all revised to a five-point scale.20
As described, research on opinion leaders has evolved, while improving the indicators of King et al., and resulted in numerous studies. For example, according to the
study by Leonard-Barton (1985), opinion leaders are known to spread not only positive information but also negative information—which is consistent with the effect
of blogs as indicated in Chap. 4.21 Aoike (2002), who studied Japanese opinion
leaders, also showed that the percentage of interpersonal communication increases
when the news value is large; opinion leaders wish to talk about it even to strangers
148
8 Studies on Connoisseurs
they have never spoken to.22 This is a survey result that indicates the development
of the Internet and the importance of the role of opinion leaders.
A variety of other things have been mentioned in studies on opinion leaders.
According to Shibuya (2002), who sorted out studies on opinion leaders, some of the
findings include “the quality of opinion leaders is not necessarily innate,” “those who
are swayed by opinion leaders have similar characteristics,” “for an opinion leader
to emerge, the individual needs to be surrounded by others who are highly interested
in a given area,” “influence flows in multiple stages of interpersonal relationship
chains,” and “in terms of the characteristics of opinion leaders, only the interest in a
given product, media exposure, and knowledge are related to opinion leadership and
no difference in personal characteristics has been identified.”23
While ideas of innovators and opinion leaders have existed for a long time, “market
maven” is a relatively new idea. Proposed by Feick and Price(1987), market mavens
are defined as “individuals who have information about many kinds of products,
places to shop, and other facets of the market, and initiate discussions with and
respond to information requests from other consumers,” while “opinion leaders”
refers to those who affect the purchasing behavior of other consumers within a
particular product group.24 According to their analysis, it was found that the scale
by King et al., which measures opinion leaders, cannot extract the same factors as
the scale that explains market maven men, and the correlation between the degrees
of being an innovator and a market maven is 0.25 with non-durable goods and −0.06
with durable goods, claiming that market mavens are trendsetters who are different
from either opinion leaders or innovators.
Market mavens are attracting attention because consumers value the opinions
of individuals in their reference group more than the information provided by the
manufacturer when obtaining information on a product or purchasing a product. This
is considered important, particularly nowadays, when information can be exchanged
on the Internet. Moreover, since there is a study result indicating that what market
mavens favor is likely to be accepted by other consumers, because market mavens
follow the norm and seek uniqueness within that norm, targeting market mavens is
considered extremely meaningful in that sense (Clark and Goldsmith 2005).25 We
can say that they have characteristics akin to that of the “active listeners” segment
described in the last chapter.
As for the characteristics of market mavens, there are various findings. The telephone interview survey conducted by Price, Feick, and Guskey-Federouch(1987), has
shown that market mavens love shopping, read new product information magazines
like Consumer Reports, are proactive to contact with many media, and are smart
consumers who frequently use coupons.26 Elliott and Warfield (1993) mentioned
the difference between general consumers and market mavens in terms of their
decision-making process.27 They followed the conceptual diagram by Brisoux et al.
and determined whether general consumers and market mavens had differences in
some stages/sets for decision-making. The study showed that while there was no
difference in size between the “consideration” set and “reject” set, the sizes of the
“awareness” set, “processed” set, and “hold’ set differed in a way such that they
were all larger than that of general consumers. These indicate that market mavens
8.2 Studies on Trendsetters
149
take a group of brands they are aware of and reduce them between the process and
consideration stages; since their hold set is large, they have more products they can
consider next time even though they do not consider them at present.
Another relatively new idea related to trendsetters like market mavens is the
research on lead users. This research, which was an idea introduced by von Hippel
(1986), is a field that has been studied to show why general consumers should be
involved in new product development rather than to show the adoption of new products.28 Lead users are individuals who find the trends early. Previous studies have
shown that lead users (1) have expertise in a specific area, (2) do not consider innovation as something that is complicated as others think, and (3) are opinion leaders
(Schreier, Oberhauser, and Prugl 2007).29
There are several research findings indicating that lead users are important at the
time of developing new products. For instance, Lilien, Morrison, Searls, Sonnack,
and von Hippel (2002) took the new product development at the 3 M Company
as an example to explain why lead users should be involved at the stage of new
product development.30 According to them, the sales of a product created based on
the opinions of lead users were shown to be eight times higher than a product created
by the regular method of surveying users at 3 M. They noted it was because the
ability of regular users to create breakthrough is weaker compared to that of lead
users. Hamaoka (2004) discusses the benefit of involving consumers in new product
development.31 Noting that consumers have the ability to not only choose from
options given by the company but also to develop and create themselves, if that is not
enough, he states consumers could develop more complex and sophisticated products
and have a huge impact on companies if they are connected to companies. Having
made a theoretical discussion, Von Hippel, Ogawa, and Jeroen (2011) performed
an international comparison on the usefulness of lead users.32 According to this,
6.1%, 5.2%, and 3.7% of the population are lead users in the UK, the US, and
Japan, respectively, and the estimated amount they invest in product innovation is a
staggering USD5.2 billion, USD20.2 billion, and USD5.8 billion in the UK, the US,
and Japan, respectively. However, the percentage of those innovations being actually
turned into products is low; it is only 17% in the UK, 6% in the US, and 5% in Japan.
Therefore, it is been concluded that the area still has a large potential.
As listed above, the main concepts for understanding trendsetting consumers
include innovators, opinion leaders, market mavens, and lead users, and we can
analogically infer that these concepts are closely intertwined rather than being independent of each other. We now have a look at the studies that compared those concepts
in order to clarify the characteristics of each concept.
Goldsmith, Flynn, and Goldsmith (2003) used data to verify the difference
between innovators and market mavens.33 The similarity is that the variables to identify innovators and market mavens are highly correlated, and they are both opinion
leaders; in terms of the difference, while innovators focus more on new products,
market mavens learn and communicate general information on the market without
being limited to new products. Ikeda (2008) explained the differences and similarities
between opinion leaders and market mavens with actual survey data.34 The results
150
8 Studies on Connoisseurs
showed that while market mavens use traditional media, online media, and interpersonal information more and have a higher number of email communication and email
newsletter genres viewed, opinion leaders have more knowledge on particular products and are more motivated to share information. This suggests that opinion leaders
are likely to possess and share narrow but deep knowledge. Yamamoto (2008) showed
the difference between innovators, opinion leaders, and market maven while sorting
key individuals in word-of-mouth communication.35 According to this, while opinion
leaders are individuals specializing in one product category, market mavens are versatile, cross-category opinion leaders. They are consumers who possess various sides
of information such as multiple products, retailers, and markets, and are capable of
directing their own conversation and answering other consumers’ questions on the
market. This also indicates that although innovators are key individuals in the sense
that they adopt (purchase) before other people, they are less influential because they
are separated from average consumers in social systems.
Based on the above, the four concepts are summarized as follows: In terms of
innovators, the focus of research includes how early the adoption timing is, and
whether they are interested in innovative products; the concept does not consider
how influential they are. In contrast, while the social position of opinion leaders
becomes extremely important, since they are very influential, they are opinion leaders
only for a particular product category because not many of them are versatile across
product categories. Market mavens are similar to opinion leaders. However, they can
be regarded as the most influential trendsetters because they cover a wide range of
information across multiple product categories, make judgments based on common
sense, and there are many of them. Lead users are individuals who are useful for
corporate product development but do not have much informational influence on
general consumers. However, since they are not intimidated by innovation, the ones
with a strong communication skill could influence general consumers when the new
products have strong technical elements.
As described, research on trendsetters has been accumulated in many previous
studies. Based on the results of these studies, we can say plenty of “connoisseur”
consumers, who can judge the right or wrong of new products, exist among general
consumers. Then again, it is probably difficult to tell them apart from ordinary people
since they are people incorporating the above four concepts intricately intertwined.
8.3 Mechanism of the Study on Connoisseurs
The study on connoisseurs shown in this chapter is a study I have been conducting
jointly with HAPIKEN since 2003 by setting up a panel of connoisseurs on the
Internet. The panel would read press releases on new products and judge whether or
not the products will sell, while we scored the judgments to forecast the actual sales
after the release. An overview follows.
First, the population of connoisseurs was created over one year by asking people
to predict the sales every time a new product was released and extracting those who
8.3 Mechanism of the Study on Connoisseurs
151
predicted correctly. Specifically, we first extracted people who were relatively good
at predicting the sales of new products released in 2003. Then, we had them predict
the success or failure of the new products released in 2004 and further qualified
consumers who had a high rate of correct predictions to create the panel of connoisseurs. The number of connoisseurs initially placed was 373, and for comparison, 922
non-connoisseurs were also surveyed each time. As for the press releases used by the
connoisseurs for judgment, the ones published in newspapers were reformatted and
used. Given that the press release of a large product is issued about a month and a half
before the product launch, in the case of beverages, we figured it would be possible
to forecast the demand before the release if judgments can be made only with that
information. Based on the reformatted press release, the panel of connoisseurs would
make assessments on sales, such as how well the new product would sell and how
well it would sell compared to the competing products and the product itself. We
clarified the reasons why it sold or did not sell and created a system that can provide
feedback to new products development. We also scored the product appeal based on
connoisseurs’ sales assessment because we needed to ultimately forecast the sales.
Furthermore, we conducted a survey one month after the new product was released
to see if the connoisseurs actually made a trial or repeat purchase and whether their
product assessments have changed in order to doublecheck if their judgments are
consistent. Since the connoisseur panel is data with no purchase history, this is to
show the number of trial and repeat purchasers—which is considered important in
the aforementioned new product prediction model—and to verify the usefulness of
the study on connoisseurs by asking about trial and repeat purchases.
As is clear from this system, the connoisseurs, extracted in this survey, are those
who can judge whether new products will sell or not; since they are there only for their
ability as critics, they are not required to purchase. Consequently, the connoisseurs
are different from the innovators who adopt early, as shown in previous studies. It is
a concept similar to opinion leaders because they look only at a particular product
category. However, since the need to accurately judge good or bad is emphasized, that
part is different from opinion leaders. While the way it is used is close to lead users,
in a sense, it can be used in new product development, because connoisseurs are not
required to have technical expertise; they do not need to know how a complicated
product works because they make assessments only as users. Because they judge
whether it would sell rather than whether they themselves would want it, they need
the ability to judge not only by their own assessment but also by guessing other
people’s assessments. In that sense, the concept of connoisseurs is closest to market
mavens who gather a wide range of information. Then again, it is not that they make
judgment on many product categories. In other words, we can regard the connoisseurs
used in this study as a new type of trendsetter that cannot be considered as completely
the same as one of those concepts but exists across the previous trendsetter segments,
even though they possess the elements of trendsetters shown in the existing previous
studies.
In what follows, I will explain forecasting that uses connoisseurs and their profile.
152
8 Studies on Connoisseurs
8.4 Analysis Results
First, I will explain the predictive power of the connoisseurs extracted in the aforementioned way. Figure 8.2 shows the six-point scale ratings based only on the press
release information prior to the product release, in accordance with the procedure
described above, as to whether a beer-based beverage (which sold extremely well)
would become a hit. The cumulative percentages of responses are shown by connoisseurs versus non-connoisseurs. Here, while 53.4% of the connoisseurs said “it will
become a big hit” or “it will become a hit,” the cumulative percentage for these two
responses among the non-connoisseurs is only 33%. This implies that the samples
of ordinary, non-connoisseur individuals could not foresee such a hit, even though it
was a product that could potentially be a huge hit. Figure 8.3 shows a similar graph
for a product that did not become a hit at all. While the cumulative percentage for
“it will become a big hit,” “it will become a hit,” and “it will become somewhat
of a hit” reaches 50% among non-connoisseurs, those three only add up to 22.4%
among connoisseurs. While the results of the survey among the sample of ordinary
120
100
96.8
100
98
96.4
84.5
100
80
78.2
60
53.4
40
33.3
20
8.6
4
0
it will become a it will become a it will become
it will not
big hit
hit
somewhat of a become much
hit
of a hit
connoisseurs
Fig. 8.2 Predictive power of connoisseurs (1)
it will not
it will not
become a hit become a hit at
all
non-connoisseurs
8.4 Analysis Results
153
120
97.5
100
100
90.7
90.3
77.4
80
60
50.1
40
22.4
15.7
20
0.3
5.7
0.7
0
it will become a it will become a it will become
it will not
big hit
hit
somewhat of a become much
hit
of a hit
connoisseurs
it will not
it will not
become a hit become a hit at
all
non-connoisseurs
Fig. 8.3 Predictive power of connoisseurs (2)
people make it seem the product might sell, if creative promotions and advertisement strategies are implemented after the release, we can see from the assessment
by connoisseurs that this is a product that will not sell at all. In fact, the product
in question was removed from the shelves in convenience stores in less than three
months; therefore, we can analogically infer that sales forecasts by connoisseurs are
useful.
Figure 8.4 goes further by scoring the responses of connoisseurs to calculate points
for each brand and show the relationship between these points and the actual sales
indicator in a certain period after the release. The horizontal axis shows the points
related to the sales forecast by connoisseurs, while the vertical axis shows the actual
figures of sales indicators. We can see from here that the two values are related; the
correlation coefficient reaches 0.91 when calculated. This result indicates that sales
can be forecast for a given product with considerable accuracy before the release if
we can successfully extract and survey connoisseurs.
If so, what do connoisseurs look at to judge whether the product will sell or
not? When we conducted a factor analysis by using the statements connoisseurs
rated regarding new products, some factors were consistently extracted from all of
154
8 Studies on Connoisseurs
Sales volume
Fig. 8.4 Connoisseurs’s evaluation score and real sales volume
the new products surveyed. When we explored elements that affect the marketable
factor among them, we found that novelty and great taste had a great impact in all new
products surveyed. Furthermore, novelty and great taste were affected by rating on
the appearance and rating on the content. Based on this, we hypothesized a structure
in which connoisseurs rate the appearance of the product package and the content
of the product from the picture and statements in the press release, and if they can
sense novelty and great taste there, they would judge the product marketable. Then,
the structure was subjected to a structural equation modeling (SEM) and confirmed.
Figure 8.5 shows the structure for Brand X which sold well, while Fig. 8.6 shows
the structure for Brand Y which did not sell. As a note, since the notations get
complicated, the error terms corresponding to the measured variables of the model
and the covariance between these error terms are omitted.
First, with regard to Brand X that sold well, all indicators representing the goodness of fit to our model, including the GFI, AGFI, and RMSEA, had good values of
0.961, 0.948, and 0.043, respectively. The GFI and AGFI, which indicate the fitness
of the model, are both above 0.9, and the RMSEA is within the acceptable range
of 0.1 or less, so all coefficients in the model being statistically significant. Given
that the goodness of fit of the model was good, we can infer that brands that sell are
the ones with which the ratings on appearance and content evoke novelty and great
taste. In other words, connoisseurs sense novelty and great taste from the picture and
contents of the press release and judge whether it will sell based on that. We can
say novelty and great taste affect how the product will sell in the case of beer-based
beverages. To speak of the Brand X surveyed, we can see that while all variables are
significant for the brand, the path of rating on the appearance driving great taste and
leading to the judgment on being marketable is extremely strong. Although novelty
affects the judgment on being marketable, the standardized coefficient shows that
the effect of the path from great taste has nearly four times the strength. Based on
8.4 Analysis Results
155
novelty 3
novelty 1
0.39
0.42
novelty 4
novelty 2
0.54
0.40
Evaluation scale1
0.36
novelty
0.56
Evaluation scale2
0.20
Rating of
appearance
0.66
0.56
marketable
0.49
0.75
0.55
0.36
appearance 2
appearance 3
0.29
0.49
Evaluation scale3
appearance1
0.42
0.78
0.34
great
taste
Evaluation scale4
Rating on
the content
0.44
0.49
Taste 4
0.62
Taste 2
Content 2
0.33
0.27
0.45
Taste 1
Content 1
0.49
Content 3
GFI:0.961
AGFI:0.948
RMSEA:0.043
Taste 3
Fig. 8.5 The structure for Brand X
novelty 3
novelty 1
0.07
0.57
novelty 4
novelty 2
0.41
0.03
Evaluation scale1
0.95
novelty
0.59
Evaluation scale2
2.11
marketable
0.44
-1.33
Taste 2
0.43
Taste 4
0.77
Taste 3
appearance 2
appearance 3
Content 1
Rating on
the content
Content 2
0.66
0.59
0.51
Taste 1
0.98
0.42
great
taste
Evaluation scale4
0.41
0.50
0.13
0.86
0.54
Rating of
appearance
0.38
0.39
Evaluation scale3
appearance1
0.18
0.46
Content 3
GFI:0.861
AGFI:0.800
RMSEA:0.091
Fig. 8.6 The structure for Brand Y
this, we can determine Brand X sold well because it had an extremely high rating
on the appearance.
On the other hand, in the case of Brand Y that did not sell, the GFI, AGFI, and
RMSEA indicators to show the goodness of fit of the model were low at 0.851, 0.800,
and 0.091, respectively, and many of the coefficients to indicate the relationship of
each variable were not significant (“broken line” denotes coefficients that are not
significant). It implies that great taste and novelty could not evoke the sense of being
156
8 Studies on Connoisseurs
marketable based on the press release. In other words, this new product was deemed
not to have the kind of great taste and novelty necessary to sell. Based on the past
studies on knowledge categories, when a new product is introduced to the market,
consumers refer to their own past experience and show interest in the product if there
is a moderate inconsistency, but show no interest and react only at a bargain price
if it is completely consistent.36 We can say that the new product in question was, in
fact, deemed not to have any more novelty and great taste than the existing products.
Next, I will show the results of the post-release survey. Figures 8.7 and 8.8 are
diagrams prepared by combining the intent for trial purchase in the pre-release survey
and the actual purchase identified in the post-release survey for the product that sold
and the product that did not. On comparing Figs. 8.7 and 8.8, the difference between
the product that sold and the product that did not sell is found to be remarkable.
First of all, the percentage of those who intended to try was already different in the
pre-release survey; while 52.9% intended to try the product that became successful,
it was only 34.6% for the product that did not sell. Of those, the percentage of
individuals who actually made a trial purchase was 49.4% and 36.4%, respectively,
showing a difference here as well. There were individuals who made a trial purchase
even though they said “neither” or “don’t want to purchase” in the pre-release survey.
However, while the percentage for the successful new product was 27.0% and 15.3%,
respectively, it was only 19.0% and 11.7%, respectively, for the unsuccessful product.
Fig. 8.7 Matching with pre-release survey and post-release survey (1)
8.4 Analysis Results
157
’
’
’
’
Fig. 8.8 Matching with pre-release survey and post-release survey (2)
Moreover, looking at the number of repeaters, while the overall percentage of those
who indicated the intention to opt for a trial purchase and indicated a repeat purchase
was 18% and those who indicated the intention to opt for a trial purchase but indicated
no intention to make a repeat purchase was 3.2% for the product that sold well, the
percentages were 7.5% and 2.0%, respectively, for the product that did not sell.
I have mentioned earlier that the Perfit & Collins model used for forecasting the
demand for new products forecasts the demand based on the number of trial and
repeat purchasers after the release, which is highly accurate. The results here, which
showed higher trial purchases and repeat intentions, are also consistent with their
claim. Since products with strong trial purchase intention before the release are strong
in actual trial purchase and repeat purchase intention, it validates the way the demand
forecast is performed based on the results of the pre-release survey.
As described above, product appeal can be measured with considerable accuracy when a pre-release survey is conducted among connoisseurs. The results of the
covariance structure analysis also provided hints about what should be emphasized
in product development. The assessments of connoisseurs are also reasonable when
looking at the concept of trial and repeat that is central to the existing demand forecasting models for new products. It can be said this is a very useful technique for
forecasting demand for new products.
158
8 Studies on Connoisseurs
In that case, what kind of people are connoisseurs who are useful in this way? In
what follows, I will look at the above 373 individuals and explore how they differ
from ordinary people, as well as how they differ among themselves.
8.5 Profile of Connoisseurs
First, I explored the difference between connoisseurs and other survey subjects in the
five major aspects: (1) demographic factors, (2) level of media contact, (3) level of
drinking, (4) lifestyle elements related to food, clothing, and housing. It is said that
the factors affecting consumer’s decision-making are roughly divided into economic
factors, social factors, and psychological factors (Shimizu 1999).37 I am going to
compare the above five elements because I felt it would allow me to include these
factors. I looked at a total of 1,295 individuals comprising 372 connoisseurs and 922
individuals who participated in the same survey but were not certified as connoisseurs.
As a note, only those aged 20 or older qualified because the surveyed product was
an alcoholic beverage.
First, there was no difference between connoisseurs and non-connoisseurs in
gender (χ2 (1) = 1.357, p < n.s.), marital status (χ2 (1) = 0.101, p < n.s.), age
group (χ2 (3) = 1.842, p < n.s.), income classification (χ2 (5) = 3.532, p < n.s.),
family configuration (χ2 (5) = 5.347, p < n.s.), and type of residence (χ2 (5) =
2.814, p < n.s.); these are demographic factors. It turned out to be a confirmation of
what I showed in Sect. 8.2, i.e., according to past studies, identifying trendsetters by
demographic factors is difficult. Likewise, no difference was shown in the time spent
on TV (χ2 (8) = 6.240, p < n.s.), radio (χ2 (8) = 1.891, p < n.s.), and the Internet
per day (χ2 (8) = 3.641, p < n.s.); this is “the level of media contact.”
Table 8.1 Level of drinking
Drinking level
Non-connoisseurs
Connoisseurs
Total
I hardly ever
drink
Drink a
little bit
Quite a bit of
drinking
Total
Frequency
192
492
289
973
Expected
frequency
191.6
504.6
276.7
973
Frequency
67
190
85
342
Expected
frequency
67.4
177.4
97.3
342
Frequency
259
682
374
1315
Expected
frequency
259
682
374
1315
8.5 Profile of Connoisseurs
159
Next, looking at the level of drinking, no difference was shown here either (See
Table 8.1: χ2 (2) = 3.309, p < n.s.). The frequency of drinking was expected to be
related because they are connoisseurs of alcohol, but that was not the case.
The items on lifestyle composed of 28 clothing items, 15 food items, and
24 housing items. When the items that differ between connoisseurs and nonconnoisseurs were explored in each, no item was differ between connoisseurs and
non-connoisseurs.
Based on the above, it became clear that there is almost no item that can successfully distinguish connoisseurs from non-connoisseurs. Despite the subject products
being beer and low-malt beer, whether the individual drinks on regular basis had
almost no discriminating power. Demographic factors, lifestyle, and involvement in
alcohol also showed almost no difference. This result indicates it is not possible to
distinguish connoisseurs from ordinary people. However, considering the discussion
shown in the previous Sect. 8.2 of this chapter, we can view it as not being able to
distinguish connoisseurs because there are various kinds of connoisseurs, rather than
not being able to distinguish connoisseurs from ordinary people. In other words,
I can think that it is because trendsetters are a mix of four—innovators, opinion
leaders, market mavens, and lead users—and these types of consumers are included
in connoisseurs. Therefore, I attempted to understand connoisseurs by dividing them
into these types. However, it is quite possible that individuals with innovator-like
elements also have market maven-like elements, since these concepts overlap. For
this reason, it is difficult to identify them by methods, such as regular cluster analysis, which force to categorize the data into groups. Therefore, I performed a pseudo
cluster analysis by using lifestyle variables used for discriminating between connoisseurs and non-connoisseurs, identified variables with strong discriminative power by
creating clusters, and prepared several overlapping groups characterized by whether
there was a reaction to those variables. As a result, three variables comprising “I
often purchase new food products,” “I eat three square meals,” and “I often look at
information on fashion” had high discriminative power among the lifestyle items.
The classification of connoisseurs was done based on the number of yeses to the
above three highly discriminatory variables. Figure 8.9 shows the size of each group
and how they overlap. Circled numbers in the figure denote the segment numbers.
There were 71 connoisseurs (Segment 8) who did not say “yes” to any of the above
three variables.
I have looked at these eight segments prepared here and explored their differences
again using (1) demographic factors, (2) level of media contact, (3) level of drinking,
just as I did earlier.
First, with regard to (1) demographic factors, no apparent differences emerged
in these eight segments: gender (χ2 (7) = 9.547, p < n.s.), marital status (χ2 (7) =
5.927, p < n.s.), age group (χ2 (21) = 23.637, p < n.s.), income classification (χ2
(35) = 35.911, p < n.s.), family configuration (χ2 (35) = 41.172, p < n.s.), and type
of residence (χ2 (35) = 33.625, p < n.s.). For (2) the level of media contact, the
time spent difference on TV between connoisseurs and non-connoisseurs becomes
significantly different (χ2 (56) = 82.441, p < 0.012), but for radio (χ2 (56) = 72.597,
p < n.s.), and for Internet per day (χ2 (49) = 62.877, p < n.s.) were not become
160
8 Studies on Connoisseurs
Fig. 8.9 Segments of
connoisseurs
Fashion information
74
31
26
23
75
36
46
Three square meals
New food product
significant. Given that previous studies showed no difference in demographic factors,
this result is reasonable.
Next, in terms of (3) frequency of drinking (level of drinking), it became clear that
those in Segment 8 drink significantly more frequently than those in other segments
(see Table 8.2: χ2 (14) = 54.721, p < 0.001). We can say that those in Segment
8, which could not be identified by the above lifestyle variables, are connoisseurs
because of their frequency of drinking. In contrast, the percentage of drinkers is low
in Segments 5 and 7. The individuals in these two segments are highly interested in
new food products, and they are the kind of people who have three square meals;
these segments have nothing to do with the frequency of drinking.
As described above, although they are summed up by one word “connoisseurs,”
they have clearly become connoisseurs for various reasons. They are probably
capable of rating products from various perspectives, because connoisseurs consist
of various types of people and that is probably why they can guess correctly.
8.6 Conclusion and Future Implication
This chapter proposed a new demand forecast model that judges the success/failure
of a given product at an earlier point in time—specifically, at the time of press
release before the product launch—by surveying information-savvy consumers
(connoisseurs). Since previous chapters have made it clear that there is a difference in information-savviness among consumers and shown that brands favored by
information-savvy consumers are strong, these concepts were incorporated in the
model. Previous studies on connoisseurs include “innovators,” “opinion leaders,”
“market mavens,” and “lead users.” While each has many findings, the concept of
“consumers who can assess products as critics,” considered in this chapter, is a
8.6 Conclusion and Future Implication
161
Table 8.2 Level of drinking between 8 segments
Drinking level
Segment 1
Segment 2
Segment 3
Segment 4
Segment 5
Segment 6
Segment 7
Segment 8
Total
I hardly ever
drink
Drink a little bit
Quite a bit of
drinking
Total
Frequency
4
21
6
31
Expected
frequency
6.1
17.2
7.7
31
Frequency
9
43
23
75
Expected
frequency
14.7
41.7
18.6
75
Frequency
6
23
7
36
Expected
frequency
7.1
20
8.9
36
Frequency
7
17
7
31
Expected
frequency
6.1
17.2
7.7
31
Frequency
17
23
6
46
Expected
frequency
9
25.6
11.4
46
Frequency
8
12
6
26
Expected
frequency
5.1
14.4
6.5
26
Frequency
13
9
1
23
Expected
frequency
4.5
12.8
5.7
23
Frequency
3
42
29
74
Expected
frequency
14.5
41.1
18.4
74
Frequency
67
190
85
342
Expected
frequency
67
190
85
342
concept that encompasses those four past concepts, suggesting that there are various
types of consumers among connoisseurs.
I have actually set up a panel of connoisseurs on the Internet for forecasting
the demand of beers. The panel of connoisseurs was created by having consumers
predict the sales of beer every time one was released over one year and retaining
the consumers who predicted with high accuracy. They were asked to forecast the
post-release sales of a new product by reading the press release for the new product. It
showed that the correlation between the average product scores predicted by connoisseur monitors and the actual sales indicators was 0.91, proving that the accuracy
was extremely high. With the survey on connoisseurs, I not only forecast the sales
but also investigated why connoisseurs predicted the product would sell. When the
162
8 Studies on Connoisseurs
factors that are consistently extracted were identified from the results and subjected
to covariance structure analysis to examine the paths through which they concluded
that the product would sell well, two factors—novelty and great taste—were noted to
be important in the product predicted to sell well; moreover, novelty and great taste
were driven by rating on appearance and rating on contents. Furthermore, it was
found after the release that while the paths between these factors became significant,
and the model was identified in the case of products that sold well, the relationships
between the factors did not become significant. The path leading to marketable could
not be determined by connoisseur consumers in the case of products that did not sell.
Although not much information can be obtained from the press releases, it was shown
that connoisseur consumers have the ability to properly recognize these factors.
Lastly, I showed what kind of profiles these connoisseurs actually have. Based
on this, no axis was found to distinguish connoisseurs from non-connoisseurs,
suggesting that there are a variety of consumers, even when they are grouped by
one word, “connoisseurs.” Therefore, I put connoisseurs through a cluster analysis
to categorize them into several types. When doing so, since multiple concepts might
be overlapping, as indicated by the previous studies, I conducted a pseudo cluster
analysis by using the lifestyle variables used for separating connoisseurs from nonconnoisseurs, rather than completely dividing them into groups. Then, I identified
highly discriminatory variables in clustering and created several overlapping groups,
based on whether there was a reaction to those variables. It revealed that connoisseurs
are roughly divided into seven groups, comprising the group that became connoisseurs because they drink beer frequently, the group that became connoisseurs despite
not drinking beer much because they are information-savvy, the group that became
connoisseurs because they like to eat out, the group that can judge beers in the market
because they actually are more likely to eat at home, and so on.
In general, conventional demand forecasts for new products often conduct a large
market testing prior to the product release or predict the post-release sales, based on
the trial and repeat purchases after release. However, it is quite difficult to measure
product appeal based on a large survey or trial and repeat purchases when many new
products are released, as many retailers are armed with information and use the POS
data to remove the products that do not sell well within only four weeks from the
release. In a way, the retail industry, being armed with information, is driving the
manufacturers into a corner. However, the spread of the Internet has increased the
number of general consumers who develop knowledge by collecting information on
the Internet and trying many new products. This study revealed that connoisseurs
who can predict how well a product will sell exist among these consumers, and
the demand can be forecast at an early stage by using them. While conventional
quantitative analyses treated consumers as the same person, and the trial and repeat
model judged the success/failure of a new product, based on the number of people
who made a trial purchase or repeat purchase, the results in this chapter demonstrated
it is not sufficient to merely understand the number of people; rather, it is important
to also consider the quality of these people. The study also showed the possibility
8.6 Conclusion and Future Implication
163
of a new marketing method that uses information disparity among consumers due to
increased information use.
Notes
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Wind, Yoram; Mahajan, Vijay; Cardozo, Richard N. (1981), ‘New-Product
Forecasting’, Lexington Books.
Katahira, Hotaka. (1983), Marketing Science, The University of Tokyo Press.
(written in Japanese).
Nakamura, Hiroshi. (2001), Marketing of New Products, Chuokeizai-sha, Inc.
Blattberg, Robert; C, Golanty J. (1978), TRACKER; An Early Test Market
Forecasting and Diagnostic Model for New Product Planning’, Journal of
Marketing Research, Vol. 15, pp. 192–202. As a note, NEWS is also wellknown as another model that combined market testing with trial and repeat
purchases. For more information, see Pringle, L.G.; Wilson, R.D.; Brody, E.I.
(1982), ‘NEWS: A Decision-Oriented Model for New Product Analysis and
Forecasting’, Marketing Science, Vol. 1, pp. 1–29.
Parfitt, J.H.; Collins, J.K. (1968), ‘Use of Consumer Panels for Brand-Share
Prediction’, Journal of Marketing Research, Vol. 5, pp. 131–145.
Sugita, Yoshihiro; Hiroshi Nakamura; Hirokazu Tajima (1993), ‘A New
Product Trial Purchase Model That Uses Unaggregated Data’, Marketing
Science, Vol. 2, No. 1, pp. 32–45 (written in Japanese).
Robertson, Thomas S. (1967), ‘Consumer Innovators: The Key to New Product
Success’, California Management Review, pp. 23–30.
Bass, Frank M. (1969), ‘A New-Product Growth Model for Consumer
Durables’, Management Science, Vol. 15, pp. 215–227.
Pizam, Abraham (1972), ‘Psychological Characteristics of Innovators’, European Journal of Marketing, Vol. 6, pp. 203–210.
McDonald, Heath; Corkindale, David (2003), ‘Behavioral Versus Demographic: Predictors of Early Adoption: A Critical Analysis and Comparative
Test’, Journal of Marketing, Summer, Theory and Practice.
Taylor, James W. (1977), ‘A Striking Characteristic of Innovators’, Journal of
Marketing Research, Vol. 16, pp. 104–107.
Venkatraman, Meera P. (1991), ‘The Impact of Innovativeness and Innovation
Type on Adoption’, Journal of Retailing, Vol. 67, pp. 51–67.
Im, Subin; Bayus, Barry L.; Mason Charlotte H. (2003), ‘An Empirical Study of
Innate Consumer Innovativeness, Personal Characteristics, and New-Product
Adoption Behavior’, Journal of Academy of Marketing Science, Vol. 31,
pp. 61–73.
Cowart, Kelly O.; Fox, Gavin L.; Wilson, Andrew E. (2008), ‘A Structural Look
at Consumer Innovativeness and Self-Congruence in New Product Purchases’,
Psychology & Marketing, Vol. 25, pp. 1111–1130.
Myers, James H.; Robertson, Thomas S. (1972), ‘Dimensions of Opinion
Leadership’, Journal of Marketing Research, Vol. 9, pp. 41–46.
164
8 Studies on Connoisseurs
16.
Summers, John N. (1970), ‘The Identity of Women’s Clothing Fashion Opinion
Leaders’, Journal of Marketing Research, Vol. 7, pp. 178–185.
King, Charles W.; Summers, John O. (1970), ‘Overlap of Opinion Leadership
Across Consumer Product Categories’, Journal of Marketing Research, Vol.
7, pp. 43–50.
Yavas Ugur; Riecken Glen (1982), ‘Extensions of King and Summers’ Opinion
Leadership Scale: A Reliability Study’, Journal of Marketing Research, Vol.
19, pp. 154–155.
Riecken Glen (1983), ‘Internal Consistency Reliability of King and Summers’
Opinion Leadership Scale: Further Evidence’, Journal of Marketing Research,
Vol. 20, pp. 325–326.
Childers,Terry L. (1986), ‘Assessment of the Psychometric Properties of an
Opinion Leadership Scale’, Journal of Marketing Research, Vol. 23, pp. 184–
188.
Leonard-Barton, Dorothy (1985), ‘Experts as Negative Opinion Leaders in the
Diffusion of a Technological Innovation’, Journal of Consumer Research, Vol.
11, pp. 914–926.
Aoike, Shinichi. (2002), ‘An Examination on Information-Sharing Behavior
of Opinion Leaders’, Bulletin of Nikkei Advertising Research Institute, Vol.
204, pp. 2–8. (written in Japanese).
Shibuya, Satoru. (2002), ‘An Examination on the Process of INFLUENCE
BETWEEN INDIVIDUALS: A View on Process Approach on Opinion
Leadership’, Keio Business Forum, Vol. 20, No. 1, pp. 1–18. (written in
Japanese).
Feick, Lawrence L.; Price, Linda L. (1987), ‘ The Market Maven: A Diffuser
of Marketplace Information’, Journal of Marketing, Vol. 51, pp. 83–97.
Clark, Ronald A.; Goldsmith, Ronald E. (2005), ‘Market Mavens: Psychological Influences’, Psychology & Marketing, Vol. 22, pp. 289–312.
Price, Linda L.; Feick, Lawrence R.; Guskey-Federouch, Audrey (1987),
Couponing Behavior of the Market Maven: Profile of a Super Couponer,
Advances in Consumer Research, Vol. 15, pp. 354–359.
Elliott, Michael T.; Warfield Ann E. (1993), ‘Do Market Mavens Categorize
Brands Differently?’, Advances in Consumer Research, Vol. 20, pp. 202–208.
von Hippel, E. (1986), ‘Lead Users: A Source of Novel Product Concepts’,
Management Science, Vol. 32, pp. 791–806.
Schreier, Martin; Oberhauser, Stefan; Prugl, Reinhard (2007), ‘Lead Users
and the Adoption and Diffusion of New Products: Insights from Two Extreme
Sports Communities’, Marketing Letters, Vol. 18, pp. 15–30.
Lilien, Gary L; Morrison, Pamela D.; Searls, Kathleen; Sonnack, Mary;
von Hippel, E. (2002), Performance Assessment of the Lead User IdeaGeneration Process for New Product Development’, Management Science,
Vol. 48, pp. 1042–1059.
Hamaoka, Yutaka. (2004), ‘Coevolutionary Marketing- New Concept of
Marketing under Age of Information Network-’, Mita Business Review, Vol.
47, No. 3, pp. 23–36. (written in Japanese).
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
8.6 Conclusion and Future Implication
32.
33.
34.
35.
36.
37.
165
Eric von Hippel; Susumu, Ogawa; Jeroen, P.J. De Jong (2011), ‘The Age of
the Consumer-Innovator’, MIT Sloan Management Review, Vol. 53, No. 1,
pp. 26–35.
Goldsmith, Ronald E; Flynn, Leisa R.; Goldsmith, Elizabeth B. (2003),
Innovative Consumers and Market Mavens, Journal of Marketing, Fall,
pp. 54–63.
Ikeda, Kenichi (2008), ‘Emergence of New Consumers: A Reexamination of
Adopter Category Factors’, ed. Miyata, Kakuko and Kenichi Ikeda. Internet
and Consumer Behavior -Research of word of mouth-’, NTT Publishing,
pp. 114–144. (written in Japanese).
Yamamoto, Hikaru. (2008), ‘Key Persons in Word-of-Mouth and How to Find
Them’, Marketing Journal, Vol. 107, pp. 125–131. (written in Japanese).
For more information, see Shimizu,Akira (2006) Strategic Consumer
Behavior, Chikura Publishing. (written in Japanese).
Shimizu, Akira (1999), New Consumer Behavior, Chikura Publishing.
Chapter 9
Brand Rating in the Age of Information
Sharing
Based on discussions thus far, it has become evident that the development of the
Internet has had a great impact on consumer behavior and marketing strategies
related to it. In particular, it has become clear that information plays an important role in generating innovative ideas for segmentation— which is a basis of
marketing strategy—just like traditional demographic and lifestyle factors. You have
to examine the process that includes post-purchase information sharing when considering consumers’ decision-making process. You also have to consider post-purchase
theories and data as well as theories that have had large impacts on brand rating,
and so on. As an illustration, I have shown examples of studies that utilized information as the basis of segmentation along with their brand ratings in Chaps. 6, 7,
and 8; changes in the decision-making process in Chaps. 2 and 5; and post-purchase
theories and realities of information-sharing in Chaps. 3 and 4.
In this chapter, I will create an online consumer panel to understand the abovementioned insights, using the panel to determine the validity of the above results obtained
from various subjects and timings. The objective is to verify the above findings in one
analysis and demonstrate their validity. The actual process included flagging online
panel members with “listener labels” such as “early listener”, “active listeners”, “passive listeners”, “uncertain listeners”, “disinterested listeners”, as shown in Chap. 7.
I have named them “the Kikimimi (it means active listener) panel,” and verified the
validity by surveying panel members on the activity of purchasing snacks. In what
follows, I will describe the relevant steps and results.
9.1 Setting of the Kikimimi Panel
First, let me explain how the “active listeners” panel is set up. The consumer panel
was created by surveying 50,000 individuals from a survey panel owned by the
online survey company MyVoice Communications, Inc. using the method shown in
Chap. 7. This method was based on the consumer study “CANVASS,” which has been
© Springer Nature Singapore Pte Ltd. 2021
A. Shimizu, New Consumer Behavior Theories from Japan, Advances in Japanese
Business and Economics 27, https://doi.org/10.1007/978-981-16-1127-8_9
167
168
9 Brand Rating in the Age of Information Sharing
conducted by Yomiko Advertising, Inc. every year for 1500 individuals, followed by
an extraction of the five segments. The characteristics of each segment, which were
based on lifestyle- related questions fielded at the same time, are as follows.
The forefront segment is “early listener”. They are very unique and strongly
motivated to express themselves; they are less likely to go along with the opinion
of others. Fashion magazines, the Internet, and mobile phones are essential for them
and they are highly informed. Around 3.4% of the overall population belongs to this
segment.
The second segment is “active listeners”-type of consumers. They are interested
in various topics such as family, hobbies, and health. They are also highly interested
in society, the environment, and community and tend to proactively absorb various
pieces of information. They differ from “early listener” in the way that they are
always thinking about their relation to society and community; they not only stay
informed, but also act by considering the balance with others around them. In terms of
past studies, this segment corresponds to market mavens. “active listeners” account
for 12.3% of the overall population.
The third segment, which has the largest sample size, is “passive listeners”type of consumers, accounting for 57.6% of the overall population. It is a very
average segment; these individuals tend to shift to “active listeners”—the segment
above them—or to “uncertain listeners”—the segment below them—based on
environmental changes, especially changes in the economy.
The fourth segment is the “uncertain listeners”-type of consumers. They do not
have any particular hobby or preference and tend to keep purchasing specific brands.
They do not focus on functions or services when consuming and do not have much
interest in fashion, home, interior, or design. Without confidence in taste, they are
likely to choose cheaper items and purchase frozen foods and breads rather than fresh
foods. Around 8.5% of the overall population belongs to this segment.
The last, fifth segment is the “disinterested listeners”-type of consumers. This is a
segment of people who are not interested in any area—not just brands and consumption but also family, leisure, health, social relationships, etc. The difference from the
aforementioned “uncertain listeners” is that while “disinterested listeners” behave
without any particular intention, “uncertain listeners” maintain strange preferences
and will not listen to other people’s opinion. Around 18.2% of the overall population
falls under this segment.
If we compare the percentage breakdowns of the subjects in the CANVASS
survey, which is discussed in Chap. 7, the subjects of the Distribution Economic
Institute of Japan (DEIJ) panel, and the panel created by MyVoice Communications, the last contained a slightly higher percentage of “disinterested listeners”
(CANVASS: 14.5% and DEIJ: 8.8% vs. MyVoice Communications: 18.2%) and
a lower percentage of “uncertain listeners” (CANVASS: 9.6% and DEIJ: 10.4%
vs. MyVoice Communications: 8.5%). However, percentages of other segments fall
between that of CANVASS and DEIJ (CANVASS: 4.2% and DEIJ: 1.6% vs. MyVoice
Communications: 3.4% for “early listener”; CANVASS: 8.3% and DEIJ: 26.7% vs.
MyVoice Communications: 12.3% for “active listeners”; and CANVASS: 63.4%
and DEIJ: 52.4% vs. MyVoice Communications: 57.6% for “passive listeners”).
9.1 Setting of the Kikimimi Panel
169
The percentage of “active listeners” under DEIJ is extremely high probably because
the institute’s panel consists of users who are loyal to a particular store and who
regularly experiment with new products and promotion, which makes them quite
well-informed. The characteristics of segments created by MyVoice Communications and applied in this study were almost the same as the percentage breakdown
of segments created in CANVASS. However, a slight difference was visible due to
the difference in the data collection method, i.e., hand delivery of self-administered
questionnaires versus online survey panel. In what follows, I will carry out an analysis
from the perspective of brand rating and verify the effectiveness of various consumer
behavior theories that have been shown in this book using separate data sets.
First, I will verify if brand power can be measured for each brand using the
percentage breakdowns of “uncertain listeners” and “active listeners”. The objective
is to demonstrate that the percentage of “active listeners” increases for products that
have strong brand power, while the percentage of “uncertain listeners” increases for
products that have weak brand power.
Next, as shown in Chap. 4, I will demonstrate that the level of affective commitment affects the power of a given brand. In this chapter, I will add calculative commitment, by which one makes a purchase based only on the discount available—which
is an even more subtle reason compared to calculative commitment—and category
commitment, by which one is familiar with a given product category, to affective
commitment, calculative commitment, and fascination commitment; ask questions
related to these concepts (Inoue 2009)1 ; and explore the relationship between brand
power and each type of commitment.
Then, as a potential and new comprehensive decision-making process, I will show
by following the AISAS® theory that there are consumers who share information.
In particular, I will explore the extent to which consumers reaching each stage of
AISAS® differs by how informed the individuals in the created panel are. As shown
in Chap. 7, “active listeners” should be more likely to move on to the last stage of
information sharing compared to other segments. I will verify to see if this tendency
is stronger for products that have stronger brand power.
Finally, I will measure a brand’s strength, which develops because of the difference
in decision-making routes described in Chap. 5. Note that, in Chap. 5, we have proven
that a product selected via the central decision-making route has stronger brand
power compared to a product selected via the peripheral decision-making route. I
will demonstrate this concept in this chapter too.
After measuring a brand’s strength by following the above-mentioned flow, I asked
identical questions to the same respondents at a different time period to examine
whether those products had brand power too. Online surveys are quite effective
because of their ability to survey the same participants at different times. Note that
brand power was measured based on loyalty; moreover, because there were three
brands that were studied in the first survey, line extension brands were observed
before the second survey. I will also conduct an analysis to verify whether factors
for rating line extension products differ by brand power. This was done because I
believe it will enable us to explain relationships between consumer behavior theories
and brand rating.
170
9 Brand Rating in the Age of Information Sharing
9.2 Brand Rating Using the Kikimimi Panel
This study focuses on the snack category. In Japan, the snack market is one in
which various manufacturers ranging in size from large and medium to small, sell
a large number of products. Therefore, depending on how the scope of a snack is
defined, the method for measuring market share also varies. More specifically, the
snack category contains multiple sub categories such as chocolate confectionery,
salty snacks, cookies, rice crackers, chewing gums, and candies. Furthermore, given
that the sub categories of snacks in recent years has seen many new snacks that
stretch across multiple sub categories (e.g., salty snack covered with chocolate), it
has become difficult to clearly draw lines. Thus, the definition of snacks differs by
manufacturer and analyst. In this study, from these diverse snack products, I selected
15 brands under the chocolate and cookie categories for the study.
The first survey was conducted in November 2010. The survey participants were
2000 women between 20 and 50 years of age, who purchased a product in the snack
category at least once a month. From this group, 1307 women, who also responded
to the second survey in March 2012, were included in the study. The breakdown was
159 “early listener”, 601 “active listeners”, 562 “passive listeners”, 239 “uncertain
listeners”, and 439 “disinterested listeners”. Compared to the overall percentage
breakdown of the “active listeners” panel, I included a larger percentage of “active
listeners” because I expected it would show more evident results as shown in the
study on communication-oriented consumers in Chap. 6. From these 15 brands, I
conducted a detailed analysis for six brands, that is, Brands R, Y, and L, which
introduced line extension brands the following year, and Brands A, B, and C, their
rivals.
First, by comparing the percentages of “passive listeners”, I examined what
percentages of “active listeners” and “uncertain listeners” account for segments that
favored each of the 15 brands such as these six brands. In particular, I subtracted
the percentage accounted by “passive listeners” from percentages accounted for by
“active listeners” and “uncertain listeners” to determine their positions, and showed
market shares by the size of the circle drawn around those positions in the center.
Figure 9.1 shows the results that were obtained. The objective is to use percentages
of “active listeners” and “uncertain listeners” to measure brand power, which cannot
be measured only by market share (just as we had seen in Chap. 7).
Thus, it was presumed that Brand B could be rated as an extremely strong brand
because it not only had a larger share but also a larger percentage of “active listeners”,
while Brands A and C had weaker brand power compared to Brand B because they
are favored more by “uncertain listeners”, although their shares are quite large. In
contrast, the potential brand power of Brand Y seems to be strong because it is
favored more by “active listeners”, although their share is not large.
Next, we measured the commitment for each brand. The objective was to verify
Chap. 4’s results that the brand power was strong when there was a large percentage
of people who purchased the brand with affective commitment, whereas it was weak
when more people with calculative commitment were observed among those who
9.2 Brand Rating Using the Kikimimi Panel
171
active listeners
Y
3.0
2.0
1.0
B
L
uncertain listeners
R
-10.0
-5.0
5.0
10.0
15.0
-1.0
A
-2.0
-3.0
C
Fig. 9.1 Who purchases? The difference between active listeners and uncertain listeners
favored the brand. In this chapter, commitment was measured on a five-point scale. In
the order of undesirable brand levels, the following five are included: (i) calculative
commitment, in which one is attracted by a discount and makes purchase decisions
accordingly; (ii) calculative commitment, which is driven by factors such as “I’m
afraid of making a mistake” and “it is too much trouble to switch the brand”; (iii)
affective commitment,, which is driven by emotional factors such as “I feel attachment or familiarity” and “it suits me”; (iv) category commitment, which is driven by
factors such as “I am a good judge for choosing this brand” and “I want to recommend
it to others” for knowing the product category in depth; and (v) fascination commitment, which is driven by factors such as “I would buy even if it was expensive” and
“I cannot think of any other product.” Figures 9.2 and 9.3 show this in terms of the
deviation value for each brand.
Thus, we can see that Brands Y and L scored high on the affective, category,
and fascination commitments; however, Brand R scored high on the fascination and
calculative commitments. Furthermore, some individuals make a purchase because
they love a certain brand and others do so because they do not wish to switch to
other brands. Therefore, when three brands that extended their product lines are
compared, the brand power of Brands Y and L as parent brands is stronger, while
the brand power of Brand R is average.
When we examine rival brands, we cannot say Brand B has strong brand power
compared to Brands Y and L since both discount and calculative commitments are
higher compared to the average, although the category commitment is also higher
172
Fig. 9.2 Deviation value for commitment
9 Brand Rating in the Age of Information Sharing
9.2 Brand Rating Using the Kikimimi Panel
173
1200
1000
800
600
400
200
0
recognion
understand the
content
Brand R
Brand Y
candidate for next
purchase
Brand L
Brand A
intenon to
recommend
Brand B
topic of
conversaon
Brand C
Fig. 9.3 Decision-making stage 6 brands
than the average. However, we can determine that the brand power of Brand A is
weak because both the discount and calculative commitments are higher compared
to the average and the remaining emotion-related commitments are below average.
Furthermore, Brand C seems to have very weak brand power because the scores for
affective, category, and fascination commitments are extremely low. Although the
share of Brand C is large, indicators for brand power are not high in terms of both
supporters and commitment.
Next, I asked questions about whether they recognized these six brands (recognition), whether the product’s characteristics were understood (understand the content),
whether they considered the product the next time they purchased (candidate for
next purchase), whether they wanted to recommend the product to others (intention
to recommend), and whether they spoke about the product (topic of conversation).
As shown in Chap. 2, the intention to recommend increases when the product is
purchased via recognition and interest. Moreover, the number of product searches
174
9 Brand Rating in the Age of Information Sharing
increases contributes to sales when there are multiple word-of-mouth communications. Thus, I have listed five items, similar to the ones described above, as
purchase decision-making processes and verified the scores for each indicator by
brand. Figure 9.3 shows this list.
Based on the figure, we can see that Brands A and B scored high on all items.
In particular, Brand A scored very high on “topic of conversation.” Even though
the supporter analysis and commitment scores were not so good, Brand A should
be quite strong in actual purchase setting because of its share. In contrast, although
Brand R scored high on “recognition” and “understand the content,” it falls at about
the same level as the other four brands for “candidate for next purchase,” “intention to
recommend,” and “topic of conversation.” Thus, while Brand R is highly recognized
because it is a brand that has a relatively long history, it does not necessarily have
brand power. As for Brand Y, all scores are low in this analysis because its share is
small, although it scored high in the analysis of supporters.
Figure 9.4 shows the results of the same analysis focusing only on the “active
listener” layer. From this chart, we can see that the percentage of people who
has intention to recommend is quite high, even if it is not in the candidate for
400
350
300
250
200
150
100
50
0
recognion
Brand R
understand the
content
Brand Y
candidate for next
purchase
Brand L
intenon to
recommend
Brand A
Fig. 9.4 Decision-making stage 6 brands (active listener only)
Brand B
topic of
conversaon
Brand C
9.2 Brand Rating Using the Kikimimi Panel
175
next purchase. In addition, the intention to recommend value of Brand B, which
is highly supported by “active listener”, is almost the same as Brand A. Although
the market share is inferior to that of Brand A, Brand B has a high level of “active
listener”support, which means that Brand B is as effective as Brand A in terms of
word of mouth.
Finally, I measured differences caused by decision-making routes. As described
in Chap. 5, when we assume a decision is made based on ELM, the information
and sources of information required by consumers differ between those who made
a decision via the central route and those who did so via the peripheral route. In
particular, consumers who made their decisions via the central route had a stronger
affective commitment and were less likely to switch brands compared to consumers
who made their decisions via the peripheral route. Furthermore, decisions are made
via the central route only when consumers are motivated to elaborate information and
have the ability to do so. Here, I have defined respondents who said “I’m very interested in snacks (25.6%)” and “I’m interested in snacks (42.9%)” as consumers that
have motivation. The respondents who checked “yes” for two or more statements out
of seven statements such as “I can tell the difference in the ingredients of snacks” and
“I know local specialty snacks in various locations very well” were knowledgeable
consumers (32.1%). As a result, 30.3% were consumers who have both the motive
and knowledge and take the central route, while 69.7% were consumers who made
decisions via the peripheral route. Figure 9.5 shows the percentage breakdown of the
central route and peripheral route calculated for each brand.
Brand R
28.3
Brand Y
71.7
39.7
Brand L
60.3
34.2
Brand A
65.8
26.4
Brand B
73.6
37.9
Brand C
62.1
28.4
Average
71.6
30.3
0
10
69.7
20
30
40
Central
50
60
70
Peripheral
Fig. 9.5 % breakdown between the central route and peripheral route
80
90
100
176
9 Brand Rating in the Age of Information Sharing
Table 9.1 Summary of each indicator
Market share
Purchaser
Commitment
Decision making
ELM
Brand R
×
×
Brand Y
×
〇
〇
×
〇
Brand L
〇
〇
Brand A
〇
×
×
〇
×
Brand B
〇
〇
〇
〇
Brand C
×
×
×
Here, note that while the percentages of people who purchase Brands Y, L, and B
via the central route are larger compared to the average, the percentages are smaller
for Brands R, A, and C. In fact, the percentage for Brand A is particularly small.
Considering these results, we can assume that Brand A is gaining market share
through promotions such as discounts. Furthermore, we can surmise that the potential
for Brand Y is large even though it scored low in the decision-making route analysis
performed earlier because of the smaller market share. As shown by the analysis here,
there are many people who purchased via the central route and many supporters are
“active listeners”, which is great.
Table 9.1 summarizes market shares of six brands and results for each diagnosis.
If we examine the level of similarity between indicators, the assessment results for
the sales share and purchase decision-making process turned out to be almost the
same. Moreover, the positioning and commitment of “active listeners” and “uncertain
listeners” and the results for assessing attitude formation in central and peripheral
routes were almost the same. While the former—assessments on market shares and
purchase decision-making process—can be regarded as indicators for the brand’s
popularity such as “how widely the brand has spread in the world” and “how often
people see it and become interested in it,” the latter assessments for “active listeners”
and “uncertain listeners” can be considered as indicators that are based on the strength
of attachment to brands, that is, “how emotional consumers are when purchasing the
brand.” Thus, it indicates that two concepts, “the width of brand” and “the depth of
brand,” are required to understand brand power.
Because these two indicators are completely different concepts, a case in which
the former’s assessment differs from the latter’s assessment inevitably occurs. For
example, Brand A was rated high for the former “width of brand,” although it was
rated low for the latter “depth of brand.” However, Brand Y was rated low for the
former “width of brand” and rated high for the latter “depth of brand.” Note that
Brands B and L scored well for both the width and depth of brand. In particular,
because Brand B scored extremely high for both, it can be considered a strong brand.
However, Brands R and C do not seem to have strong brand power because they do
not score high on either parameter, although some scores are partially quite high.
To determine whether these diagnostic results are correct, I compared the surveys
conducted in 2010 and 2012. The 2012 survey’s outline will be explained in the
next section; however, the subjects of this comparison are the sample population
9.2 Brand Rating Using the Kikimimi Panel
177
Table 9.2 Brand switch matrix
Brand R
Brand Y
Brand L
Brand A
Brand B
Brand C
Brand R
32.4
4.4
11.8
5.9
14.7
1.5
Brand Y
9.8
61
4.9
0
4.9
0
Brand L
1.4
5.1
59.4
1.4
5.8
1.4
Brand A
3.2
1.2
2.8
59.7
10.5
5.2
Brand B
1.9
0.6
5.1
7.7
60.9
3.2
Brand C
2.3
1.5
2.3
14.4
12.1
33.3
that responded to both surveys. In terms of a specific method of comparison, it was
performed by tabulating individual data (how the brand selected as “the brand you
want to purchase most” in 2010 changed in the 2012 survey) and creating a brand
switch matrix. The 2010 and 2012 survey results are demonstrated in rows and
columns, respectively (see Table 9.2). Note that the values of these six brands do not
add up to 100 because the tabulation was performed by including all the 15 brands
that were surveyed.
Based on this figure, we can see that while four brands, i.e., Brands Y, L, A, and
B, which have been measured to have brand power at least on one indicator out of the
two (the width of brand and depth of brand), have a loyalty percentage of about 60%,
Brands R and C, which were deemed to have weak brand power for both indicators,
have a loyalty percentage of merely 30%. Thus, while a brand can secure about 60%
of loyalty if it is recognized for one of the two types of indicators, the loyalty level
reduces by half to 30% if it does not do well on either indicator.
Furthermore, if we examine Brand B, which was highly rated on these two major
indicators, in addition to having a high loyalty rate, the switching rate from other
brands was also the highest among these six brands. In particular, customers came
in from Brands R, A, and C at the rate of >10% while leaving Brand B at the rate of
<10%. The brands that were strong for both the indicators—the width of brand and
the depth of brand—promoted inflow and prevented outflow of customers. This is
different from Brand A, which has the same 60% loyalty and strongly indicates the
meaningfulness of measuring feelings of consumers rather than merely measuring
the market share, brand recognition rate, and purchase experience.
Thus, based on the abovementioned data, it was demonstrated that brand assessment via the “active listeners” panel, which is a brand assessment based on consumer
behavior theories, was extremely effective. In particular, who the purchasers are, how
committed they are, and which route they took for attitude formation demonstrate
another aspect of brand strength that cannot be measured via market share or brand
recognition. In fact, it was made clear that customers leave when only the width
of brand is strong; however, they do not leave as much when both the indicators,
i.e., “the width of brand” and “the depth of brand,” are strong. This is consistent
with an extremely valid suggestion that supports the idea of CRM, which is more
advantageous for maintaining relations with existing customers.
178
9 Brand Rating in the Age of Information Sharing
9.3 Effect of Product Line Extension and Its Measurement
As described previously, the strength of each brand could be measured with two
indicators: the width of brand and the depth of brand. From these six brands, three
brands, i.e., Brands R, Y, and L, introduced product line extensions one year after
the survey. Among the three parent brands, Brand R had a loyalty rate of 30% and
scored high on both the indicators, namely, the width of brand and the depth of brand.
However, Brand Y scored high only on the width of brand indicator and had a loyalty
rate of 60%, while Brand L scored high on both the width of brand and depth of brand
indicators and had a loyalty rate of 60%. What type of difference would product line
extensions cause when parent brands have a different brand power?
Product line extension is very popular in the United States and Europe. For
example, according to Aaker’s 1991 book, 89% of new products launched in the
United States were product line extensions (Aaker 1991)2 and, according to Keller
(2008), 80–90% of new products were product line extensions.3 Thus, product line
extension has been used quite extensively as a new product strategy for a number of
decades.
A concept similar to product line extension is brand extension. In terms of research,
there are overwhelmingly many more studies on brand extension. The primary difference is that while brand extension indicates entering a product category that is
different from that of the parent brand, product line extension indicates entering
the same category as the parent brand. Thus, as success factors, previous studies on
brand extension first note that strong equity of the parent brand is a prerequisite and
list compatibility between the product categories of the new extension product with
that of the parent brand (Aaker and Keller 1990).4 Although there are multiple discussions as to what constitutes compatibility, similarities in product characteristics and
consistency in brand concepts are considered important (Park, Milberg, and Lawson
1991).5 This is because consumers would be able to easily process information on
the new extension products by using their own attitude and existing knowledge about
the parent brand, if they are highly consistent.
However, most studies on brand extension were experimental studies that used
imaginary brands. Volckner and Sattler (2007), who reviewed and attempted to generalize past papers on brand extension, stated that the results obtained from studies on
imaginary brands or supposed brand extensions cannot be generalized because study
participants were forced to make judgments with limited information such as the
brand name and product category of the extension product.6 Therefore, studies on
brand extension using actual brands are desirable; however, it is not very feasible
since you would have to conduct a joint study with a company and there would be
confidentiality obligations.
In addition to the strength of brand equity mentioned in brand extension studies,
studies on product line extension report the importance of considering cannibalization
of the parent brand (Caldieraro, Kao, and Cunha 2015),7 and the timing of product
release because it is an extension of the same product category (Wilson and Norton
1989).8 That said, many studies are experimental and assume imaginary extensions,
9.3 Effect of Product Line Extension and Its Measurement
179
similar to studies on brand extension; not many studies examine the phenomena
of real product line extension. The following studies examined real line extension
products and analyzed the factors that make a product line extension successful.
Reddy, Holak, and Subodh (1994) examined 75 brands of cigarettes and tried to
understand whether their product line extensions succeeded by considering cannibalization of their parent brands and using the pooled data.9 Note that market share
was used as a dependent variable and multiple regression analysis was used for
modeling. The results showed that market share was affected by the parent brand’s
strength (market share and tenure), the symbolic value of the parent brand, early
timing for market entry, company’s size, distinctive marketing competencies, and
advertising expenses. Moreover, it showed that cannibalization of the parent brand
does not affect the tobacco industry much and, even if the parent brand was cannibalized, a product line extension generated more value. Furthermore, this behavior
shows that the strength of the parent brand drives the product line extension’s success.
Sinapuelas, Clark, and Sisodiya (2010) used IRI’s data for 20 different product
categories to verify whether the product line extension brand affected the parent brand
and whether the equity of the parent brand affected the introduction of a product
line extension.10 Based on this study, it was shown that the number of product line
extensions and the profit generated from them increased when the equity of the parent
brand was strong. Moreover, a highly rated advertisement of a product line extension
was proven to positively affect the equity of its parent brand, and advertising the
extension brand alone would not be effective when the equity of the parent brand
is strong. Moreover, the equity of the parent brand was unaffected even when the
product line extension brand was highly innovative. Therefore, it became clear that
there is a solid relationship between the parent brand and the brand of its product
line extension. Also, while they are interdependent, technological innovation of each
brand of extension will not affect the equity of the parent brand unless something
unusual happens, because the parent brand’s equity is strong.
Hypothesizing that the attitude towards the parent brand and knowledge of the
parent brand also interact with product line extensions just as they did in brand
extension, Akamatsu conducted an analysis by considering two elements, namely,
factors related to the parent brand’s functions and factors related to the parent brand’s
relationship. Note that the sample size was 771 and they examined beer drinks that
actually existed. In terms of questions, they asked about past purchase history and
intention to purchase the brand of product line extension by asking questions such as
whether they have purchased the parent brand and whether they would purchase the
brand of new product extension. The results demonstrated that while consumers who
purchased the parent brand decided whether to purchase the product line extension
brand based on the strength of their relationship with the parent brand, consumers
who have not purchased the parent brand focus both on their relationship and the
functionalities of the new extension brand of the parent brand. Furthermore, this
study demonstrated that the purchase history of the parent brand affected the rating
of the product line extension brand.
Thus, from the above literature review, three points can be clarified: (1) strong
brand equity of the parent brand is essential for the extension brand’s success, (2)
180
9 Brand Rating in the Age of Information Sharing
as a product line extension strategy, you should take cannibalization of the parent
brand into consideration, and (3) the items for rating a product line extension brand
vary by whether consumers have purchased the parent brand.
In this chapter, the purpose of analyzing product line extension is to explore how
the brand equity levels of the parent brand affect its product line extension brand.
Based on the results of analyzing the cookie category, Brand R’s equity was not
strong, although the brand equities of Brands Y and L were strong. Therefore, it
seemed that the factors that drive consumers to select the extension brand of Brand
R are largely unrelated to the equity.
In March 2012, a second survey was conducted. Product line extensions were
released in October 2010. Typically, new products are released in fall and spring in
Japan; products that have poor sales would disappear by the time the next product is
released. Because the product line extensions covered in this study were still in the
market at the time of new product release in the following season, we can assume
they are reasonably rated in the market.
I added certain survey questions to the first set of questions related to product line
extensions of the three brands. In particular, I added questions about how they rated
product line extensions (taste, quality, advertisement, etc.) and related to the parent
brand (“I reevaluated the parent brand,” “I don’t eat the parent brand,” etc.). The
sample population is those participants who responded to the first questionnaire. The
survey was designed in such a way that I could compare the previous survey results
such as how the same individuals thought of brand extensions. I asked 2000 subjects
of the first survey to participate and obtained responses from 1307 of them, that is,
65.35% responded to the second survey. The breakdown was 85 “early listener”, 390
“active listeners”, 372 “passive listeners”, 165 “uncertain listeners”, and 295 “disinterested listeners”. Moreover, while the percentage of “early listener” that cooperated
with the survey was lower (53.46%) compared to the overall percentage that of the
remaining segments was about the same as the overall percentage.
From the question items, purchase intentions for each product line extension brand
were measured via dummy variables that were used as dependent variables; however,
the rating and the relationship with each product line extension brand were used as
independent variables to create binomial logit models. The objective was to verify
the statement “(1) strong brand equity of the parent brand is essential,” which was
proven in the earlier literature review. An analysis was conducted for the product
line extensions, i.e., Brands R_ex, Y_ex, and L-ex for three Brands R, Y, and L,
respectively. Table 9.3 shows the summary of the results.
First, it was shown that the statements “it is a leading shortbread/cookie product,”
“the quality is good,” “it tastes great,” “it’s plain (in a negative sense),” “it suits me,”
“I’m interested,” “I would like to try it,” “I can imagine how it tastes,” and “the
quantity is just right” were common factors that affected the three brands. Because
“it tastes great” and “the quality is good” are basic qualities of food products, it is
understandable that such factors affect purchase intention first. Moreover, since “I
can imagine how it tastes” is derived from trying out the parent brands, the meaning
of product line extension could be identified.
9.3 Effect of Product Line Extension and Its Measurement
181
Table 9.3 Reason for choosing the extension brand (only variables that are significant at the 5%
level are listed)
Brand R_ex
Brand Y_ex
Brand L_ex
It’s a leading shortbread/cookie product
0.714
0.459
0.523
The quality is good
0.616
0.445
0.683
It tastes good
1.119
1.183
1.104
I often see it at the store
0.446
I often see the advertisement
0.511
It’s likable
I’m used to eating it
0.551
1.055
0.184
It’s familiar
0.144
It’s plain
−1.227
−0.820
−1.116
It suits me
1.412
1.236
1.298
−1.146
1.176
−0.833
It’s for older people
It’s old fashioned
It’s luxurious
−1.026
It’s convenient
0.529
It’s sober
−1.462
It’s very particular
−1.496
0.998
−0.922
−1.854
1.130
0.659
0.907
0.528
0.752
I want to try it
1.226
0.735
0.988
I want to recommend it to others
1.285
0.836
I can imagine how it tastes
0.507
0.591
0.443
The volume is just right
0.699
0.491
0.487
It’s not really new/it seems common
−0.913
−0.751
It doesn’t feel right
I’m interested
It’s interesting
It makes me want to eat the original cookie
0.408
I have a renewed appreciation for the original
cookie
0.445
The characteristics of the original cookie is
leveraged
0.633
−0.836
It improved my impression of the original cookie
The image of the original cookie has changed
0.737
−1.369
I want to try it along with the original cookie to
compare
−1.052
0.569
0.680
Constant
−1.608
−1.486
−1.776
Nagelkerke R-square
0.445
0.379
0.436
182
9 Brand Rating in the Age of Information Sharing
Next, I will compare Brand Y-ex, which is an extension brand with strong brand
equity; Brand L_ex, which is an extension brand; and Brand R_ex. which is an
extension brand with weak brand equity. “It makes me want to eat the original cookie”
and “I have a renewed appreciation for the original cookie” are relevant for Extension
Brand Y-ex and “I have a renewed appreciation for the original cookie” and “the
characteristics of the original cookie is leveraged” are relevant for Extension Brand
L_ex. Although Extension Brand L_ex. came out to be positive and was statistically
significant to make the factors representing the relationship with the parent brand
important variables, those variables representing the relationship with the parent
brand were not significant for Extension Brand R_ex, which had weak brand equity.
The variables that become significant only under Extension Brand R_ex—“I often
see it at the store” and “I often see the advertisement”—were effective factors when
a completely new product was introduced to the market. Thus, unlike Extension
Brands Y_ex and L_ex, Extension Brand R_ex is rated based on promotions and
advertisements that are unique to this brand rather than the brand equity of the parent
brand. Furthermore, it became clear that while brand equity affects consumers’ choice
of extension brand when the equity of the parent brand is strong, advertisements and
promotions affect the choice when the equity of the parent brand is weak, similar to
the case of a completely new product.
In addition, given that Extended Brand Y_ex was still in the market, we could
observe that product line extension was feasible when purchased by well-informed
people whose affective commitment was strong even if the market share of the parent
brand was small. When rated based on factors related to market environment, such
as market share and purchase intention, Brand Y’s brand power was not necessarily
strong. In fact, we would have judged that product line extension is difficult. However,
as shown in this chapter, because Brand Y scored high on supporters, commitment,
and decision-making route, we can judge that their brand power was strong if we
consider the consumer behavior theories. Thus, even if the market share is small,
a product line extension is feasible and the extended brand can leverage the parent
brand’s equity if the brand can score high on the brand rating indicators that are
based on the theory of consumer behavior, which has been repeatedly discussed in
this book. This shows that you don’t have to have a strong concept of “the width of
brand” to have a strong concept of the depth of brand,” to expand the line. It shows
that brand strength needs to be viewed not only in terms of “the width of brand”, but
also in terms of “the depth of brand,”.
Next, I will examine the parent brand’s cannibalization by its extension brand.
The purpose is to verify the statement “(2) as a product line expansion strategy, you
should take cannibalization of the parent brand into consideration for the outcome,”
which has been mentioned in previous studies.
Purchase patterns of a sample consumer population can be classified into eight
types based on three questions: (1) whether the 2010 survey indicated they purchased
the parent brand, (2) whether the 2012 survey indicated they purchased the extended
brand, and (3) whether the 2012 survey indicated they purchased the parent brand.
For each pattern, the number of respondents is shown by brand in Table 9.4. Here,
Patterns 1, 2, 5, and 6 are those who had purchased a parent brand in 2010. In total,
9.3 Effect of Product Line Extension and Its Measurement
183
Table 9.4 Purchase pattern groups
2010 parent
brand
2012 extension
brand
2012 parent
brand
Brand R
Brand Y
Brand L
Pattern 1
◯
◯
◯
189
173
208
Pattern 2
◯
◯
×
67
52
44
Pattern 3
×
◯
◯
92
73
71
Pattern 4
×
◯
×
179
205
146
Pattern 5
◯
×
◯
84
85
126
Pattern 6
◯
×
×
60
43
68
Pattern 7
×
×
◯
55
48
52
Pattern 8
×
×
×
486
533
497
there were 400 respondents for Brand R, 353 for Brand Y, and 446 for Brand L.
Furthermore, Pattern 1 is a pattern in which they began purchasing extension brands
as well as parent brands in 2012. Pattern 2 is the cannibalization pattern, that is, they
switched from the parent brand to the extension brand; Pattern 5 is the pattern in
which they continued purchasing only the parent brand; and Pattern 6 is the outflow
pattern in which they no longer purchased either the parent brand or the extension
brand in 2012.
The remaining Patterns 3, 4, 7, and 8 are of those who had not purchased the
parent brand in 2010. Pattern 3 is inflows of those who began purchasing the parent
brand and the extension brand in 2012; Pattern 4 is those who purchased only the
extension brand or new customers who were acquired via extensions; Pattern 7 is
those who became new purchasers of the parent brand in 2012; and Pattern 8 is those
who did not purchase this brand at all in either 2010 or 2012. Incidentally, in 2010,
regardless of the actual situation, those who purchased the parent brand in 2012 fall
under Patterns 1, 3, 5, and 7. In total, there are 420 purchasers of Brand R, 373
purchasers of Brand Y, and 457 purchasers of Brand L.
First, I will examine the parent brands. As mentioned above, the number of parent
brand purchasers increased, although slightly for all the three brands. When the
loyalty rate (the percentage of those who had also purchased in 2010 out of those
who purchased it in 2012) was measured for each brand, it was 65.0% for Brand R,
68.07% for Brand Y, and 62.28% for Brand L. Table 9.2, which was prepared based
on the question regarding “the brand I want to purchase the most,” shows the value of
32.4% for Brand R; however, we can see that the value considerably improved when
the brand switch matrix was prepared based on the question related to “the brand I
purchased.” Possibly, the question on “the brand I want to purchase the most,” which
considers the individual’s strong intention, is more suitable for measuring the brand
strength compared to the question related to the fact that they purchased.
Next, I will examine extension brands. Purchasers of extension brands fall under
Patterns 1, 2, 3, and 4 and the number of purchasers for each brand is 527, 503, and 469
for Brands R_ex, Y_ex, and L_ex, respectively. Because the number of parent brand
purchasers in 2012 was 420, 373, and 457 for Brand R, Y, and L, respectively, the
184
9 Brand Rating in the Age of Information Sharing
penetration rates were higher compared to the parent brands; moreover, the extensions
could be considered as successful too. From the extension brand purchasers, those
who did not purchase the parent brand in 2012, although they did in 2010—purchasers
resulting from cannibalization—fall under Pattern 2. Note that there are only 67,
52, and 44 of such consumers for Brands R_ex, Brand Y_ex, and Brand L_ex,
respectively. However, because there are Patterns 3 and 4 for those who purchased
the extension brand in 2012, although they did not purchase the parent brand in 2010,
the total was 271 for Brand R_ex, 278 for Brand Y_ex, and 217 for Brand L_ex, which
makes the number of newly acquired customers overwhelmingly larger compared to
the numbers reduced by cannibalization. Therefore, in terms of the three brands that
were surveyed, product line extensions were successful even when cannibalization
was considered.
Finally, I will verify “(3) the points for rating the brand of product line extension
vary by whether the consumer has purchased the parent brand” shown in the literature
review. Among the abovementioned patterns, Pattern 1 is those who purchased the
parent brand in 2010 and 2012 and purchased the product line extension brand in
2012, whereas those who did not purchase the product line extension brand in 2012
are Pattern 5. If we compare these two patterns, it will reveal the difference in rating
between the purchasers and non-purchasers of line extension brands from those who
purchased the parent brand. Similarly, Pattern 4 is those who purchased the line
extension brand in 2012 without purchasing the parent brand in 2010 or 2012, while
Pattern 8 is of those who did not purchase the line extension brand. Therefore, a
comparison of these two patterns will why those who have not purchased the parent
brand purchased the line extension brand. As part of the first analysis, I analyzed the
reason for purchasing the extensions of three brands using binominal logit analysis in
which whether the respondent has purchased the brand was the dependent variable.
Table 9.5 shows the results in which “Yes” is the result of the analysis on factors
affecting the purchase of extension brands when the individual has purchased the
parent brand and “No” is the result of the analysis on factors affecting the purchase
of extension brand when the individual has not purchased the parent brand. Note that
only variables that are statistically significant at the 5% level are shown.
First, we can see here three factors, namely, “I’m interested,” “I think it’s interesting,” and “I want to try it,” which drive individuals to purchase extension brands
of all three brands regardless of whether they have purchased the parent brand.
Furthermore, we cannot say that these three factors are prerequisites for selecting an
extension brand.
Next, if we closely look at each brand, we can clearly see that there are differences.
First, reasons for purchasers of parent brand to purchase the extension brand as well
as reasons for non-purchasers of parent brand to purchase the extension brand do not
overlap except for the above three cases. In the case of Brand R, which was deemed
to have weak brand power, there were multiple overlapping factors and Brands Y and
L were deemed to have stronger brand power compared to Brand R. In particular, in
the case of Brand L, all factors are identical except for one (“I want to try it along with
the original cookie to compare”). Thus, the analysis seemed to indicate that while
having purchased the parent brand makes a difference to determine the subsequent
purchase of the extension brand when the brand power of the parent brand is not
9.3 Effect of Product Line Extension and Its Measurement
185
Table 9.5 Whether the parent brand was purchased and rating of the extension brand
Brand R
Brand Y
Brand Y
◯→◯ ×→◯ ◯→◯ ×→◯ ◯→◯ ×→◯
−1.230
It doesn’t feel right
−1.339 −1.855
I’m interested
1.525
1.027
1.032
0.654
1.081
0.942
It’s interesting
0.783
0.456
0.816
0.729
0.636
0.975
I want to try it
1.336
1.455
0.947
1.021
1.302
1.181
I want to recommend it to others
1.171
I can imagine how it tastes
0.597
The volume is just right
1.368
It’s not really new/it seems common −0.763
0.850
0.662
0.649
0.592
0.944
1.392
0.534
0.641
0.902
0.659
0.924
1.012
0.847
−0.817 −0.832
It makes me want to eat the original 0.604
cookie
0.550
I have a renewed appreciation for
the original cookie
1.211
The characteristics of the original
cookie are leveraged
0.622
It improved my impression of the
original cookie
−1.150
There is something good that wasn’t
in the original cookie
0.663
0.802
−1.404
2.062
It’s more suitable to the present day
compared to the original cookie
0.531
0.570
I want to try it along with the
original cookie to compare
0.883
0.747
Constant
−0.329 −1.611 −0.322 −1.541 −0.648 −1.891
Nagelkerke R-square
0.290
0.302
0.303
0.236
0.285
0.304
strong, that difference is smaller when the brand power of the parent brand is strong.
This is probably because brand power of the parent brand has a significant impact at
the time of selecting an extension brand; therefore, I was not able to verify with this
data “(3) the points for rating the brand of product line extension vary by whether
the consumer has purchased the parent brand,” which was mentioned in the literature
review.
9.4 Conclusion and Future Implication
In this chapter, I created an online monitor to survey real products and verify whether
the consumer behavior theories in the era of the Internet as discussed in the previous
186
9 Brand Rating in the Age of Information Sharing
chapters are really effective in brand rating. The primary objective was to comprehensively judge the effectiveness of theories using one data set for one target product
and to re-examine the new findings obtained from different data for different target
products.
Thus, it became clear that four indicators, namely, “how far along is the consumer
in the decision-making process,” “who purchases it,” “type of commitment,” and
“central vs. peripheral attitude formation route (ELM),” could be used for brand
rating. Here, it was revealed that while results of the analysis for the stage of decisionmaking process were about the same as the market share of a given brand, the
results of the remaining three indicators were completely different. However, the
three indicators showed similar trends, which made it clear that brand rating has
two major angles, namely, product power from the perspective of “width of brand”
such as market share and recognition about whether it is often seen by consumers,
and product power from the perspective of “depth of brand,” which is the degree of
strong connection with the product, that is, purchaser’s attachment or interest and
the quality of purchasers.
In fact, when all six brands are rated by these two indicators, two brands were
found to not have strong brand power because the abovementioned two indicators,
“the width of brand” and “the depth of brand,” were low. Furthermore, one of the
remaining four brands showed extremely high scores for both these indicators. When
I prepared a brand switch matrix between two points in time—2010 and 2012 when
the surveys were conducted—and observed the trend, while the loyalty rates of these
two brands with poor scores for the above two indicators remained at 30%, the
loyalty rates of the remaining four brands reached 60%. Thus, we can say it shows
the effectiveness of indicators that were prepared based on the theories of consumer
behavior. Moreover, even when the loyalty rate was the same at 60%, brands that
scored high on both indicators had both high loyalty rates and high rates of inflow
from other brands and actually low rates of outflow. Therefore, it became clear that a
truly strong brand is strong not only in terms of the indicators for brand width, such
as market share and recognition, but also in terms of the indicators for brand depth,
representing the degree of attachment to the brand.
Next, three brands out of those that I used for the analysis launched product
line extensions after the first survey, I conducted a detailed analysis of the power
of the parent brand as well as consumer’s choice of product line extension brand
by surveying the same respondents again. Consequently, while consumer’s relationship with the parent brand became important for choosing the line extension brand
when the parent brand’s power was strong, the same general reasons for selecting a
new product such as advertisement and exposure at the store emerged as reasons for
selecting the line extension brand when the parent brand’s power was weak. Furthermore, it was shown from here that a brand has strong brand power even if the parent
brand’s market share is not large when there were multiple consumers that had a
strong attachment to the brand. This allows leverage of the brand power of the parent
brand when extending the product line. In terms of cannibalization of the parent
brand, which had been noted in previous studies for product line extensions, it was
not an issue in the analysis. In fact, it actually helped increase the penetration rate for
9.4 Conclusion and Future Implication
187
the entire brand. Furthermore, when I examined the relationship between purchase
experience of the parent brand and the reason for purchasing the line extension brand,
the reasons did not differ much when the parent brand’s power was strong, although
they differed when the brand power was not strong.
As described, when I considered the theories of consumer behavior, which had
been individually measured in this book and analyzed them with one database, I
was able to confirm that all of them were valid concepts. The fact that we could
demonstrate that brand power could be understood by two aspects—the width of
brand and the depth of brand—was particularly important. Previously, it was thought
that consumer behavior research was effective for marketing strategy to search for the
axis of segmentation as well as for the decision-making process and media. However,
as it was shown to be effective for brand rating, we can say it newly demonstrated
the importance of consumer behavior research for marketing strategy.
Notes
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Inoue, Atsuko (2009), ‘The Relationship of Brand Commitment and Consumer
Behavior’, Journal of Marketing & Distribution, Vol. 12, No. 2, pp. 3–22.
Aaker, David A. (1991), Managing Brand Equity. New York: The Free Press,
p. 208
Keller, K.L. (2008), Strategic Brand Management: Third Edition, PrenticeHall.
Aaker, David A.; K.L. Keller (1990), ‘Consumer Evaluations of Brand
Extensions’, Journal of Marketing, Vol. 54, January, pp. 27–41.
Park, C. Whan; Milberg,S.; Lawson, R. (1991), ‘Evaluation of Brand Extensions: The Role of Product Feature Similarity and Brand Concept Consistency’,
Journal of Consumer Research, Vol. 18, pp. 185–193.
Volckner, Franziska; Sattler Henrik (2007), ‘Empirical generalizability of
consumer evaluations of brand extensions’, International Journal of Research
in Marketing, Vol. 24, pp. 149–162.
Caldieraro, Fabio; Kao, Ling-Jing; Cunha, Marcus, Jr. (2015), ‘Harmful
Upward Line Extensions: Can the Launch of Premium Products Result in
Competitive Disadvantages?’, Journal of Marketing, Vol. 79, pp. 50–70.
Wilson, Lynn O.; Norton, J.A. (1989), ‘Optimal Entry Timing for a Product
Line Extension’, Marketing Science, Winter, pp. 1–16.
Reddy, Srinivas K.; Holak, Susan L.; Subodh Bhat (1994), ‘To Extend or Not
to Extend: Success Determinants of Line Extensions’, Journal of Marketing
Research, Vol. 31, pp. 243–262.
Sinapuelas, Clark, Ian; Sisodiya, Sanjay Ram (2010), ‘Do line extensions influence parent brand equity? An investigation of supermarket packaged goods’,
Journal of Product & Brand Management, pp. 18–26.
Chapter 10
A New Decision-Making
Process—A Circulating-Type
Communications Model
We can summarize the relationships between consumer behavior and marketing
strategy in the history of research into two broad areas: research into segmentation
and research into the consumer decision-making process. Segmentation is important
in a marketing strategy as the pre-stage when determining the targets and thinking
about positioning. While working on the issues of robustness and accessibility, it
has been used as the factor axis, such as for demographic and lifestyle factors, and
moreover for purchase histories representing FSP data. In contrast to this, research
on the consumer decision-making process focuses on the sequence of events from
perceiving consumers’ needs through their actual purchases. Within this time series,
it captures how consumers’ product evaluation criteria change and how they narrow
down products to a single, final product. It has been proven useful for understanding
the changes to the sources of information that consumers encounter within this time
series and the order of information deemed necessary. In terms of the research lineage,
there are two categories. From the standpoint of consumer behavior theory, there
exist the stimulus–response type and the information-processing type, while from
the standpoint of research into the flow of information, there has been research into
AIDA, AIDMA, and other aspects.
The use of the Internet has spread rapidly since 2000. As has been outlined in
the previous chapters, it is naturally having a major impact on both these research
areas. For example, the research described in Chap. 6 on communications-type
consumers and the research in Chap. 7 on uncertain listeners show that the information gap is effective as a new axis for segmentation. The research on changes to the
decision-making process described in Chap. 2 shows that the scope of this research
is expanding to one step further after the purchase stage as far as considering the
impact on potential customers after a purchase.
Previously, as for the flow of these two research areas, segmentation research
ascertained consumers as a segment in the developments from mass marketing. On
the other hand, decision-making process research targeted individual consumers for
research, which it considered to be a separate research area. However, as has been
clarified so far in this book, an information-sharing approach that should be applied to
© Springer Nature Singapore Pte Ltd. 2021
A. Shimizu, New Consumer Behavior Theories from Japan, Advances in Japanese
Business and Economics 27, https://doi.org/10.1007/978-981-16-1127-8_10
189
190
10 A New Decision-Making Process—A Circulating-Type …
the decision-making process is a concept in which one consumer affects many potential customers, and there are limits to ascertaining the decision-making process under
the conventional, individual consumer concept. We need to consider the fact that this
information sharing is performed by people who belong to a segment with comparatively high information sensitivity and that it is not performed by all members. In
that context, we can see that the conventional consumer decision-making process,
in which the focus is placed on the individual up to the purchase stage and the
process is completed by that individual, is insufficient to explain consumer behavior
at the current time. It is also a fact that this purchase experience affects other potential
customers and there occur differences in the extent of this effect due to the differences
in consumers’ information sensitivity.
While it is obvious that the Internet has a significant effect on people’s lives, when
speaking from the area of marketing research, it has been clarified that an individual’s
behavior does not end within that individual. The behavior also influences society,
and that the extent of this effect on society depends on the individual’s information
sensitivity. The situation provides opportunities for mass marketing to have points
of contact with research on the behaviors of individual consumers.
This is the perspective taken in this chapter, and I will confirm the possibilities for
this concept using experimental data, while simultaneously showing the concept for
a new consumer decision-making process in the Internet era. My aim is to propose a
new theory to serve as the basis on which to consider the synergies between marketing
strategy and consumer behavior theory.
10.1 What Has Been Clarified in This Chapter Up to This
Point
From among the points initially clarified in this book, in this chapter, I will organize
the theories useful in deriving a new concept for the decision-making process and
its introduction.
First, in Chap. 1, I described the rapid development of the Internet in the last
10 years and the resultant emergence of an information gap between consumers.
When the Internet was introduced, differences occurred in the amounts of information
depending on whether or not a person used the Internet. At present, there are many
people in a group wherein the amount of information they acquire from the Internet
and existing media is well balanced. Even on the Internet, it is understood that the
amount of information of a group that only uses the mobile phones to go online is
smaller than that of a group that mainly uses the mass media. Thus, it can be said
that we have changed to a more comprehensive understanding from the presence or
absence of tool use.
This difference in the amount of information has been shown to be effective
as a segmentation axis in the research on the emergence of communications-type
consumers in Chap. 6, and the research on uncertain listeners in Chap. 7.
10.1 What Has Been Clarified in This Chapter Up to This Point
191
In Chap. 6, we explored the effectiveness of the information gap from the viewpoint of CRM, which applies the segmentation theory to actual practice. Here, it was
explained that segmentation through RFM analysis, which has been the conventional
method until now, is effective in terms of improving an individual customer’s lifetime
value. However, in the information-sharing era, it is necessary to consider the effect of
the individual on potential customers, and the effectiveness of creating messages for
communication-type consumers, who have high commitment to the relevant product
and high information sensitivity, has been confirmed by actual data. As a result, it has
been shown that the products preferred by this segment have strong brand power, and
also that a segment combining information sensitivity and commitment is effective.
In Chap. 7, we saw that, from creating a specific reference group as the brand itself
as the target, segments can be created from consumers’ individual brand preferences.
We also found that consumer’s information sensitivity can be measured from the
newness or oldness of the brands they prefer. Based on this, I created five segments. I
focused on the active listener segment with high information sensitivity, which clearly
communicates information and listens with a commonsense way of thinking. I also
focused on the uncertain listeners segment, which has low information sensitivity and
whose sense of being outdated is undeniable. From the actual data, it was found that
for a brand to endure, it is important that it is supported by the active listeners segment
but not supported by the uncertain listeners segment. Furthermore, this rule applies
not only to existing products but to new products as well. Moreover, it was shown
that the active listener segment shares information on the relevant hit product from
the beginning and it has the power to make a new product a topic of conversation. The
more they support a brand, the more this brand is preferred, because they can make the
relevant product a topic of conversation. We can see here the point of contact between
segmentation from information sensitivity and an individual’s information-sharing
behavior.
In terms of showing the relationship between the consumer decision-making
process and the Internet, the changes to the comprehensive decision-making process
were shown in Chap. 2; measuring the effects of blogs was considered in Chap. 3; and
then the mechanism of attitude formation in consumers who disseminate information
was discussed in Chap. 5.
First, in Chap. 2, in the conventional, stimulus–response type decision-making
process and the information-processing-type decision-making process considered
from the perspective of consumer behavior theory, we saw that the main research
themes have been on the stages up to an individual consumer’s purchase. However,
after the purchase, although research on consumer satisfaction has been conducted,
the effects of the purchase on other potential customers have not been considered at all. In the AIDMA-related research focusing on the flow of information, in
every case, the research findings have come from Japan, where information-sharing
concepts such as AISAS®, AIDEES, and SIPS are being advocated. In addition, in
the consumer decision-making process considered in the research area of consumer
behavior theory, the necessity of including this way of thinking has been theoretically
clarified. Data have been collected in accordance with the flow of AISAS®, which
shows that there certainly exists a segment that shares information on the relevant
192
10 A New Decision-Making Process—A Circulating-Type …
product after a purchase. It also shows that this information sharing is more likely to
occur by consumers who are initially aware of and have an interest in the product,
and who collect information that leads to a purchase.
In Chap. 3, we confirmed the ways in which word-of-mouth information, which is
widespread on the Internet, is useful for consumers’ decision-making both in terms
of quantity and quality. When the quantity of blogs increases, the number of searches
also increases, which has the effect of increasing sales. For the quality aspect of blogs,
it has been clarified that reading a blog strengthens a consumer’s attitude, and that
in particular, reading a blog that is favorable to a product improves their evaluation
of that product and increases their intention to purchase it. Currently, the sharing of
consumer information is changing from blogs to SNS, such as Twitter and Facebook.
Moreover, the quantity of this information is increasing. In other words, the extent
to which the sharing of information by consumers affects potential customers is
increasing, and the significance of this being incorporated into the decision-making
process has been shown.
In Chap. 5, I discussed the conditions for consumers’ sharing of information in
accordance with the ELM approach, which uses conventional consumer behavior
theory. Here, as the condition for consumers to disseminate information, the case of
decision making on the central route deemed to be logical was cited, and it was shown
that in this case, the affective commitment toward the relevant product becomes
higher. Affective commitment, as was also described in Chap. 6 on segmentation,
shows the strength of consumers’ emotional ties to a product, and word-of-mouth
communication plays a key role in strengthening the ties. As clarified in the detailed
analysis in Chap. 4, affective commitment declines when purchases are continuously made in bargain sales. Considering this in conjunction with what was stated
in Chap. 2, when purchases are continuously made during bargain sales, consumers
come to respond to the stimulus of the bargain sale without needing to have an awareness of or interest in the relevant product. Therefore, it is considered that in such
cases, attitude formation does not occur on the central route. As a result, affective
commitment declines and information sharing does not take place. This proves the
notion that excessive promotions cause brand power to decrease.
On taking an overview of segmentation and consumers’ decision-making process,
as described above, we can see the following: consumers with high information sensitivity collect information on products that they are aware of and have an interest in;
their attitudes are determined and they then purchase products for which they have
a strong attitude mainly through the central route; and after that, they share information. However, it is considered that consumers who do not have high information
sensitivity purchase a product having awareness of or interest in it, form weak attitudes mainly through peripheral routes, and therefore do not share information. In
particular, the sharing of favorable information has a major effect on determining
customers’ attitudes toward a product. It is necessary for people with powers of
discernment to identify good products and form a strong affective commitment to
them, and then to spread information on them.
10.2 A New Theory—Toward Circulating-Type Marketing
193
10.2 A New Theory—Toward Circulating-Type Marketing
Based on the above discussion, we will next consider a comprehensive consumer
decision-making process that should be proposed.
A point emphasized in past research on a comprehensive consumer decisionmaking process is the point of creating a model that is appropriate for any situation
and for any consumer. For example, in the Howard and Sheth model (1969), which
is a classic comprehensive decision-making process model, the relevant product is
changed to adapt to consumers’ decision-making process in each period of its lifecycle. Namely, if the product is in its launch period, extensive problem solving is
conducted that places the focus on information-search activities; if it is in its growth
period, limited problem solving is conducted; if it is in its mature period, routine
problem solving is conducted. It means this model makes possible a response in every
period.1 In the Bettman model (1977), which is a typical information-processing-type
model, the focus is placed not on the product, but on the differences in consumers’
information-processing capabilities, and it shows that in the flow from consumers’
creation of a purchase target through to a purchase, if their information-processing
capabilities are controlled, the decision-making process changes due to these differences.2 Furthermore, in ELM presented in Chap. 5, which is one more comprehensive
decision-making model, when consumers’ motivations toward the relevant product
differ, the route that determines their attitude will also be different depending on
whether or not they are aware of the relevant product. Specifically, in the event that
there are motivation and awareness, decision making is carried out on the central
route, while in other cases, decisions are made on peripheral routes (Petty and
Cacioppo 1983).3 In other words, a comprehensive decision-making model must
be a model that can be applied regardless of the relevant product and differences in
consumers.
So far from this book, it is clear that the consumer decision-making process does
not end with the purchase and that it must be understood by adding the viewpoint of
information sharing after the purchase. However, as we saw in Chap. 1 on segmentation through the information gap, there are still many people who make decisions
through the conventional method of coming into contact with the media. In the event
of a situation of a purchase made from a careful consideration in accordance with
the information-processing-type, it can be thought that there will be many cases of
stimulus–response type purchases made on sales floors.
Figure 10.1 shows an information-circulating-type decision-making model that
proposes this situation. The features of the model are (1) it is not the conventional
one-way model that starts from awareness and ends in the purchase, as it assumes that
information sharing after the purchase affects the following information search and
that information circulates; and (2) it assumes that the decision-making process is not
completed only within the individual and that individuals affect the market as a whole.
For example, in the information-processing-type, it is considered that the decisionmaking consumer decides in the initial stage and then makes the purchase, and that
within this sequence, the “active listener” type people disseminate information, as
194
10 A New Decision-Making Process—A Circulating-Type …
Pre-purchase behavior
Post-purchase behavior
Formation of consideration set
and information contact points
Contact with media
Commitment research
Consumer satisfaction
Loyalty
SNS
Word of Mouth
Online media
Mass media, such as TV
commercials
Circulating-type marketing
theory and strategy
How does a 4P strategy change?
New customer management
Behavior at places of purchase
Changing / strengthening attitudes and in-store marketing
Intentions when purchasing a product
A non-planned purchase
Promotions
Consumers come to
make a purchase
from anywhere
Fig. 10.1 A communication framework that considers circulation
assumed by AISAS® and AIDEES. Furthermore, it seems that “uncertain listener”
type consumers mainly make stimulus–response type purchases and participate in
the purchase at the sales venue, and their behavior ends with only the purchase.
People whose medium of awareness is normally from the use of SNS ought to enter
the information-circulation wheel from the information-sharing stage, and then at
the information-search stage, to confirm information by referring to mass media and
other sources of information, to make the purchase, and then to act in SIPS type ways
to disseminate information. On assuming this sort of circulating-type information,
it becomes possible to cover not only the conventional stimulus–response type and
the information-processing-type approaches but also the AISAS®, AIDEES, and
moreover the SIPS approaches. It can be said to be a conceptual model that can be
applied to whatever the situation and consumer, which were the previously stated
requirements of a comprehensive decision-making model.
If we organize a framework of the respective media that the consumer comes into
contact with before the purchase, at the place of purchase, and after the purchase, they
are mainly the mass media before the purchase, promotions carried out by retailers at
the place of purchase, and SNS by consumers after the purchase. Furthermore, from
the academic point of view, considering that there have been a lot of research into
information-processing-type consumer behavior, research studies on aspects such as
involvement, knowledge, and consideration sets can be arranged into pre-purchase
behaviors. Since the appearance of POS data, many marketing-science researchers
are participating in this area, and research on promotions, which is flourishing, has
come to be positioned as research into the purchase-behavior stage. Furthermore,
10.2 A New Theory—Toward Circulating-Type Marketing
195
research into aspects such as individual consumers’ satisfaction, commitment, and
SNS are post-purchase consumer behaviors. It is considered that assuming such
an information-circulating-type, comprehensive decision-making model makes it
possible to organize the elements academically.
This flow is not limited to the consumer behavior theory. In Fig. 10.2, it is expanded
to the flow for a hit product. As shown in Chap. 7, among the group of people who
share information on hit products, the percentage of active listeners is extremely
high. In Chap. 9, we saw that there are many active listeners among the group who
support strong brands, and they have high commitment toward the relevant product.
It was also clarified in Chap. 2 that people who make a purchase with awareness
and interest are more likely to share information. Furthermore, in Chap. 3, we saw
that information that has been shared is useful at the time of making a purchase.
When taking together these research findings with the above-mentioned informationcirculating-type consumer behavior, we can assume that a framework for continuous
sales in contemporary society is completed from the following sequence; first, the
high-information sensitivity, active listener segment obtains information on a new
product, they become interested in that product, and after purchasing it, they share
favorable information on it and this shared information affects people other than those
in the active listener segment at the time they make a purchase. The assumptions up to
the spread of a product can be said to be theoretically explained using this consumer
decision-making process.
The key for this framework of circulating, continuous sales is whether the relevant
product contains the elements that will make it a topic of conversation in the active
listener segment. As clarified in the research on judgments in Chap. 8, several types
of people are able to judge that a product is good. In this sense, it is a mistake to have
only a single topic of conversation. It is important to have multiple topics of conversation for realizing information sharing. The content of the topics of conversation is
The key to a product remaining in the market is whether the
relay of mass media→sales floor→WOM→mass media→sales floor
→WOM→ is skilfully created and it becomes a topic of conversation.
Launch
Growth /
mature periods
It is important to propose a media mix to create
the optimal circulation.
Fig. 10.2 Information circulation and spread of products
196
10 A New Decision-Making Process—A Circulating-Type …
considered to be judgments that pertain to factors relating to feelings to promote the
sales of new products.
From the broad categorizations of “feelings on appearance” and “taste” in the case
of a beer-type drink, it can be assumed that the important topics of conversation will
be the “feelings on appearance” and “the essential qualities of the product” in the
case of a new product. For existing products, this point cannot be confirmed using the
data in this book. However, communication topics that emerged in the summer from
TV celebrities have become a topic of conversations. In addition, when considering
examples of ice cream TV commercials aired in this period and that of long-selling
chocolates whose sales were revived by communications that appealed to the essential
quality of their content, there is good reason to assume that we can apply “feelings
on appearance” and “the essential quality of the product” to the existing products
too.
Next, I shall clarify whether the circulation of information assumed here actually
takes place, from the results of an experiment survey carried out at the time of a new
product launch.
10.3 Overview and Results of the Experiment Survey
I conducted the experiment survey in combination with a sampling experiment for
a new product: a non-alcoholic beer. Specifically, I carried out surveys thrice; (1) a
preliminary survey before the launch of beer-type drinks, including a non-alcoholic
beer; (2) the first survey, on a sample distribution and drinking experiment carried
out at the same time as the new product launch, and (3) the second survey asking
about usage conditions one month after the launch. When conducting the sample
distribution experiment, for confirming whether or not information was circulating
properly on SNS, the sampling-experiment participants disseminated information on
blogs and Twitter, and the results were observed. The target product, a non-alcoholic
beer, is a successful product that has a high market share even a year and a half after
its launch. I then ascertained and substituted the pre-purchase conditions in sampling
the purchases in the preliminary survey, the conditions after the purchase in the first
survey immediately after the sampling, and the conditions for repeat purchases in
the second survey.
For the survey subjects, I used the Kikimimi panel created in Chap. 9. This was for
confirming the circulation of information by people with high-information sensitivity.
From this panel, I selected 1,500 people after obtaining their permission to view the
content of their blogs and Twitter accounts and chose those people who had actually
written about the new product to be sampled as subjects. Furthermore, for ascertaining
the actual purchase conditions, separate to the Kikimimi panel, I conducted the same
survey and sampling using the Distribution Economics Institute of Japan’s panel
(DEIJ) with purchase histories. As shown in Chap. 7, the DEIJ’s panel, the same
as the Kikimimi panel, flags-up aspects like active listeners and uncertain listeners.
However, information sharing, such as via blogs, was not questioned in this sample.
10.3 Overview and Results of the Experiment Survey
197
In total, 836 subjects (719 people on the Kikimimi panel and 117 people on DEIJ’s
panel) answered the three surveys, including the preliminary survey. Of these, 120
people were active listeners (86 people on the Kikimimi panel and 34 people on
DEIJ’s panel). The preliminary survey was conducted from January 27 (Friday) to
February 12 (Sunday) in 2012; the samples were sent on February 17 (Friday); the
first main survey was conducted from February 24 (Friday) to March 4 (Sunday); and
the second main survey was conducted from March 30 (Friday) to April 8 (Sunday).
First the subjects were asked whether they already know of the relevant product
before sampling it. If yes, they were also asked from which media did they acquire that
information. The results were aggregated according to the Kikimimi panel listening
segments, as illustrated in Table 10.1. The figure shows that the percentage of people
who answered that they had heard about the product at an early stage. We can see from
the size of Pearson chi-square value that there are no significant differences among
these five segments (χ2 (8) = 8.799, p < n.s.). The brand targeted for sampling was
a product that had become a topic of conversation through the mass media before its
launch, and information on the relevant brand was fairly prevalent before its launch
even in the segment that did not actively collect information. Figure 10.3 shows the
awareness pathway before the launch, indicating that the source of information was
TV in an overwhelming majority of cases, followed by newspapers and company
Table 10.1 Knowledge of the product
Knew of it before
Remembering the
product after seeing
it
Do not know
Frequency
11
4
17
Expected
frequency
7.4
5
19.7
Frequency
40
23
85
Expected
frequency
34.1
23
90.9
Frequency
133
82
347
Expected
frequency
129.4
87.4
345.1
Uncertain
listeners
Frequency
17
18
58
Expected
frequency
21.4
14.5
57.1
Disinterested
listeners
Frequency
27
27
101
Expected
frequency
35.7
24.1
95.2
Total
Frequency
228
154
608
Expected
frequency
228
154
608
Early listeners
Active listeners
Passive listener
198
10 A New Decision-Making Process—A Circulating-Type …
18
21
Conversaons between friends and acquaintances
6
community site
4
Individual SNS
21
15
28
manufacture's Homepage
8
magazine
52
19
22
newspaper
55
2
5
radio
84
TV
0
20
40
remembering the product aer seeing it
60
80
146
100
120
140
160
knew of it before
Fig. 10.3 The awareness pathway
websites. The results show that the product became a topic of conversation in the
mass media.
Next, I confirmed whether or not the subjects were aware of the product before
the sampling, their subsequent information-sharing behavior, and the extent of the
influence of this sharing. Tables 10.2 and 10.3 show the results on combining the
results of the first and second surveys. The statistical significance confirmed that
people who were aware of the product before the sampling talked to people who had
not consumed it (χ2 (2) = 6.007, p < 0.050) and also that they talked enthusiastically
about it (χ2 (2) = 11.037, p < 0.004). Based on the results shown in Fig. 10.3, we can
consider that the following flow was created: the people who knew of the product
from before the launch were highly interested in it; so they carefully examined the
sampling product; as a result, they disseminated information on it; and then this
Table 10.2 Whether talked to people who had not consumed it
Whether talked to people who had not
consumed it
No
Yes
27
33
Knew of it before
Frequency
Expected frequency
33
27
Remembering the product after seeing it
Frequency
21
8
Expected frequency
15.9
13.1
Do not know
Frequency
52
41
Expected frequency
51.1
41.9
Total
Frequency
100
82
Expected frequency
100
82
10.3 Overview and Results of the Experiment Survey
199
Table 10.3 They talked enthusiastically about it
They talked enthusiastically about it
No
Yes
Frequency
13
33
Expected frequency
33
27
Remembering the product after seeing it
Frequency
21
8
Expected frequency
15.9
13.1
Do not know
Frequency
52
41
Expected frequency
51.1
41.9
Knew of it before
Total
Frequency
100
82
Expected frequency
100
82
information became a topic of conversation. In fact, the relationship between the
non-alcoholic-beer drinking frequency surveyed in the preliminary survey and the
awareness rate of this product (Table 10.4) leads to the statistical confirmation that
among the people who drank a high percentage of a non-alcoholic beer brand before
the launch, many were aware of the relevant brand before the sampling compared to
the people belonging to the low-percentage group (χ2 (6) = 43.264, p < 0.004). The
fact that if people have an interest in the relevant product category, they will search
for information on it, confirms the statement made in Chap. 2.
Table 10.4 The non-alcoholic-beer drinking frequency and the awareness rate of this product
Knew of it
before
Remembering the
product after
seeing it
Do not know
Total
More than once
a week
Frequency
57
29
77
163
Expected
frequency
37.5
25.4
100.1
163
Two or three
times a month
Frequency
24
13
20
57
Expected
frequency
13.1
8.9
35
57
About once a
month
Frequency
24
17
66
107
Expected
frequency
24.6
16.6
65.7
107
Less than or
equal to
Frequency
123
95
445
663
Expected
frequency
152.7
103.1
407.2
663
Total
Frequency
228
154
608
990
Expected
frequency
228
154
608
990
200
Table 10.5 Relationship
between satisfaction with the
sampling product and actual
purchases
10 A New Decision-Making Process—A Circulating-Type …
Satisfaction
Score
Frequency
Standard error
Already
purchased
2.12
178
1.029
Plans to
purchase
2.46
187
1.012
No plans to
purchase
3.67
395
1.359
Average
3.01
760
1.393
Next, we will discuss in more detail the behaviors after the sampling. First, as
outlined in Chap. 4, in consumers’ behaviors after a purchase, affective commitment
is fostered from their satisfaction in the purchase, which then leads them to disseminate information. Tables 10.5 and 10.6 provide an overview of their satisfaction
with the brand sampled and their subsequent purchases and information sharing.
Table 10.5 compares sampling satisfaction scores by whether or not the sampled
products were purchased. Satisfaction scores are measured on a six-point scale, with
“very satisfied” as 1 point and “very dissatisfied” as 6 points. From this, it can be
seen that those who are more satisfied with the sampled products are more likely to
have already purchased or intend to purchase them (χ2 (2) = 366.334, p < 0.001). In
other words, it seems that there is a relationship between a person’s satisfaction and
purchase behavior. Next, Table 10.6 shows that an overwhelmingly high percentage
Table 10.6 Satisfaction level and intention to write SNS
Satisfaction level
Writing to blogs and Twitter
No
Greatly so
Yes
Total
Frequency
60
50
110
Expected frequency
71.36
38.64
110
We would think so
Frequency
104
79
183
Expected frequency
118.71
64.29
183
I think so
Frequency
151
74
225
Expected frequency
145.95
79.05
225
Frequency
98
30
128
Expected frequency
83.03
44.97
128
I don’t think so
Frequency
39
22
Expected frequency
39.57
21.43
61
I don’t think so at all
Frequency
41
12
53
Expected frequency
34.38
18.62
53
Somewhat disagree
Total
61
Frequency
493
267
760
Expected frequency
493
267
760
10.3 Overview and Results of the Experiment Survey
201
of people who were satisfied with the sampled product wrote on blogs and posted
on Twitter (χ2 (5) = 22.160, p < 0.001). That is, it was statistically confirmed that
product satisfaction is useful for purchases and information sharing.
For affective commitment, which shows the emotions after a purchase, I prepared
question items based on the results of Chaps. 4 and 9. When carrying out the first
survey, I obtained scores for each item on a five-point Likert scale for the products
that they purchased most. I conducted a factor analysis (principal factor analysis with
varimax rotation) of these evaluations and extracted the results (eigenvalue > = 1),
which are shown in Table 10.7. From these results, I extracted three commitments:
affective commitment (I want to talk to others about what I’m drinking and how
good it is, I expect others to think highly of me., etc.), fascination commitment (I
trust this brand, It’s the right brand for me, etc.), and calculative commitment (I buy
them because they’re often discounted.
I buy them because I can’t be bothered to consider other brands, etc.). The Cronbach’s α coefficient, which shows the cohesiveness of the scale, was above the
reference value (over 0.7), and it was understood that this scale is effective.
For these commitment scores, as it was only possible to collect data on the “brand
purchased the most,” from among the brands, for the three brands with high affective
commitment (brands A, B, and C) and the three brands with high calculative commitment (brands X, Y, and Z), I calculated the percentages for which postings on blogs
and Twitter existed. Table 10.8 shows each brand’s score for affective commitment,
calculative commitment, and the percentage of social networking messages sent to
Table 10.7 Factor analysis to extract three commitments
I feel a kind of attachment and familiarity.
I trust this brand.
It's the right brand for me.
I'm buying, and I'm a good judge of character.
I want to talk to others about what I'm drinking and how good it is
I expect others to think highly of me.
I can't think of anything else to do with this brand.
I'll buy it even if it costs a li le more than other brand
I buy them because they're often discounted.
I buy them because I can't be bothered to consider other brands
I'm afraid to buy a different brand and fail.
I'm buying it for no great reason
Table 10.8 Commitment and
sharing behavior
Factor 1
0.302
0.321
0.478
0.731
0.794
0.764
0.700
0.681
0.040
0.104
0.201
-0.041
Affective
commitment
Factor 2
0.786
0.785
0.692
0.425
0.273
0.310
0.182
0.185
0.037
-0.019
0.088
-0.147
Factor 3 Cronbach'sα
-0.011
0.871
-0.053
0. 008
0.019
0.081
0.889
0.093
0.232
0.026
0.539
0.910
0.788
0.686
0.662
Calculative
commitment
Communality
0.709
0.722
0.707
0.715
0.712
0.688
0.577
0.499
0.294
0.840
0.518
0.462
% of sharing
behavior
Brand A
0.177
0.006
4.6
Brand B
0.343
−0.516
4.7
Brand C
0.369
−0.190
4.7
Brand X
−0.402
0.290
1.7
Brand Y
−0.071
0.171
1.8
Brand Z
−0.027
0.127
2.7
202
10 A New Decision-Making Process—A Circulating-Type …
social networking sites. As the targets were existing products, the targets of analysis
were beers that are comparatively inexpensive. Inevitably, there were not that many
postings on blogs and Twitter about them. However, I found that the strength of
information sharing onto blogs and Twitter was higher for the three brands with high
affective commitment than the brands with high calculative commitment. In other
words, when the affective commitment toward the relevant brand is high, information
is shared even for the existing, inexpensive products. Although the number is small,
it is higher for these brands than for brands with high calculative commitment. This
confirms the need to conduct sales that increase affective commitment.
From the above analysis, it was clarified that among people who used the relevant
product category before the launch, the majority were aware of the relevant new
product before they sampled it. The people with awareness tended to talk about the
product they had sampled to other people. It was also found that there are relationships between satisfaction and purchases, satisfaction and information sharing, and
affective commitment and information sharing. This single-source data confirmed
the data shown separately in each chapter of this book.
Next, I examined the relationship between the differences in consumers’ information sensitivity and information sharing. This confirmed that as shown in Chap. 9,
people with high information sensitivity tend to share information. Therefore, we
explored whether there were differences in post-purchase behavior between listener’s
segments that I made in Chap. 7. For their behaviors after the launch, I asked 4 questions: “I bought another one.” “I’ve talked to people who haven’t taken this product.”
“People who have taken this product have discussed it with each other.” “I’ve recommended this product to others.” For the effects of word-of-mouth information, I also
asked 4 questions: “My story was the starting point of the conversation.” “There
were retweets and other responses.” “Someone actually bought this brand when they
found out their story.” “It was a great conversation.” Two of these eight questions
were statistically significant: “People who have taken this product have discussed
it with each other.” (χ2 (4) = 19.177, p < 0.001) and “Someone actually bought
this brand when they found out their story.” (χ2 (4) = 15.038, p < 0.001). They are
shown in Tables 10.9 and 10.10. The results showed that not only active listeners
made the relevant new brand a topic of conversation but also their comments affected
the purchases of potential customers. We can understand that the percentage of those
who circulate information is extremely high compared to people in general.
Furthermore, I confirmed purchases one month after the sampling according to the
listening segment. This is shown in Table 10.11. It is clear that there are differences
according to the listening segment, in that the active listener segment purchases or
plans a purchase at an early period (χ2 (8) = 33.005, p < 0.001). In contrast, only
a few people in the uncertain listener and disinterested listener segments planned
to make a purchase in the future. In Chaps. 7 and 9, we saw that good brands are
supported by many in the active listener segment but few in the uncertain listeners
segment; the results here follow it. Chapter 6 also showed that it was important to
be able to word-of-mouth to potential customers as well as having a high purchase
volume as a target for CRM. The analysis here reveals that the active listener segment
combines both elements. For this reason, capturing the active listener segment was
10.3 Overview and Results of the Experiment Survey
203
Table 10.9 People who have taken this product have discussed it with each other
No
Yes
10
0
Total
Early listeners
Frequency
Expected frequency
8.4
1.6
10
Active listeners
Frequency
21
13
34
Expected frequency
28.6
5.4
34
Frequency
89
15
104
Expected frequency
87.4
16.6
104
Uncertain listeners
Frequency
16
0
Expected frequency
13.5
2.5
16
Disinterested listeners
Frequency
17
1
18
Expected frequency
15.1
2.9
18
Passive listener
Total
10
16
Frequency
153
29
182
Expected frequency
153
29
182
Table 10.10 Someone actually bought this brand when they found out their story
No
Yes
3
2
Total
Early listeners
Frequency
Expected frequency
3.3
1.7
5
Active listeners
Frequency
8
15
23
Expected frequency
15.1
7.9
23
Frequency
42
17
59
Expected frequency
38.8
20.2
59
Uncertain listeners
Frequency
5
1
Expected frequency
3.9
2.1
Disinterested listeners
Frequency
11
1
Expected frequency
7.9
4.1
12
Total
Frequency
69
36
105
Expected frequency
69
36
105
Passive listener
5
6
6
12
shown to be important both in terms of circulating information and in obtaining the
correct evaluation of new products.
Next, I conducted a detailed analysis of the information actually sharing on Twitter
and blogs. First, we looked to see if there was a difference in the number of people
sending out to social networking sites by listening segment. There is no difference
between segments (χ2 (4) = 0.825, p < n.s.). Comparing the average number of intraperiod posts per capita, we found that disinterested listeners has a very high number of
outgoing messages. The high number of non-influential people sending out messages
shows that we should not evaluate senders based on the number of messages alone
(Fig. 10.4). Also, seeing the average number of followers according to listening
204
10 A New Decision-Making Process—A Circulating-Type …
Table 10.11 Actual purchasing conditions by listener segment after the sampling
Already
purchased
Plans to
purchase
No plans to
purchase
Frequency
10
6
13
29
Expected
frequency
6.3
7
15.6
29
Frequency
34
42
44
120
Expected
frequency
26.1
29.1
64.7
120
Frequency
104
119
254
477
Expected
frequency
103.8
115.8
257.3
477
Uncertain
listeners
Frequency
16
14
49
79
Expected
frequency
17.2
19.2
42.6
79
Disinterested
listeners
Frequency
18
22
91
131
Expected
frequency
28.5
31.8
70.7
131
Total
Frequency
182
203
451
836
Expected
frequency
182
203
451
836
Early listeners
Active listeners
Passive listener
40
Total
***
35
30
25
20
37.59
15
28.75
25.48
25.45
20.91
10
5
0
early listeners
acve listeners
passive listener
uncertain listeners
disinterested
listeners
p<0.001
Fig. 10.4 Average number of posts per capita in the period
10.3 Overview and Results of the Experiment Survey
205
segment, while the differences in the number of followers are not significant, the
lowest were the active listeners (131) and the highest were the disinterested listeners
(323). Incidentally, the number of followers in the other segments is as follows: early
listeners (156), passive listener (214), and uncertain listeners (201). As the difference
was not significant, it cannot be stated that the number of followers is that big for
either segment. However, research on SNS has often been conducted by looking at
the number of followers of influential people. Based on the above result, at the very
least, it cannot be said that this approach is the correct one.
As it is difficult to measure postings numerically, next, I will explore the content
of what was actually posted. Among the subjects undergoing the sampling, in the
period from after the sampling up to the second survey, 174 people had posted on a
blog about the relevant new brand. I conducted text mining on the content of these
posts and classified the content. For the analysis, I used Staffs (name of software for
text mining) of IBM-SPSS Inc.
When conducting text mining, it takes a lot of time to extract the keywords and
create a dictionary. Therefore, in this chapter, I created a dictionary as much as
possible from the items used for the judgments to evaluate new product sales in
the judgments survey on beer drinks conducted in Chap. 8. In total, I extracted 35
keywords and segmented them into six categories. The newly created keywords were
“Evaluation of taste” (“the aftertaste was good,” “a clean taste,” etc.), “Comments
on usage scenes” (“good for a lunch beer,” “for when you are driving a car,” etc.),
“Comments on having the same impressions as a beer” (“it has the feeling of being
a beer,” “it smells the same as beer,” etc.), “Similarity with the beers of the relevant
company” (“the taste is similar,” “the packaging is similar,” etc.), “Low evaluation”
(“I would not buy it,” “the taste is bad,” etc.), and “High evaluation” (“I would buy
it,” “the taste is good,” etc.). The first four keywords are about the product, and
the remaining two keywords are on the product evaluation. Table 10.12 shows the
number of occurrences of these keywords.
Next, I investigated the relationship between these classified keywords and the
result variables. The analysis included 149 respondents who responded about their
purchases of the target products after the sample was distributed. We first explored
the relationship between these keywords and the actual post-sample presentation
purchases history of those who uttered them. Of the above six keywords submitted,
three of the keywords, “Evaluation of taste” (χ2 (2) = 4.645, p < n.s.), “Comments
Table 10.12 List of
Keywords
Number
Evaluation of taste
55
Comments on usage scenes
53
Comments on having the same impressions as a beer 95
Similarity with the beers of the relevant company
79
Low evaluation
37
High evaluation
64
206
10 A New Decision-Making Process—A Circulating-Type …
on usage scenes” (χ2 (2) = 1.299, p < n.s.) and “Comments on having the same
impressions as a beer” (χ2 (2) = 2.285, p < n.s.), had nothing to do with purchasing,
the other three had anything to do with purchasing. Table 10.13, 10.14 and 10.15
show the relationship between purchase history and these keywords: “Similarity with
the beers of the relevant company” (χ2 (2) = 6.791, p < 0.034.), “Low evaluation”
(χ2 (2) = 12.591, p < 0.001), and “High evaluation” (χ2 (2) = 11.108, p < 0.004.).
The person who said the key phrase “similar to the same company’s beer” is likely to
be someone who is familiar with the company’s beer and therefore appreciates and
purchases this new product as well. Also, the fact that people with lower ratings are
less likely to buy and people with higher ratings are more likely to buy proves that
they are responsible for what they say.
Figure 10.5 is percentage of keyword dispatch by segment. The Figure shows
that while active listeners has relatively more to say about the drinking scene of
the product, passive listeners has a more positive evaluation of the product, while
disinterested listeners has a more negative evaluation of the product. After the launch
of the relevant brand, this brand held a high market share in the non-alcoholic market
Table 10.13 Similarity with the beers of the relevant company
No Comments
Yes
Total
Already
purchased
Plans to
purchase
No plans to
purchase
Total
Frequency
14
18
49
81
Expected
frequency
19
20.7
41.3
81
Frequency
21
20
27
68
Expected
frequency
16
17.3
34.7
68
Frequency
35
38
76
149
Expected
frequency
35
38
76
149
Table 10.14 Low evaluation key words
Already
purchased
Plans to
purchase
No plans to
purchase
Total
Frequency
33
33
51
117
Expected
frequency
27.5
29.8
59.7
117
I do not rate this
product
Frequency
2
5
25
32
Expected
frequency
7.5
8.2
16.3
32
Total
Frequency
35
38
76
149
Expected
frequency
35
38
76
149
No Comments
10.3 Overview and Results of the Experiment Survey
207
Table 10.15 High evaluation key words
Already
purchased
Plans to
purchase
No plans to
purchase
Frequency
15
20
56
91
Expected
frequency
21.4
23.2
46.4
91
I rate this
product
Frequency
20
18
20
58
Expected
frequency
13.6
14.8
29.6
58
Total
Frequency
35
38
76
149
Expected
frequency
35
38
76
149
disinterested
listeners
Total
No Comments
Total
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
early listeners
acve listeners
passive listener uncertain listeners
Evaluaon of taste
Comments on usage scenes
Comments on having the same impressions as a beer
Similarity with the beers of the relevant company
Low evaluaon
High evaluaon
Fig. 10.5 Percentage of keyword dispatch by segment
segment, so it was noted that no matter what disinterested listener segment says, it
has little impact on actual sales because it has little influence.
This series of experimental surveys clarified that the active listener segment
certainly shares and circulates information, and that the people in this segment
affect potential customers. In other words, this single-source data confirms what
was discovered in each of the chapters in this book; namely, the important thing for a
brand to be successful is that consumers with high information sensitivity disseminate
information on it and support it.
208
10 A New Decision-Making Process—A Circulating-Type …
10.4 Conclusion and Future Implication
In this chapter, the information present in the booking is summarized, showing
conceptual diagrams of a new, information circulating-type decision-making process,
and there is information to confirm its validity through experimental data. Ultimately,
the results of these experiments do not stretch beyond their respective experiment
areas. However, through data analysis, it has been observed that instead of a one-way
consumer decision-making process, the process can be organized extremely well by
understanding it using the circulation concept. For consumers with varying levels of
interest in information, it is possible to capture it within this framework by changing
the starting point for a purchase, which is highly versatile. Furthermore, specifying
the roles of the various types of media facilitates the use of this conceptual framework in the future to formulate a communication strategy leading to the creation of
a successful product. When combining the findings of consumer behavior research
with the present, having the present-day scenario, we observe that this circulatingtype consumer behavior model is extremely effective conceptually. The experiments
in this chapter also clarified that for a successful new product, information is circulated through “active listeners,” and the words being circulated provided information
through the judgment survey where words relating to “feelings on appearance” and
“taste” are the factors that cause a product to be successful. It can be understood as
a feasible concept.
Businesses too constantly monitor whether this circulation wheel is turning
correctly. Should the performance deteriorate or become sluggish, improving the
product or modifying communication may increase the probability of the product
being a long-term seller. Information is circulated by the active listener segment with
high affective commitment, which plays the primary role in ensuring that the product
keeps selling in the long term, and so it is important to focus on the behavior of these
people. Doing this enables advertising agencies to plan media strategy based on the
primary goal of determining how to ensure that effective circulation happens. Additionally, a company may choose to integrate together media strategy positioning and
the chosen path until the product is deemed successful.
However, a single experiment cannot establish everything. One challenge for
current research is the lack of a field that can describe this information circulatingtype consumer behavior model in entirety. Data analyses on consumers till now have
relied on purchase history in many cases, and despite the achievement of results to
a certain extent, today’s information environment is drastically different. Therefore,
it will be necessary, in the future, to establish and confirm a field that can ascertain
single-source data not only on based on purchase history but also based on media
exposure and information-sharing conditions. Fortunately, in recent developments
in mobile technology, techniques are being formulated that can measure the abovementioned information circulation. Such a field does not exist overseas. In the near
future, we may accurately prove this circulation in Japan before the rest of the world.
Currently, the research style in Japan has been to use foreign research. Hence, in order
to transition to a system for developing and disseminating theory originating from
10.4 Conclusion and Future Implication
209
Japan, this book concludes that a probable way to address this situation is developing
a theory, which is a research issue that will be tackled in the future.
Notes
1.
2.
3.
Howard, John; Sheth, Jagdish (1969), The Theory of Buyer Behavior, Jon Wiley
& Sons.
Bettman, James (1977), ‘Data Collection and Analysis Approaches for Studying
Consumer Information Processing’, Advances in Consumer Research, Vol. 4,
pp. 342–348.
Petty, Richard; Cacioppo, John (1983), Central and Peripheral Routes to
Advertising Effectiveness: The Moderating Role of Involvement’, Journal of
Consumer Research, Vol. 10, pp. 135–146.
Index
A
Active listeners, 118, 126, 127, 129, 130,
132–134, 136, 137, 148, 167–170,
174–177, 180, 191, 193, 195–197,
202, 205–208
Affective commitment, 61–63, 65, 67–73,
75, 80, 82, 83, 85, 86, 93, 97, 98,
101, 102, 104, 105, 107, 109, 113,
114, 117, 169–171, 175, 182, 192,
200–202, 208
AIDEES, 26, 36, 37, 39, 62, 67, 75, 76, 94,
191, 194
AIDMA, 25, 26, 189, 191
AISAS® , 26–29, 31, 36, 37, 60, 67, 75, 76,
134, 137, 169, 191, 194
Attitude, 58, 61, 62, 71, 74–78, 80, 93, 94,
97, 103, 122, 176–179, 186, 191–193
B
Blog, 20, 24, 28, 107, 137, 147, 191, 192,
196, 201–203, 205
Brand, 38, 39, 74, 94, 138, 139, 163, 164,
187
Brand loyalty, 133
C
Calculative commitment, 61, 62, 69, 70, 125,
169–171, 173, 201, 202
CANVASS, 15, 124, 125, 128, 130, 131,
167–169
Central route, 19, 76–83, 85–90, 93, 94, 97,
109, 175, 176, 192, 193
CHi-squared
Automatic
Interaction
Detector (CHAID), 89, 93, 99
Circulating-type communications model,
189
Communication-oriented consumer, 15, 97,
98, 101–104, 109, 110, 112–115,
117, 170
Connoisseurs, 141, 142, 150–154, 157–162
Consideration set, 110, 194
Customer
Relationship
Management
(CRM), 98–102, 104, 107, 109,
112–115, 177, 191, 202
D
Decision-making process, 7, 17–21, 25–27,
29, 75–78, 80–82, 85–87, 94, 134,
148, 167, 169, 174, 176, 186, 187,
189–193, 195, 208
Disinterested listeners, 126, 127, 129, 130,
134, 136, 167, 168, 170, 180, 202,
203, 205–207
E
Early adopter, 145
Early listeners, 126, 127, 129, 130, 136, 167,
168, 170, 180, 205
Elaboration likelihood model, 19, 75, 76, 94
F
Facebook, 4, 20, 24, 27, 79, 123, 192
I
Influencer, 29, 30, 36, 39, 87, 94
Information processing model, 17–21, 26,
27, 31, 36, 37
© Springer Nature Singapore Pte Ltd. 2021
A. Shimizu, New Consumer Behavior Theories from Japan, Advances in Japanese
Business and Economics 27, https://doi.org/10.1007/978-981-16-1127-8
211
212
Information sharing, 20, 26–28, 31–37, 60,
62, 67, 73, 97, 102, 103, 107–109,
167, 169, 190, 192, 193, 195, 196,
200–202
Innovator, 136, 145, 146, 148–151, 159, 160,
163, 165
Interest, 7, 10, 23, 25, 26, 28–32, 36, 37, 58,
60, 61, 65, 75, 80, 81, 87, 94, 97, 119,
120, 123, 126, 129, 134, 136, 138,
145–148, 156, 168, 173, 186, 192,
195, 199, 208
K
Kikimimi, 167, 170, 196, 197
L
Lead user, 149–151, 159, 160, 164
Lifestyle, 117–122, 125–131, 136–139, 141,
145, 158–160, 162, 167, 168, 189
Life time value, 97, 99–101, 115, 191
List Of Value (LOV), 120, 121, 138, 139
Long selling brand, 63–67, 72, 73, 86, 97,
105, 109, 110
M
Market maven, 23–25, 27, 38, 134, 148–151,
159, 160, 164, 165, 168
Media value, 4, 8, 12, 15
O
Opinion leader, 23, 146–151, 159, 160, 164
P
Passive listeners, 126, 127, 130, 136, 167,
168, 170, 180, 205, 206
Peripheral route, 19, 38, 76–83, 85–90, 93,
94, 175, 176, 192, 193, 209
POS, 63, 69, 70, 72, 101, 115, 125, 131–133,
137, 162, 194
Promotion, 12, 16, 18, 29, 57–63, 65, 67–74,
99, 134, 137, 153, 169, 176, 182, 192,
194
Index
R
Recognition, 7, 20, 22, 23, 25, 27–32, 36–
38, 60, 61, 75, 78, 87, 97, 143, 173,
174, 177, 186
Reference price, 58–60, 70–73, 137
Relationship marketing, 61, 74, 114
RFM, 60, 191
ROM, 23
S
Satisfaction, 30–37, 39, 40, 57, 68, 69, 72,
75, 78–80, 94, 97, 98, 100, 146, 191,
195, 200–202
Segmentation, 15, 17, 74, 113, 117–119,
121, 122, 124, 126, 136, 138, 139,
167, 187, 189–193
SIPS, 27, 134, 191, 194
Social Networking Service (SNS), 4, 7, 8,
11–13, 20, 24–27, 36, 37, 79, 93, 100,
101, 104, 123, 192, 194–196, 200,
205
Stimulus-response, 17–20, 27, 29, 36, 65, 76
Structural equation modeling, 30, 70, 73, 154
T
Twitter, 4, 20, 24, 27, 79, 123, 192, 196,
201–203
U
Uncertain listeners, 117, 118, 126, 127, 130,
131, 133, 134, 136, 137, 167–170,
176, 180, 189–191, 194, 196, 202,
205
V
Values & Lifestyle (VALS), 120, 121, 139
W
Word of mouth, 20–23, 25–27, 30–32, 37–
39, 65, 78–80, 82, 85, 94, 114, 165,
175