Document 11049643

advertisement
HD28
.M414
no.
ALFRED
P.
WORKING PAPER
SLOAN SCHOOL OF MANAGEMENT
Multi-Project Strategy and Organizational
Coordination in Automobile Product Development
Kentaro Nobeoka and Michael A. Cusumano
MIT Sloan School of Management
WP#3487-92/BPS
Date: November 5, 1992
MASSACHUSETTS
INSTITUTE OF TECHNOLOGY
50 MEMORIAL DRIVE
CAMBRIDGE, MASSACHUSETTS 02139
Multi-Project Strategy and Organizational
Coordination in Automobile Product Development
Kentaro Nobeoka and Michael A. Cusumano
MIT Sloan School of Management
WP#3487-92/BPS
Date: November 5, 1992
M.I.T. LIBR^.u.
NOV
1
3 1992
1.
Introduction
Since the management
numerous academic
competition,
effective
of
and
new
product development lias
researcliers
have been
efficient projects
in
central issue
recent years have undertaken studies of
various industries.
in
become a
Most
in
global
how
of the empirical research
has
focused on the innovation process and on managerial or organizational approaches as well as
performance measures
Fujimoto, 1991
;
for individual projects (Imai et
Cusumano,
1991). At the
Japanese manufacturers tend
European competitors and
skills, for their
et
1988;
al.,
in
that this
Womack et
al.,
efficiency.
In
in
of the
much more
major reasons, along with manufacturing
time,
products and variations more
over time at the firm
in
and manufacturing so many new products and product
money
of
Japan has increased due
in
efficiently is
level.
individual projects
This
may
is
the
management
products unnecessarily, requires
or
and drops
essential missing area in
how
of multiple
to
produce multiple
new-product development
important because, while high levels of engineering
contribute to
designs and components
demand has slowed
to rising interest rates
An
making a
firm overall
development, to devetop a successful stream of new products over
of
1985; Dertouzos
recent years they have faced severe profitability problems related at
research on product development that relate directly to the issue of
advantage
Stalk,
even Japanese manufacturers became more
the stock market and real estate values (Business week, 1992).
productivity
and
markets, such as automobiles and consumer electronics, where
even declined, while the cost
efforts
frequently than U.S. or
global markets (Abegglen
same
1990). At the
part to the high costs of developing
variations
in
has been one
in
1985; Henderson, 1990; Clark and
time, there are various reports that leading
develop new products
strong growth performance
concerned with
least
to
same
al.,
in
more
many
efficient in
product
years, as well as to take
more than one product without compromising the
some degree
of planning
and coordination above the
final
level of the
individual project.
In
industries
where manufacturers
multiple projects in parallel
two elements
new
offer multiple products to the
market and undertake
product development strategies and organizations must take at least
into consideration.
Rrst, they
need
to plan for the
frequency
of
new development
and
projects, both to replace existing products
(Utterback and Abernathy, 1975;
to
expand the breadth
1988; Kekre and Srinivasan, 1990; von Braun, 1991). This
Miller,
frequency becomes a central competitive dimension because
much more
prolific in their
related they
want products
the coordination process
new
prefer to
1991
;
among
such as
et
1990).
al.
terms
in
of
components
of different products that
each
in
same
share the
little
be
reflect
decisions
how
and manage
basic design, while others
(Clark and Fujimoto,
made above
yet they affect not only the project organizations but also a firm's competitiveness.
there has been
to
some manufacturers
new products
of their different
These differences may
or design features,
For example,
multiple projects as necessary.
use unique designs more often
Womack
some manufacturers appear
product introductions than others. Secondly, firms need to plan
to be,
develop an extensive number
of available product lines
the project
level,
Nonetheless,
empirical research that explores the interrelationship of these factors
and
their
impact on either market or organizational performance.
Numerous
performance
for
new
and Nobeoka 1992
University
and the
differences
(Clark et
al.,
in
studies
in
recent years have examined differences
product development
for
a detailed review of
International
among worldwide
this literature).
Motor Vehicle Program
at
Womack et
al.,
strategy, structure
auto manufacturers (see
In particular,
and
Cusumano
Clark and Fujimoto at Harvard
MIT have found several important
management and performance among Japanese,
1987; Sheriff, 1988;
in
U.S. and European manufacturers
1990; Clark and Fujimoto, 1991). Clark and
Fujimoto conducted the most thorough study, focusing on 29 projects from 22 producers. They
concluded that the Japanese
measured by design
quality, lead time,
volume producers, three
manager
firms, in general,
were
new
better at
product development as
and productivity defined by engineering hours.
Among
factors also contributed to better project performance: heavier project
responsibility, higher supplier involvement in engineering,
and more overlapping between
stages such as product planning, product engineering, and process engineehng. Clark elaborated
on these data
in
a 1989 paper that focused on the
result
showing
unique parts than U.S. or European firms, which theoretically
add time and cost
in
development, unless
fitting
that
Japanese projects used more
may inaease
old parts into
new designs
design quality but also
creates additional
coordination that increases engineering time (Clark, 1989).
had more unique parts and higher engineering
counterparts primarily because they
sample consisted
Fujimoto's
project-level analysis
of
and comparisons,
between
project-level
each
firm,
is difficult
performance and
and European
they limited their study to a
with statistical analysis, of regional
it
Japanese projects
that
extensive use of suppliers. Since Clark and
projects from
European, and U.S. producers. Therefore,
linkage
productivity than their U.S.
made more
one or two
He concluded
from
this
sample
performance
firm-level
averages
for
to generalize
Japanese,
about the
the marketplace. Nor were
in
they able to explore the potential impact of different inter-project strategies and
management
approaches on organizational and market perfonnance.
As
part of the
introductions
1987
(Sheriff,
MIT
and average
study, Sheriff
measured differences
project complexity for
1988; also reported
in
Womack
et
for
each
data confirmed that Japanese firms introduced
European
fimns.
As a
result, the
Japanese
al.,
Project complexity
1990).
made
additional
in
major exterior,
body
that the
firms maintained
product
was
interior,
wheelbase
calculated
and platform
variation.
These
frequently than U.S. or
much newer products
in
In addition. Sheriff's
the market and
measurements
European projects had the highest average complexity, followed by the Japanese
and then the U.S. producers. Fujimoto and
interrelationships
style or
new products much more
increased the number of product offerings during this period.
showed
new
the frequency of
25 major auto manufacturers between 1982 and
through an index that assigned weights to changes
components, with adjustments upward
in
and found
Sheriff then
positive correlations
as well as engineering hours
at
compared
their
data to explore
between productivity measures such as lead time
the project level and the performance variables at the firm level
(Fujimoto and Sheriff, 1989). They also found a positive correlation between the rate of
product introductions and market-share growth, although
project complexity
this
new
paper did not explore the impact of
on market performance.
The purpose
of our
study
is
to build
on
this
research and explore product-development
strategy and organization for multiple projects within the firm.
The underiying hypotheses are
that.
.
apart from differences
project strategy
firm
new
is in
organizational performance for individual projects, differences
in
and management should
product development, and
share or sales growth
if
significantly influence
this effectively
how
efficient
and
should have an impact
a firm introduces more and newer products
competitors. Specifically, this paper examines two questions.
into
First,
in inter-
an entire
effective
at least
on market
the marketplace than
we propose a
typology of
product development projects based on inter-project linkages to discuss an effective multi-project
strategy
in
between
in this
the market.
In this section,
multiple projects within the firm
way firms can
analyzing 223
new
Secondly,
2.
we
and
financial resources.
multiple projects within firm
Japan and three
in
in
is
We
and can
investigated this question by
manage
interactions
of
225 engineers
New Product Management and Research Questions
project within a firm
technically
the market, because
the U.S.
products to replace existing products or to add
organizationally.
in
based on a questionnaire survey
Large automobile manufacturers have several product
Each
interactions effectively
create competitive advantage
discuss the organizational coordination required to
automobile firms
Firm-Level
managing concurrent
car products introduced at 21 automobile manufacturers between 1980 and
between concurrent
at six
may
that
transfer technology and design quickly across multiple projects
effectively leverage their engineering
1990.
we argue
new
lines
and constantly develop new
product lines over time as
shown
in
Rgure
1
has some linkages with other projects both technologically and
Managing the way
different projects interact organizationally or relate to
by no means a simple matter.
Consideration of multi-project
management
each other
includes
both linkages between different product lines and linkages between past and present projects.
example,
some
projects
use the basic design framework
designs from other product
lines,
and design from saatch. A
different
have
different influence
while
some
way
projects
of interaction
of their
previous models and others use
may choose
to
develop a new technology
a project chooses with other projects
on the new product competitiveness and may impose
requirements on the project management.
In addition,
For
because managing the
may
different
inter-project linkages
effectively
may
require an extensive integration across a firm,
inter-project linkages
may have an
tfie
patterns firms choose regarding
influence on their organizational competitiveness as a whole.
order to simplify the complicated multi-project strategy and
management
at
the firm level, the next
section proposes an analytical framework by decomposing the multi-project patterns into four
different types of inter-project relations.
Rgure
1
.
Dimensions
of Multi-project
Management
Evolutional Linkage (Time)
n
Inter-Project
Linkage
Firm/
In
Rgure
2.
Typology
of the Inter-project Strategy
— —
New
Type
*^m
New Design
1:
Project
Weiv Project
Type
Concurrent Design Transfer
2:
/
On-going Other Project
Uew
*^m
^
Type
—
Project
Sequential Design Transfer
3:
Past Other Project
Weiv Project
Type
Design Modification
4:
Predecessor
The
first
type,
New design,
thus refers to the development of a
new
design produced primarily from scratch, without a preexisting base design.
there
is little
project
relatedness or interaction with any other projects within the
product with a core
In this
firm.
Members
can concentrate on creating a new design and a new product. While the
engineering task requirements should be high because the design
is
new
type of project,
of the
project's
(Clark, 1989), both
coordination costs and design constraints are low because the project does not have to be
coordinated with other projects or follow design constraints derived from an existing design base.
The next two types
the firm.
In tfie
second
of projects transfer
and share a core design from other projects
type, concurrent design transfer, a
new
design from a base project before the base project completes
-
new
project begins to transfer
its
and
potentially costly
coordination because (1) they must overlap chronologically, (2) the
new
project
the
project
and the base
project
--
incorporate a design from the base project while the design
is still
a core
design engineering. These two
require extensive
projects
within
needs
to
under development or
relatively
new, and thus
perhaps
mutual adjustments
(3)
in
design between the two projects are
possible and
still
likely.
The
third type, sequential
base model's development
design that
is
core design
compared
finished.
is
Because
this
type of project basically reuses an existing
"off-the-shelf," inter-project coordination
in this
to
design transfer, transfers a design from a base model after the
is
not needed.
manner, however, the design being transferred
designs transferred more concurrently.
In addition,
is
When
new
a
direct
last type,
design modification, refers to a
predecessor product as
need any
is
a
relatively
The
difference
may be
design constraints
new development
high
possible. ^
project that modifies a
minor model change. This type of project does not
inter-project coordination either, but
the current model.
transfer
in
uses the
already relatively old,
because mutual adjustments between projects on the core design are no longer
The
project
has
to consider constraints
from the core design of
between the design modification and the sequential design
thus the source of the base design. Modifications
in this
type
may be
easier than with a
sequential design transfer, which transfers a core design between different product lines. Another
difference
is
modification
One
that sequential design transfer
is
can be used
add a new product
line,
while a design
only for replacement projects.
of the strong points of this typology
strategy type, which
the difference
to
is
the age of the core design a
new
time between the introduction of the
in
which the product was based was
utilizing transfer
scheme
strategies
is
introduced.
new
new design
it
determines design age of each
project uses.
Design age
new product and when
For example, design age of a
strategy
is
new
project that develops
the smallest, zero. Design age of a
is
determined by
the original design on
new
product
was introduced
project using the concun-ent design transfer
design transfer. Design age of a
scratch using the
that
the time that has passed since the base product
market. Thus, design age of a
of the sequential
first
is
is
to the
smaller than that
a core design from
new
product using the
1. This discussion of hypothetical differences between concurrent and sequential design transfer
are partially based on Thompson's distinction between "long-linked technology" and "intensive
technology," where the latter also requires mutual adjustments and higher coordination costs. See
Thompson, 1967
design modification strategy
is
the
same as
the product
can even be older than with a sequential design
3-2.
New Product Rate,
number
transfer.
rate here
defined as a ratio of the
is
of product offerings in a
increase the
(Miller,
new product
rate, firms
result in products that suffer
do not want
of
new product
need
to invest
to or
from problems
cannot increase
rate requires a decrease
in
more
may
financial
reflect
In
order to
incomplete development efforts and
in
the market.
resource investments, then increasing the
their
it
and engineering resources.
design quality and perfonn poorly
in
introductions
new market segments more frequently than
1988; Fujimoto and Sheriff, 1989; Kekre and Srinivasan, 1990).
Otherwise, frequent new-product introductions
firms
number
base year. A higher new product rate makes
possible for a firm to replace existing products or enter
competitors
predecessor model, which
its
Average Design Age, and Different Types of Inter-project Strategy
The new product
adjusted by the
cycle of
life
new
If
product
task requirements, because as Clark (1989) illustrated, a project that
develops more new components generally requires more lead time and engineering hours. Thus,
may
not
be a reasonable choice
for
a
pursues a high new product rate to
firm that
utilize
the
it
new
design strategy, extensively.
Firms have
at least
decrease the average
repeat the
same
two choices
to
decrease engineering tasks
such as the number
intra-project variations
design
among
different
new
a negative impact on market competitiveness
projects.
if
of
also
new components:
body and
styling types, or
Decreasing intra-project variations
same
design or components
have a negative impact on market competitiveness, because the purpose
product introductions
an old design may
is
to capture
changes
in
conflict with this objective.
new technology may
we can
competitiveness should depend to
of frequent
of
However, the rapid reuse among multiple projects of
newness
or technological sophistication of
a
hypothesize that the negative impact on a firm's market
some
new
customer needs with new technology, and reuse
actually improve the overall
product offerings. Therefore,
may have
products appear too similar to consumers and lead to
a reduced coverage of market segments. The repeated use of the
may
for
extent on the average design age of
new products
firm's
introduced into the marl<etplace. There should generally be a tradeoff between increasing the
product rate and incorporating
the
same
new designs
into
each new product, rather than extensively reusing
design. Successful automobile firms regarding market share growth
new products
without introducing older designs than their counterparts.
example, implied
that, in
some
order to avoid this tradeoff,
manufacturers depended more on outside suppliers
Our study explores the idea
in
new
product rate.
may
technologies
both increase
in
A new
have
also
different inter-project
of the transfer strategies,
same
in
new product
Hypotheses on
on
concurrent design
developing more
technology or design
is
new
shared by
the product than those using concurrent design transfer. Therefore,
rate
low average design age but
3.
may
(Clark, 1989).
product using sequential design transfer necessarily incorporates older
and keep average design age
concurrent design transfer strategy.
Rgure
Japanese
new component designs
provide firms with an advantage
products than the other two project types, because
multiple projects.
of Clark's findings, for
influence of different inter-project types
An extensive usage
transfer or sequential design transfer,
may develop more
order to mitigate this tradeoff.
Rgure 3 shows hypotheses regarding the
design age and
for
One
of the successful
that successful manufacturers
strategies from low performing manufacturers
new
An extensive use
may have
Inter-project
High
Types
Modification"
\
Average
Design Age
Low
(Design:
New)
a
fimri
new design
a negative impact on the
"Design
(Design: Old)
of
low,
in
order to
may choose to use
strategy
new product
may end up
rate.
with a
3-3.
Sample
Characteristics
The sample
study covers the 21 largest auto manufacturers
in this
seven Japanese, three
and Measurements
U.S.,
Data on new product development
in
the industry were
in
Auto Review an annually published industry joumal
,
features and introduction dates for
engineers
the world, including
and eleven European producers, and the 223 new car products they
introduced between 1980 and 1990.
primarily collected from
in
all
new products
covers design
that
worldwide. Unstructured interviews with
these firms were also conducted when needed
for clarification.
We used the
cumulative worldwide market (unit-production) share growth of each manufacturer during this period
and low performers.
to categorize firms into three groups: high performers, middle performers,
Inter-Project Strategy
As a firm-level
individual firms during
and A verage Design Age
analysis of inter-project strategies,
1980-1990
project's inter-project type
and
measured the new product
introductions
defined a
and
its
we
classified
into four inter-project strategy types.
base model,
rate for
its
design age
is
each manufacturer by the
new
the
By
this definition,
of
offehngs
new
in
show
variable for intra-project variation to
developed within individual
projects.
the average
projects
and the nature
and Sable, are openly discussed
to rely
in
number
is critical to this
variations
study,
of their interrelationships.
product
new
is
we used
body types and
were
because
We
1979.
interior
not
such as the Ford
the other hand,
of different
Whether two or more new
together within one project or separate projects
new
On
project,
in fact
another
stylings
developed
this affects
the
11
total
Most cases, such as the Taurus
Auto Review or other industry journals. For unclear cases
on interviews with company engineers.
We
earlier.
a new product with minor cosmetic modifications
Taurus and Mercury Sable, count as only one new product.
have had
number
of product
counted as a new product. Product variations designed within a single
of
a new car
identifying
product as a model designed within a single project and with completely
exterior stylings.
number
By
car projects done within
determined as defined
ratio of
between 1980 and 1990 divided by the number
new
we
The core design used
in
the present data analysis
is
the platform design, which determines
the basic characteristics of other major component designs, including the body and engine size,
drive-train type,
and the general
level of
design sophistication. Designing a
scratch requires both financial and engineering resources as well as
determine whether the platform
preceding products,
new
we
new
project
type,
new
one
well
new
of the other three categories.
other three types, design age
is
based on changes
new products based on
Design age
measured by the
the base product
or transferred from
platform design
1
for
order to
In
more
between the
in
the
details).
platform designs are automatically categorized as the
design, while those developing
when
product,
in
as the suspension design (see Appendix
projects that developed
product and
changes
platform from
technology.
was newly developed
of
new
product and preceding products similar to the
New
new
a certain
assigned points to the extent
wheelbase and tread as
into
of
new
new
was
first
in this
type
difference
is
in
platforms from other projects
zero as defined
earlier,
while
in
fall
the
time between the introduction of the
Projects that developed a
developed.
first
new
product based on the platform design of the predecessor model are categorized as design
modifications, while those which shared platfonn designs with any preceding projects are either
concurrent design transfers or sequential design transfers.
between concurrent and sequential
same as
transfers
is
As
indicated earlier, the distinction
determined by the transfer time
which
lag,
is
the
the design age defined above.
Rrst,
projects that
we compared
the average transfer time lags for the Japanese, U.S., and European
were not new designs or
modifications.
We then
defined concurrent design transfers
as a transfer between two projects occurring within 2.0 years of the introduction of the core design.
Visual analysis of a frequency distribution for
transferred designs within
years (Appendix
and
is
dose
2).
to the
The
1
projects indicated that
.25 years, while another group transferred
figure 2.0 years
midway
new
was
also the
point (2.25 years) for the
median
by using 1.5 years and 2.5 years as
:
73).
cutoff points, with
12
of projects
most designs
after
2.25
transfer lag time for the entire
average lead time (4.52 years)
development as calculated by Clark and Fujimoto (1991
division
one group
for
new
sample
car
We also tested the sensitivity of this
no
significant
change
in
the results.
we
In addition,
much
believe that
if
the time lag
is
longer than two years, then there does not
need
to
be
overlapping or coordination between projects.
3-4. Results
Rrms
and Discussion
in
the high
new product
rate/small average design
market share than the others as shown
in
Rgure
4.
age region tend
Low performers
including
two European firms developed fewer products with older designs than most
to
all
have gained more
three U.S. firms and
of high performers.
Middle performers dominated by European firms tended to develop fewer products with newer
designs.
Rgure 4 also suggests
that
European and U.S.
firms, but not
Japanese
firms,
experienced a tradeoff between the new product rate and average design age.
Rgure
4.
New
Product Rate and Average Design Age
Average
Design Age g
2
4
3
New Product Rate
^
ZA
(J:
:
High performers
:
Low performers
Table
1
---
performers
Market share increase by
utilized
firms, E:
European
firms)
35% or higher
Market share deaease
summarizes differences
1980 and 1990. Three
Japanese firms, U: U.S.
in
usage
of different inter-project strategy types
different patterns are evident in
each group. Most importantly, high
concun-ent transfer strategy more than the other firms, which,
contributed to both higher
new
words, instead of developing
product rates and relatively
many
completely
new average design
new core designs
13
between
to achieve
we
believe,
ages.
In
other
these two key
objectives, high performing firms created a few
new core designs and
other product lines, while the designs were
relatively
firms
used concurrent design
Table
transfer, at least
project from which the core design
Comparison between
1.
Number
Number
of
Manufacturers
of
New
Inter Project
46%
new. Since
of their projects
23%
of
new
may have
projects at these
required extensive
because each concurrent transfer involves some overlapping with
inter-project coordination
one other
still
quickly transferred these to
is
U.S., European,
at least
transferred.
and Japanese Manufacturers
Low
High
Performer
6
Middle
Performer
9
84
78
6
61
.44 (.09)
.23 (.05)
.69 (.07)
.09 (.04)
.49 (.09)
.05 (.04)
.16 (.08)
.17 (.06)
.16(06)
.29 (.08)
.18 (.06)
Projects
Performer
Strategy (%)
New
Design*
Concurrent Design Transfer**
Sequential Design Transfer**
Design Modification
New Product Rate**
Average Design Age***
Average Intra-project Variations
Market Share Growth***
(Standard deviations are
Statistically Significant at:
in parentheses)
***
**
1%
The middle performing
Level,
5%
Level,
new
products.
3.58 (.52)
1.62 (.43)
2.09 (.19)
1.78 (.42)
1.81 (.52)
0.81 (.35)
1.89 (.16)
3.00 (.43)
2.14 (.19)
.78 (.11)
.15 (.09)
-.24 (.11)
*
10%
Level (One-way
ANOVA)
between 1980 and 1990, mostly European makers, are
firms
characterized by an extensive use of completely
design age of their
.06 (.05)
new
designs, which explains the low average
Even though they developed fewer new products, by
concentrating on these products, they developed newer designs than other producers.
Low
performing makers tended to have more sequential design transfers than other firms, which
resulted
in
older designs
in their
new
products as high performers or as
products. Accordingly, they did not develop either as
many new designs as
middle performers. These
is
no
many
significant
difference in average intra-project variations, which suggest that increasing the complexity of
project
by adding variations
in styling
and body types did not have a
significant
each
impact on market
share.
We realize that variables here are too limited to predict market performance.
for
example, ultimately should
result
from the
ability of
14
a
firm to design
and
Sales growth,
build products that
customers want
to buy,
availability, service,
framework on
and
price-performance, advertising, product
this relates to quality,
and other
factors.
inter-project strategy
Our primary
and
to
show
intention here
is
to
propose a conceptual
that high performing firms
seem
to
adopt a different
product-development strategy that also has specific organizational implications. Specifically, high
performing firms
to both higher
seem
new
to transfer
new designs among
product rates and relatively
multiple projects quickly, which contributed
new average design
ages.
We
also
assumed
that in
order to implement concurrent design transfer, two or more different projects have to coordinate
with
each
other.
In
the next section,
we examine
organizational requirements to
inter-project interactions.
15
manage
concurrent
4.
Organizational Requirements for Managing Inter-proiect Interactions
4-1. "Project Coordination"
One
and "Functional Coordination"
of the central issues in
managing a new product development organization
complicated product such as an automobile that consists
functions
is
coordination
numerous studies
among
different
many
different
components and
groups within the organization. There have been
focused on importance
that
of
a
for
of coordination in
a
new product development
organization, although few study explored the influence of inter-project interactions on coordination
requirements
we
section,
1977; Tushman, 1978; Galbraith, 1982; Clark and Fujimoto, 1991). This
(Allen,
start with discussing
conceptual frameworks and empirical findings from past studies,
even though they do not consider
for
inter-project interactions,
because they are necessary as a basis
our later discussions regarding the influence of the inter-project interactions.
Product development organizations
goals:
One
manage
is
to
manage
for
a complicated product basically have two major
the organizational inputs of technical knowledge and the other
organizational outputs of designs for
new
products.
In
is
to
order to increase the quality and
quantity of inputs of technical knowledge, a high degree of coordination around technical
component as
specialties including
On
the other hand,
in
order to integrate
degree
of coordination within
1985).
Managing each
central issues in
New
well
of
as functions such as design and manufacturing
all
and around a
needed.
inputs toward well-defined outputs effectively, a high
project
is
needed (Marquis, 1965;
Galbraith, 1974; Katz,
these two types of coordination and the balance between them are
managing product development.
product development organizations
at large
automobile firms appear to be a matrix
organization to take care of these two types of coordination (Clark and Fujimoto, 1991
shows a
is
simplified
model
of
in
functional
manager
Figure 5
such a matrix product development organization, positioning design
an engineering function
engineers
).
at
the center of the matrix. Design engineers work both for a
primarily regarding technical or
component issues and
regarding the integration of inputs and intermediate outputs into
engineers formally or informally interact with engineers
16
in
final
for
a project manager
products.
other functions
In addition,
who work
for the
many
same
new
product project to integrate technical outputs across functional areas.
also want to maintain a
including those
dose working
who work
for other projects, to
"Other engineering functions"
manufacturing engineers.
relationship with engineers in the
in
In this
framework, project coordination
technical discipline,
manager and her
is
components and
defined as the degree of
staff
as well as engineers
Functional coordination refers to the degree of coordination
in their
same
for other project
teams.
Figure
Product Development Organization and Definition
Engineering
Other
Engineenng
Function
Funcrions
Project
in
other
between
technical function
of "Project
"Functional Coordination"
Manager
& Staff
may
update and refine "state-of-the-art" technologies.
engineers and a functional manager as well as engineers
5. fvlatrix
same
Figure 5 consist of design engineers of other
coordination between engineers and a project
engineering functions.
Furthermore, they
who work
Coordination" and
functional organizations as well as of matrix organizations.
(1965), by investigating thirty eight
conducted the
autonomy
first
R&D
For example, f^arquis and Straight
projects under contract with a
government agency,
extensive study regarding this issue. Using two dimensions
manager, and the form
of the project
the authority and
-
of organizational reporting relations
-
they
categorized the form of project organizational structure into project, functional, and matrix
organizations.
They concluded
that functional organizations tend to
performance, while project organizations tend to be more successful
and Gobeli (1988) conducted a mailed questionnaire survey
variety of industries including pharmaceutical, aerospace,
United States.
teams tend
to
They found
all
project
R&D
projects
in
technical
cost and lead time.
540 development projects
and computer
in
both
Larson
a
in
Canada and
the
schedule, cost and technical performances, project-oriented
nine major U.S. organizations to examine the relationship
performance and the
concluded that performance reaches
relative influence of project
highest level
its
when
and functional managers. They
organizational influence
influence over technical details of the work
manager and
the project
in
in
be more successful than function-oriented organizations. Katz and Allen (1985)
studied eighty-six
between
that in
for
be more effective
is
centered
is
centered
in
the functional
in
manager.
In
these empirical studies, project-oriented structures, rather than function-oriented
structures, resulted in higher performance, especially in cost
functional orientation
was
Yet,
it
is
inter-project interactions
in
can allow firms
new technology across
impose a new dimension
frameworks
some cases
in
design or engineering either conceptually or
important to study the influence of inter-project interactions on organizational
requirements, because as discussed
transfer of
in
appropriate for technical performance. However, no study has explicitly
treated questions of inter-project interactions
empirically.
and schedule, while
of
of past studies
the
first
part of this paper,
an effective management
to leverage their engineering resources
multiple products.
In addition,
because
by
of the
facilitating
quick
inter-project interactions
contingency on product development organizations, the findings and
may have
to
be modified.
18
In thie
next section,
we
discuss the potential
influence of inter-project interactions on organizational requirements for project and functional
coordinations
in
new
product development organizations.
4-2. Hypothieses: Inter-project Interactions, Organizational Coordination,
Based on the past studies discussed above, we
may have
interactions, project coordination
performance such as cost and schedule.
tfiat,
without inter-project
a particularly strong positive influence on operational
In addition, functional
project coordination regarding technical performance.
shows possible
hypotfiesize
and Performance
On
coordination
may be
as important as
the other hand, the model
in
Figure 6
influences of inter-project interactions on the degree of organizational coordination,
which are indicated by the dotted
develops a design
in
In this
lines.
model, an engineer
conjunction with an other project,
in
the
new
which the engineer
in
product project
is
not directly
involved.
In this
case,
it
two projects regarding
design
may
engineers
in
not be the
is
assumed
that there is
at least this particular
same between
may
may be needed
also have to be well
other words, the degree of functional coordination
product performance
in this
some
extent between these
component design. Requirements
these two projects. Therefore,
these two different projects
successful. This coordination
an interdependency to
for the projects
managed by
may have a
kind of design work than
in
some
for the
component
coordination between
and the products
to
the functional manager.
be
In
stronger influence on project and
a project without any inter-project
interactions.
In addition to this direct
requirement for the functional coordination between engineers
the two projects, requirements for intra-project coordination
without inter-project interactions.
A
may
also be higher than
product development project
is
in
projects
a system consisting of closely
coupled multiple engineering functions (Rosenberg, 1982). Uncertainty
in
part of the
system
increases requirements of coordination as a project (Rosenberg, 1982; Tushman, 1979).
case, uncertainty
interactions,
in
the engineer's task
because
of the
is
higher than that
in
in
In this
a project without inter-project
interdependency with the other project. For example, suppose there
19
is
a design change caused by interdependency between the two projects. The change must be
incorporated into a
hypothesize that
final project,
in
which
may
require coordination within the project. Therefore,
a component design that has interactions with another project, the influence of
the project coordination on design performance
project interactions.
is
also stronger than
in
projects without inter-
Therefore, requirements for both project coordination and functional
coordination around the engineers
may be
significantly higher in projects with inter-project
interactions than in those without inter-project interactions.
Rgure
6.
we
Influence of Inter-project Interactions on the Degree of Coordination Requirements
Other
Project
Manager
&
Staff
Engineering
Engineering
Function
Functions
purpose
we
of this study is not
a comparison of performances between
believe that this doesn't affect the theoretical discussion
in this
the questionnaire, each respondent chose one specific
In
worked on
a specific product development
for
general use.
One
development project
that
was
research.
component
project, rather than basic
asked whether there was
of the questions
U.S. and Japanese firms,
tfie
at least
he or she
that
research or components for
one other product
using similar component technology or designs
in
conjunction with
the project for which the respondent worked. Respondents were asked to think only about other
projects in which they
appeared
difference
to
have
were not
at least
one other
between the two
Among 225 component
directly involved.
project with which they
interacting projects in these
and the mean was 9.6 months. 13
of the
had
we
inter-project interactions.
32 U.S. component developments and 93 out
193
of
with inter-project interactions.
analyze data separating these two sample groups to explore
organizational requirements differ between these two
Time
responses ranged from zero to 28 months
Japanese component developments are categorized as those
following sections,
developments, 106
In
the
how
component development types.
Performance Measurements
The questionnaire asked respondents
component development performed above
quality,
and the degree
averaged
to
measure design
Measurements
There
coordination
focused
match
with
of
is
of
schedule, cost, design
customer needs. Cost and schedule performance data were
performance
(principal
component loading =
0.83).
(principal
component loading =
0.87).
Degree of Project and Functional Coordination
not a single best
among
in
design quality and the degree of match with customer needs were averaged
quality
ttie
on a 7-point Likert-type scale whether each
or below their expectation
measure the operational performance
Performance ratings
to
of
to rate
measurement
of the
degree
different groups, rather than specific
in this particular analysis.
The degree
of
of coordination.
means
The degree
of coordination,
needs
communication has often been used
21
in
to
of
be
the past
studies to
measure coordination
communication
needed
be an
to solve
in
not a
good measure
problems or
This
of the
degree
The degree
conflicts.
However, the degree
when communication
of coordination,
among
of goal sharing
of
different
is
groups could
as Lawrence and Lorsch (1967) used as a measurement of the degree of
alternative,
integration.
are
is
1977; Tushman, 1978).
(Allen,
is
not a
good measurement
in this
because
study, either,
a specific new product development project and there
may
not be
all
groups
enough
in
a response
variations in their
goals. Thus, in this study, the degree of satisfaction in working relationship on a particular
engineering task that each respondent chose
between
was used as a proxy
Respondents rated the
different groups.
engineers
in their
same
in
in different
to
degree
0.86).
coordination with a project
and
with other engineers
in
the
in
component loading =
in
shown
other functions (principal
Rgures 5 and
6,
were calculated. The degrees
of
in
the
other functions were averaged into the
0.86), while the
same
and manufacturing engineers
engineers
the model
with engineers
component loading =
functional coordination (principal
in
functional coordination
manager and
project coordination (principal
manager and
of coordination with
Secondly, as indicated
of project coordination
functional
groups: a functional manager,
technical function working for other product projects.
measure the degree
component loading =
of coordination
other functions, and manufacturing engineers, as well as
Rrst, ratings regarding product engineers in other functions
were averaged
degree
satisfactory level of working relationship
regarding a specific component development with people
a project manager, product engineers
for the
function
0.83).
degrees
of coordination with
were averaged
In addition,
we
also
a
to obtain the
measured
total
coordination by averaging the degrees of project coordination and functional coordination (principal
component loading =
0.96).
Control Variables
Other project characteristic variables that
may
affect the relationship
between component
development performance and organizational coordination are measured as control variables.
the percentage of the
component design
that
Rrst,
was newly designed was asked regarding each new
22
component development. On average, 80%
were newly designed, and
project interactions
project interactions.
measured
for
design
87%
of
in
the
design were
33%
of
design
with inter-project interactions and
was done by
34%
a component's interdependency with other parts
in
Lastly,
respondents were asked to rate the extent
project
manager has as opposed
Rgure
6.
that
to the functional
has been explained
Analytical
new
in
projects without inter-
was
also
in
those without the interactions.
were measured by asking, on
affects the other part of the product.
of authority regarding
design work that the
manager. Figure 6 summarizes the analytical
for this research.
Framework
With or Without
Inter-project Interactions
Control Variables
with inter-
suppliers on average
of products
a 7-point Likert-type scale, the extent the component design
framework
component developments
Secondly, the percentage of design that suppliers engineered
each component design;
component designs
Thirdly,
of
interactions, schedule/cost
while design quality
coordination
is
is
significantly correlated witfi only project coordination,
significantly correlated witfi only functional coordination.
rated fiigher
is
performance
in
component design without
that achieving a strong coordination
is
generally
more
Organizational
inter-project interactions,
difficult in
component design
which indicates
with inter-
project interactions.
Table
2.
Descriptive Data and Correlation Matrix
With Inter-project Interactions (N=106)
Mean§
Performance (Sched./cost)
2 Performance (Design Quality)
1
S.D.
12
3
4
5
6
7
.06
.01
.10
.14
-.06
3.42
3 Total Coordination
4
Project Coordination
5
Functional Coordination
6 Project
N/lgr
Authority
7 Component's Interdependency
8
New
4.79
1.72
.34
.24
Design Ratio
9 Supplier's Design Contribution
-.16
-10
-.02
perform well
in
schedule and cost. Secondly,
functional coordination
is
important to
in
manage
component design
with inter-project interactions,
inter-project coordination
performance. Thirdly, the influence of project coordination as well as
performance
is
interactions.
In addition,
stronger
in
component design
with
even
total
for
schedule/cost
coordination on
inter-project interactions than in
project manager's authority contribute to
performance
those without
significantly in only
those projects with inter-project interactions.
In
addition to the differences
in
the influence of organizational coordination variables, other
design characteristic variables also affect performance differently between these two types of
component design. A component's interdependency
with other parts of the product
and the
extent of supplier's contribution have a significant negative effect on performance only
component design
with inter-project interactions. Accordingly, respondents at the U.S. firms
tended to rate the performance higher than the Japanese respondents. This
caused by the low
return rate from the U.S. firms,
component design
projects as pointed out earlier.
who may have chosen
In
regarding the general theoretical propositions posed
Table
Regression Analysis
3.
in
for Project
Independent variables
any case,
in this
Performance
in
this bias
may have been
only high-performing
does not
affect the results
paper.
Schedule and Cost
With Inter-project
Without Inter-project
Interactions
Interactions
(N=106)
(N=119)
Design Quality Performance
There are smaller differences between the two kinds
of
component design regarding the
influence of organizational coordination on design performance than on schedule/cost
performance. Design quality performance
is
both types of component design. However,
significantly affected only
total
by functional coordination
coordination has a significant influence on design
quality only in
component developments
contribution
design also has a stronger negative influence on design quality performance
in
component design
Table
4.
Total Coordination
Project Coordination
Functional Coordination
New
Mgr
Performance
for Project
Independent variables
Constant
Project
with inter-project interactions.
In addition,
supplier's
with inter-project interactions.
Regression Analysis
Authority
Design Ratio
Component's Interdependency
Supplier's Design Contribution
Nation (US;0, Japan; 1)
in
Design Quality
With Inter-project
Without Inter-project
Interactions
interactions
(N=106)
(N=119)
3.67
***
in
in
in addition, project
with inter-project interactions,
performance
in
design with multi-project interactions.
Complexity caused by other
interdependency with other parts
design,
seem
to
and
for
and thus
it
of the
elements, such as component
product and the degree of supplier's involvement
inter-project interactions.
design without inter-project interactions
may be
supplier's involvement
The
tasl< characteristic
impose more penalty on component design with
may be because component
interactions,
coordination has a stronger influence on
easier to
more
manage
the complexity of
stronger project coordination
This
simpler than design with
component interdependency
effectively.
results of this survey indicate that, in order to effectively
component design across
is
in
manage schedules and
costs
multiple projects, not only stronger functional coordination but also
is
needed.
In addition,
other factors that impose further complexity on
the organization, such as component interdependency and supplier's involvement, tend to cause
difficulties to
Table
5.
the organization
Summary
of
in
component design
the Regression Analyses
with inter-project interactions.
proposition,
by managing
theoretically the
inter-project interactions effectively, concurrent design transfer is
most appropriate way both
design of these products that are relatively
The second
component designs
part of this
with
develop multiple products quickly and to maintain
new under
limited financial
and organizational resources.
paper examined how organizational requirements
and without
these two types of component design
-
differ for
The questionnaire survey
inter-project interactions.
component engineers provided evidence
differ, particularly
to
that organizational coordination required to
with
and without
inter-project interactions
-
of
215
manage
significantly
with respect to schedule/cost performance. While only project coordination
significant influence
on schedule/cost performance
in
has a
design without inter-project interactions,
both functional coordination and project coordination have a strong impact on performance of
design with inter-project interactions. Moreover, the magnitude
of the
impact of project
coordination on the performance of design with inter-project interactions
without interactions.
factors that
We also found that inter-project interactions make
impose complexity on the organization such as
intra-project
is
it
bigger than
difficult to
in
those
deal with other
component
interdependency and supplier's involvement. This result theoretically implies that a different model
is
required to predict the relationship between project strategy, organizational coordination, and
performance
for projects with inter-project interactions.
These
findings imply that effective
projects, rather than focusing
on
management
of multiple
individual projects separately,
new product development
can be a competitive advantage
the market on the presumption that financial eind engineehng resources are limited.
also found evidence that
managing
for
component design
in
we
inter-project interactions effectively is organizationally difficult
and requires a substantially higher degree
across functions than
In addition,
in
a project that
is
of organizational coordination both within
a function and
independently managed. The coordination requirements
with inter-project interactions are so different from those without inter-project
interactions that different organizational structures
and processes are
likely to
be needed. This
paper has not discussed specific processes or mechanisms with which project organizations actually
manage
inter-project coordination.
In this
area, there are
28
many
questions to be explored regarding
appropriate
means to
coordinate across multiple projects.
Rrst, there is
an issue
of task partitioning
regarding specific engineering tasks that are related to multiple projects (von Hippel, 1990). For
example, components
like air
conditioners that are relatively easy to share
projects without substantial modifications
projects.
Second,
different
groups
may be designed by
people
of
may have
to
the
same engineers across
be responsible
coordination effectively, depending on the nature of coordination.
coordination
functional
may be well-managed by
managers
in
direct coordination
each engineering function or
aaoss a number
for
multiple
inter-project
For example, inter-project
between engineers
project
managing
of
managers
in
in
multiple projects,
multiple projects, or by an
independent coordinating group. Third, selecting appropriate coordination means and effectively
implementing them are also important issues, which include formal or informal meetings, long term
planning for sharing components across multiple projects, and computer systems such as
may
facilitate
design transfer between projects.
order to explore such coordination processes,
In
of different
component design tasks
coordination.
Project
it
is
essential to analyze in detail the nature
that affect requirements of different types of organizational
and functional coordination requirements depend on a component's cross-
functional interdependency and inter-project interdependency.
Rgure 7 categorizes
component belongs
different types of
is
component. However,
to design
CAD that
components
Using these two dimensions,
into four groups.
A group
to
which a specific
conceptually determined by a firm's inter-project strategy for a specific
it
also at least partially
depends on the nature
of the
component
with respect
interdependency with other components, and on the benefits of perceived differentiation
from other products
in
the market.
determined here by the degree
The degree
of differentiation benefits for
of contribution the
component has
in
perceive one product as different from other products the firm offers.
design directly visible to the customer
is
a specific component
persuading customers to
For example, the upper-body
usually distinctive to each product rather than shared
across multiple projects. Therefore, upper-body design need not to be coordinated with other
projects.
of tfie
However, the upper body design should be extensively interdependent with other parts
automobile design, such as the suspension system and
29
is
interior that also
need
to vary with
each product
to
make
components, need
Rgure
7.
to
it
distinctive.
be
well
These type
components, which
of
managed through a
we
call differentiated
system
project-oriented organization.
Design Interdependency and Organizational Coordination
Shared system
components
system
components
Differentiated
(e.g.,
High
upper body, seats)
(e.g.,
platform,engine)
System-level
Project -oriented
^
^
Organization
Inter-project
coordination
Within project
Cross-functional
Interdependency
Function-oriented
Low
Organization
Differentiated functional
Standardized functional
components
components
gamisii)
(e.g., int./ext.
(e.g.,
audio sets,
Low
Ixtttery)
High
Inter-project
Interdependency
On
the other hand, there are
market are
relatively small.
batteries for
projects.
design.
each
some components
for
which the benefits of differentiation
in
the
For example, firms do not extensively differentiate audio systems or
different product
and may want
This type of component design
to standardize
is relatively
among
different
independent of other parts of the automobile
We call this type standardized functional components,
through a function-oriented organization.
these designs
which
may be managed
effectively
We also label the type of components that have a strong
interdependency along both dimensions, intra-project and
components. There are many components
of this type,
inter-project,
which range from major components such
as platforms (underbodies and suspension systems) and engines,
brakes and door-lock systems.
30
as shared system
to small
components such as
This framework for different types of components raises two related questions. The
how
firms
different
can structure an organization
to
manage
effectively the
development
first is
of significantly
design tasks, which also have to support the project strategy. Since simple matrix
organization does not
mechanisms
seem
to
be adequate, there may have
within a matrix organization.
Secondly,
how
function-oriented matrix organizations, because this type of
addressed
this
problem
mechanisms above the
of coordination
Few
if
be extensively differentiated
firms can
shared system components, which cannot be coordinated by
the context of a specific project as a system.
to
manage
the development of
either traditional project-oriented or
component must be coordinated
within
any empirical or theoretical studies have
between multiple systems. Companies need either strong
matrix organization such as, executive-level long-term planning offices, or
organizational structures and processes that enable system-level coordination across multiple
projects.
In
order to analize this and related issues,
appropriate as a
project
first
step,
we
and we are thus continuing
managers and engineers
at
t>elieve that in-depth
this
research through extensive interviews of
major automobile manufacturers.
31
case studies are
References
Abegglen, James C, and George
Basic Books).
Allen,
Thomas
J.
(1977).
Stalk, Jr. (1985).
Managing the
Row
of
Kaisha:
The Japanese Corporation (New
York,
Technology, Cambridge, Mass, MIT Press.
Automobil Revue (Bern, Switzerland). Annual.
Business
Clark,
Week
Kim B
(1992). "A
New
Era for Japan."
Scope and
(1989). "Project
April 17:
50 &
Project Performance:
Supplier Involvement on Product Development,"
"Detroit's Big
The
Chance." June 29: 82.
Effect of Parts Strategy
Management Science,
and
35, 10 (1989),
1247-1263.
Clark, K. B., B.
W. Chew and
Activity, 3,
Clark,
Kim
B.,
Development in the World of Auto
and Performance," Brookings Papers on Economic
T. Fujimoto (1987). "Product
Industry: Strategy, Organization
729-771.
and Takahiro Fujimoto (1991). Product Development Performance: Strategy,
Management in the World Auto Industry (Boston, MA, Harvard Business
Organization, and
School Press.
Cusumano, Michael
A. (1991). Japan's Software Factories:
York, Oxford University Press).
A Challenge to
U.S.
Management (New
Cusumano, Michael
A., and Kentaro Nobeoka (1992). "Strategy, Structure and Performance in
Product Development: Observations from the Auto Industry," Research Policy 21, 265293.
Dertouzos, Michael L., Richard K. Lester, and Robert M. Solow (1989).
(Cambridge, MA, MIT Press).
Made
in
America
Fujimoto, Takahiro, and Antony Sheriff (1989). "Consistent Patterns in Automotive Product
Strategy, Product Development, and Manufacturing Perfonnance - Road Map for the
1990's," Cambridge,
MA, MIT
International
Motor Vehicle Program, International Policy
Forum.
Galbraith, J. R. (1974). "Organization Design:
An
Information Processing View", Interfaces, 4, May,
1974. 28-36.
Galbraith,
R. (1982). "Designing the Innovating Organization." Organizational
J.
Dynamics. (Winter):
5-25.
Henderson, Rebecca M., and Kim B. Clark (1990). "Architectural Innovation: The Reconfiguration
of Existing Product Technologies and the Failure of Established Rrms," Administrative
Science Quarterly, 35, 9-30.
Imai, K.,
Nonaka and H. Takeuchi (1985). "Managing the New Product Development Process: How
Japanese Learn and Unlearn," in K. B. Clark et al., ed.. The Uneasy Alliance: Managing the
Productivity-Technology Dilemma (Boston, MA, Harvard Business School Press).
I.
Katz, Ralph
and Thomas
R&D
Matrix",
J.
Allen (1985). "Project Perfonnance and the
28, 1,67-87.
Locus
of Influence in the
Academy of Management Journal,
Kekre, Sunder, and Kannan Srinivasan (1990). "Broader Product Line:
Success," Management Science, 36, 10, 1216-1231.
32
A
Necessity to Achieve
W. and David H. Gobeli (1988). "Organizing for Product Development Projects",
Journal of Product Innovation Management, 5,180-190.
Larson, Erik
Lawrence, Paul R. and Jay W. Lorsch (1967). Organization and Environments: Managing
Differentiation and Integration. Homewood, IL, Irwin.
Marquis, Donald G. and D.
L.
Straight (1965). "Organizational Factors in Project Performance",
Unpublished Working Paper, MIT Sloan School
Miller,
Alex (1988). "A
Taxonomy
of
of
Management.
Technological Settings, with Related Strategies and
9, 239-254.
Performance Levels," Strategic Management Journal,
Sheriff,
Antony (1988). "Product Development in the Automobile Industry: Corporate Strategies
and Project Performance," Cambridge, MA, Unpublished Master's Thesis, MIT Sloan
School of Management.
Thompson, James
D. (1967). Organizations in Action
(New
York, McGraw-Hill).
L. (1978). 'Technical Communication in R&D Laboratories: The Impact of Project
Charactehstics", Academy of Management Journal, Vol. 21, No. 4, 624-645.
Tushman, Michael
Work
Rosenberg, Nathan (1982). Inside the Black Box: Technology and Economics. Cambridge. MA,
Cambridge
Tushman,
L.
University Press.
Mk;hael (1979). "Managing Communication Network
37-49.
in
R&D
Laboratories", Sloan
Management Review,
von Hippel,
Eric (1979). 'Task partitioning:
An
innovation process variable", Researcfi Policy, 19,
1990. 407-418.
Tushman,
L.
Michael. (1979). "Managing Communication Network
37-49.
in
R&D
Laboratories", Sloan
Management Review,
Utterback, J. M., and W. J. Abernathy (1975). "A
Innovation," Omega, 3, 6.
Dynamic Model
von Braun, Christoph-Friedrich (1991). 'The Acceleration Trap
Management Review (Summer), 43-52.
Womack, James, Daniel Jones, and Daniel Roos
(New York, Rawson Associates).
(1990).
33
in
of Product
and Process
the Real World," Sloan
The Machine
that
Changed the World
:
Appendix
Change
1
in
.
Change Index
of Platform
Design
Wheelbase and Treads
Points
Both wheelbase and tread are the same
Only either wheelbase or tread are new
Both wheelbase and tread are new
Change
in
Suspension Design
Points
Suspension system and design are the same; modification
Suspension system is the same, but design is new
Suspension system is new
0:
1
2:
If
a
sum
of the points in
Appendix
both areas
2. Distribution of
is
three or more, platform design
is
in
geometry
defined as new.
Transfer Lag
"Concurrent" "Sequential"
10-
Count
5
-
4
3
5
6
Transfer Lag (Years)
34
7
8
L
92b
2U
Date Due
i^^ ft^
S^lfe. «i
Lib-26-67
3
9080 007 212 076
Download