A-Z of (Better)
Brand Health
Tracking
Jenni Romaniuk
About the A-Z of (Better)
Brand Health Tracking
Following the launch of my third book,
Better Brand Health: Measures and Metrics
for a How Brands Grow World, in 2023 for
the month of February I posted the A-Z
of Brand Health Tracking on LinkedIn each
day, covering the little facts and useful
tips around each letter.
Here is a compilation of those posts.
Professor Jenni Romaniuk is a Research Professor of
Marketing and Associate Director (International) at the
Ehrenberg-Bass Institute – the world’s largest centre for
research into marketing.
Thank you to everyone who has been
on this fantastic ride, and particularly to
my Ehrenberg-Bass Institute workmates
who have provided such great support.
Creating these posts has been much
more work, and much more fun,
than I ever anticipated!
– Professor Jenni Romaniuk
As the key architect behind the Ehrenberg-Bass approach
to Distinctive Asset, Category Entry Point and Mental
Availability measurement, Jenni has worked with
companies all over the world to help them build
stronger brands.
Jenni has written three books:
Professor Jenni Romaniuk
Associate Director (International)
Ehrenberg-Bass Institute,
University of South Australia
LinkedIn
Website
Building Distinctive Brand Assets, which helps marketers
to future-proof their brand’s identity, How Brands Grow
Part 2 which builds on the knowledge revolution started in
How Brands Grow and her new book, Better Brand Health
provides a valuable resource for those looking to get the
most out of their brand health tracking.
Ehrenberg-Bass Institute
The Ehrenberg-Bass Institute is the world’s largest
centre for research into marketing, based at the
University of South Australia in Adelaide. The team
of 60+ marketing scientists are advancing marketing
knowledge, busting pseudo-science and marketing
myths, and teaching marketers how marketing really
works and how brands grow. We help Ehrenberg-Bass
Sponsors all over the world to develop and benefit
from a culture of evidence-based marketing.
Jenni’s expertise spans mental and physical availability,
brand equity, brand health tracking, word-of-mouth and
advertising effectiveness. She was editor of the Journal of
Advertising Research from 2014-2016, and now sits on the
Journal’s Senior Advisory Board.
2
A-Z for (Better) Brand Health Tracking
Ehrenberg-Bass Institute for Marketing Science
3
Specialist Research Services
Distinctive Asset
Measurement
Category Entry Point
Identification and Prioritisation
Laws of Growth
Analysis
Media Planning
Review
Better Brand Health:
The Workshop
Distinctive Assets are the non-brand name triggers
that remind category buyers of your brand. They play
an important role in building Mental and Physical
Availability and need to be developed and protected
over the
long-term.
Category Entry Points (CEPs) are the building blocks of
Mental Availability — they capture the thoughts that
category buyers have as they transition into making
a category purchase.
The Ehrenberg-Bass Institute has conducted decades
of research into marketing. This large body of research
includes the discovery of a number of law-like patterns
of buyer behaviour and brand performance.
The Media Planning Review allows you to ensure your
media decisions are optimal.
The Ehrenberg-Bass Institute run a two-stage project
which identifies CEPs for your category, benchmarks
your brand’s current performance and identifies priority
CEPs to develop for the short
and long term.
To check if the Laws of Growth apply in your categories/
countries, the Ehrenberg-Bass Institute will analyse
your data (e.g. standard panel data) to document the
fundamental laws-of-growth patterns, and highlight any
meaningful deviations that may exist.
This is a full day event designed to improve your brand
health tracker. We cover data collection frequency,
sampling issues and Key Performance Indicators (KPIs).
All key areas you need to better track brand health. This
is an opportunity to improve how you measure, collect
and analyse data and align your team on this key area
of measurement.
The Ehrenberg-Bass Institute has an empirically
validated approach to assessing the strength of
potential Distinctive Assets, and will advise on the
opportunities and threats for building a strong longterm brand identity.
Learn more
Learn more
Based on this research, we will tailor recommendations
outlining the key steps you should take for profitable
brand growth.
Evidence-based media decision making ensures that
this audience has the right characteristics and is
exposed to your communication when and where
it matters most.
The Ehrenberg-Bass Institute Media Planning Review
provides the practitioner with the established evidence
of how media works and brings clarity to the often
murky waters of media decision making.
Read more
Read more
Read more
4
A-Z for (Better) Brand Health Tracking
Ehrenberg-Bass Institute for Marketing Science
5
Attitude
A
B2B
B
How do I like the, let me count the measures?
Valentine’s Day* advertisements are
here which means soon thoughts
turn to love. Questions such as How
much does he/she/they love me?
swirl in the air. However, for brand
health trackers every day, many
times a day, is Valentine’s Day.
This is great for getting gifts, but
not so great for having an efficient
and effective brand tracker.
Attitude questions, which ask category buyers
how they feel about a brand, can appear in
questionnaires as:
∙ rating scales (e.g., rating the brand on
a scale from love to hate, like to dislike,
or terrible to perfect),
∙ attributes (e.g., is a brand I feel close to,
a brand I love, a brand I care about) or;
∙ future plans/intentions (e.g., how do you
feel about buying the brand in the future?).
If you are unsure which one to keep, or are lucky
enough to only have one measure, then do a quality
check on that measure. Does it capture the full
range of attitudes? In particular does it have a home
for those with no attitude at all, which is often a
very popular response from a brand’s non-buyers
(more on this in later posts).
*This post was written in Feb 2023
Three steps to improve
your tracker
Identify likely
attitude measures.
Analyse category buyer
response patterns to test
how similar they are.
If there is duplication,
whittle down the list to
one attitude measure.
B2B buyers have the same brains as
everyone else. Analysis of the responses
from B2B respondents to brand health
questions reveals they follow the same
patterns and suffer from the same biases
as responses from B2C customers.
For example see:
Romaniuk, J., S. Bogomolova and F. Dall’Olmo Riley (2012).
“Brand image and brand usage: Is a forty-year-old empirical
generalization still useful?” Journal of Advertising Research
52(2): 243-251.
The challenge with B2B is often reaching a quality sample, so
make sure you use a panel provider with this expertise. Given
the cost and difficulty of getting a good B2B sample, it’s even
more important not to waste time on useless measures.
The CEP paper and Better Brand Health both contain alternative
approaches to identify Category Entry Points. This is helpful
to get the right inputs, if you can’t easily survey your B2B
customers. Unfortunately, I don’t have an easy non-survey
solution for actually assessing brand health. In Better Brand
Health, we have a chapter on the use of online data scrapping
for brand health assessment purposes. The challenges we
discuss there are even more likely to hold back online data
quality in B2B categories.
This means we can use the same question styles, the same
response styles and the same analyses.
However the inputs are likely to vary. For example, in Mental
Availability measurement you will have different Category Entry
Points (attributes) and different brand lists, but you can still use
the same ‘free choice, pick any’ measurement approach.
See my paper on Cateogry Entry Points in a B2B world for more
on this:
Read more
6
A-Z for (Better) Brand Health Tracking
Ehrenberg-Bass Institute for Marketing Science
7
Category
C
No this is not about Category Entry
Points, but rather about one of the
key underlying principles for better
brand health tracking, which is to
Design for the Category.
Regardless of your brand’s current size, you need to
understand the whole category.
If your brand is bigger, you not only need to
understand your brand’s performance but also its
threats, which are likely to be new/small brands
nibbling at your market share, but also other big
brands and medium ones too.
Also, in the future you might launch a new brand in
the category and then suddenly be looking through
the smaller brand lens again.
If you are a small or new brand, you need to
understand your brand’s performance but you also
need to understand where your future sales will
come from, which the Duplication of Purchase Law
tells us will most likely be the larger brands.
8
A-Z for (Better) Brand Health Tracking
Designing for the category includes:
∙ When Recruiting all category buyers to be part of
the sample, in line with the typical buying weight
distribution.
∙ Including all the major competitors and smaller
competitors (or a representative sample of smaller
brands if necessary).
A tip to check your brand health tracker is to look
at your current measures though the lens of very
different brand in the category (e.g., if you are a
big brand, imagine a small brand using the same
questionnaire) - what questions would you need/
want to change?
These are the questions that have a bias in their
design and that you should rethink now.
Distinctive Assets
Yes this topic is a bit predictable, but
perhaps the point I want to make
is less so. While Distinctive Assets
are really valuable brand memories,
they don’t need to be part of a brand
health tracker.
First, the tracker is a poor place to do the strategic
research that you need to identify the brands’
longer term Distinctive Asset Palette. In strategy/
benchmarking research you should test many more
assets than you will end up building for the brand.
Therefore this is better tackled as a stand alone
piece of research.
A key outcome of strategy/benchmarking research
is a subset of assets for the brand’s Distinctive Asset
Palette. These assets are either currently close to
100% Fame and 100% Uniqueness, or the assets to
build to get to 100% Fame and 100% Uniqueness.
Therefore, you only need to monitor this subset of
assets, and the (rare) instance where you might
want to introduce a new asset.
D
Ongoing Distinctive Asset tracking should check you
are protecting strong assets (staving off memory
decay) and making progress on building the next
wave of assets. You will need to also throw in
some key competitor assets to mask the brand
of interest, but this helps you keep track of key
competitor activities. Your first follow up after the
benchmarking/strategy piece is usually better after
a year (at minimum) so you have time to remove
inconsistencies and execute the asset building
tactics.
If you include Distinctive Assets in your tracker,
this section needs to be placed at the start of
the questionnaire, before any brand names are
revealed. However, it may just be easier to cut down
the strategy/benchmarking questionnaire and track
this as a stand alone study. That gives you more
flexibility in frequency and timing.
Regardless of your tracking approach, do invest
in ongoing systems to ensure that you create
opportunities to use and build assets. This is better
done before marketing activities go into the field.
Ehrenberg-Bass Institute for Marketing Science
9
Exposure to Executions
Brand health trackers often help diagnose
the longer-term impact of recent past
marketing activities on category buyer
brains. This makes it useful to assess the
exposure of category buyers to the
brand’s marketing activities.
To measure exposure, there are a range of advertising
memorability measures, a common example being, Which
brands in <insert category> have you seen advertising for
recently? The strong presence of the brand in this advertising
retrieval cue introduces such a bias that the results from this
measure usually devolve to simply a test of brand awareness,
whereby big brands score more while small brands score less.
Indeed some researchers have noticed the correlation to be so
strong, they use unprompted advertising awareness as a proxy
for unprompted brand awareness, which is a detrimental to
understanding either concept.
An alternative approach is to use an execution-cued exposure
test, whereby you test memory for exposure to the execution
stripped of branding first, and then memory for the brand
as a second step.
10
A-Z for (Better) Brand Health Tracking
While this is more fiddly to do, it has the advantages of:
1.
Providing a more specific, and richer execution-based cue
to for buyers to access their memory.
2. Allowing you to capture brand memory separate from ad
memory.
This is useful because marketing can be ineffectual due to
lack of reach or lack of branding, but the remedy for either
differs.
An execution-based exposure still has a slight brand buyer bias,
as brand buyers notice advertising for their brand more than
non-buyers, but has more variability across executions, and so
can have greater diagnostic value than brand-cued approaches.
E
Let’s Forget Funnels
I don’t understand the appeal of
purchase funnels because they are:
False – Funnels give a misleading view of the
buying process, as they make it seem as if buyers
should naturally follow a (similar) path from say,
Awareness to Recommendation/Advocacy. Any
observation of cohort buying data over time or
word-of-mouth data quashes this assertion.
Note: the argument that ‘we don’t really mean
it that way’ doesn’t hold water with me as if you
don’t mean it that way, why have a visual image
that shows it that way - picture, 1000 words,
and all that…
Fraught with irrelevance – Funnels make simple
data unnecessarily complex by turning numbers
into ratios. Ratio metrics are a blessing for the
insignificant as you can only get high ratios with
low incidences.
Futile – The vast majority of measures included in
funnels are correlated with brand share/penetration,
and so there are easier ways to see if you hare
higher or lower than expected for your size for
any metric (a scatterplot against brand size will
often do).
Frittering away your time – Ratios put a barrier
between you and insight. The scores can lift or
decline due to changes in the numerator or the
denominator or both, so if your funnel ratio
changes you then have to reverse engineer
the calculation to work out what happened.
This just wastes your time.
F
Awareness
Interest
Desire
Action
Loyalty
Advocacy
Ehrenberg-Bass Institute for Marketing Science
11
Good Measures
If you want to bake a cake you need
the ingredients to make a cake. If
your ingredients are for a stir fry you
are unlikely to end up with a cake,
no matter how good a chef you are*.
Sometimes we are presented with measures that
have not been fully tested. How do we know if
something new is a good measure to include in a
brand health tracker?
We can improve our odds of good measurement
focusing on those that have the raw ingredients to
be a brand growth indicator, and discarding on the
ones that don’t.
Here are some ingredients to look for:
∙ Measures that cover all category buyers, but
particularly a brand’s non-buyers, and not just
focused on a brand’s current buyers.
∙ Measures that draw on latent and/or nascent
brand memories, and not just on strong
emotions.
∙ Measures that can change without a great deal
of buyer thought, and not just when a buyer has
a ‘road to Damascus’ conversion.
∙ Measures that can capture a small change even
when distributed across a wide group of people,
and are not only relevant to a small (usually
weird) group of buyers.
There might be other filters we can use, but these
can get you started. Any measure you see as a
leading indicator for growth - how does it stack up
on these criteria?
G
Handling the Haters
Legend has it that if a customer is
happy with a brand they tell three
or four other people, but if they are
unhappy with a brand they tell a
whopping 10 to 20 other people.
By implication, the world is therefore
swamped with negative brand
sentiment that marketers need
to continually find and quash.
Online reviews can easily give a similarly distorted
view, because people typically only comment
when they experience something very good or very
bad. This means a large amount of rating data are
missing - that of the pretty good, OK, did the job.
This is illustrated in Better Brand Health, where we
compare the sentiment distribution from an online
rating scale with that generated from a survey of
category buyers.
This heightened attention on the haters can waste a
lot of time and effort, and can (even unconsciously)
have a detrimental affect on strategy if marketing
activities are always viewed through the lens of
avoiding or removing negative WOM.
12
A-Z for (Better) Brand Health Tracking
The good news is that empirical evidence from
Professor Robert East shows negative WOM for brands
is given by fewer people than positive WOM, but at
much the same rate*. The world (and pretty much
every brand) has much more positive WOM than
negative WOM. The legend is indeed a false tale.
In brand health tracking, it’s worthwhile to
benchmark WOM to get the full, unbiased, picture.
You can check if the brand/company’s negative WOM
(and positive WOM) is normal. This usually has the
side benefit of showing you that negative WOM is
insignificant and your should direct your efforts to
more useful areas of brand management. Remember,
in a world awash with metrics, knowing what metrics
are not important is very helpful!
This benchmark data can also help determine
whether WOM requires close monitoring and so
need ongoing tracking or is something to only
monitor by exception, and only triggered by an
event likely to stimulate WOM.
H
Heavy Buyers
No one disputes heavy buyers are important
for your brand’s current sales. Indeed its
something of a circular argument because
it is via their past sales that they usually
get classified as heavy buyers (despite that
in many CPG categories, only around 50%
continue to be heavy buyers in a subsequent
time period). So heavy buyer responses are
important to capture in a Category Buyer
Memory tracker.
The issue arises when we mix all buyer groups together,
particularly heavy buyers with very light/non-buyers.
It becomes like trying to hear someone whispering in one
ear when someone else is shouting in the other. Splitting
out buyers from non-buyers is useful for most metrics, but
splitting out heavy buyers from light buyers is useful for
more difficult/more extreme measures. For example,
we find heavy buyers have significantly higher Top-of-mind
awareness scores than lighter category buyers,
but not significantly higher spontaneous or prompted
awareness (Hogan, 2015).
H
Something I wrote
but did not post.
Intentions-to-Buy
Manna from heaven or the road to hell?
Intentions-to-buy measures crop up in all
sorts of research from advertising ‘brand
lift’ studies to pack testing to, of course,
brand health tracking.
This all assumes, of course, a strong relationship
between buyers’ stated intentions and their future
buying behaviour.
A typical intention-to-buy question looks something
like this:
Q: How likely are you to choose each of these brands
next time you buy <insert category>?
With five response categories:
1.
Definitely will buy
2. Probably will buy
3. Might or might not buy
4. Probably will not buy
5. Definitely will not buy
14
A-Z for (Better) Brand Health Tracking
I
In the 1960’s consumer follow-up studies found that
more consumers with no previously stated intention
purchased a category than consumers with a previously
stated intention. In reaction to this Thomas Juster
created the Juster scale, which captures purchase
probabilities rather than intentions, to improve the
accuracy of future buying forecasts ( Juster 1966),
something confirmed by a more recent meta-analysis
(Wright & McCrae 2007).
What is the difference? Intentions speak to the
consumer having latent future plans, while probabilities
are calculated on the spot, by the consumer, based on
their current available knowledge. For example:
Do you intend to get sick next winter? You would
probably answer “Definitely not”. But if I asked you
to assign a number to the probability you will get sick
next winter, this is probably not zero, but would vary
depending on your own immune system, whether you
have kids, where you work etc. As Thomas Juster found,
many people with zero plans to act can have a non-zero
probability of acting.
If you do want to continue using intentions-to-buy,
Better Brand Health outlines key conditions where
intentions-to-buy are thought to be more accurate.
This could help you improve your current measure.
For those looking to upgrade, the chapter also contains
more detail on the wording of the Juster Scale.
Now you might say, I know things can interfere
between expressing and intention-to-buy and acting on
it. I am just measuring the current consumer mindset.
That is fine, just then appreciate this is attitude-tobuying measure, with more emphasis on the attitude
than the buying, and don’t kid yourself it is an accurate
measure of future behaviour.
This measure is an easy way to ‘test’ if
something could be linked to higher sales,
for example:
Did those exposed to this ad
have higher purchase intent?
Will this pack change lower
purchase intent?
Is perceiving the brand as innovative
linked to higher purchase intent?
Ehrenberg-Bass Institute for Marketing Science
15
Jobs To Be Done (JTBD)
A question I am often asked is what is the
difference between JTBD and Category Entry
Points (CEPs)? Let me first start with what
they have in common - they are both about
understanding the category from a categorybuyer point of view, rather than the brand.
Therefore, applying either approach should
improve your attribute list and avoid the
brand-based myopia that often dominates.
A CEP approach draws from how our memory works, under
the Associative Network Theories of Memory. It is based upon
empirically observed facets of memory that affect the brand’s
chance of being accessible in category buyer memory. This
includes that even existing memories are not fixed, but subject
to natural decay unless refreshed. Sometimes all we need to do
to grow a brand is to turning a decaying memory into an easily
retrieved one. We might not need to change the brand, but
rather change category buyers memories for the brand.
When combined with the framework of the W’s to give us a
multifaceted perspective, a CEP approach can give us a pretty
comprehensive view of how category buyers interact with
the category.
16
A-Z for (Better) Brand Health Tracking
JTBD and the idea of ‘hiring’ a brand to do a job can be a
useful metaphor to help marketers think creatively about their
category, and identify opportunities to innovate. But it is just a
metaphor, and if we take it too seriously we can be fooled into
thinking of the category buyer as a more logical, thoughtful
actor than the the real world would show. I did not ‘hire’ Sushi
Train for lunch today, I thought of options my parents would
like, on a warm day, that was on the way home from the
vet. Visiting Sushi Train was one option, ordering Uber Eats
from Sushi Planet was another, but Sushi Train was going to
be quicker (better physical availability) and so got selected
at this time.
At their best, CEPs provide insight to improve how you
manage a brand now, while JTBD provide ideas for innovating
the brand to change its future trajectory. But CEPs can stimulate
innovation and JTBD can provide insight to current category
interactions.
I prefer CEPs because it is based in a real memory process
that occurs, largely without us knowing. Therefore, in Better
Brand Health, there is a chapter on methods to get CEPs.
However, I think taking a JTBD approach is going to be
better than doing neither.
J
JTBD and the idea of ‘hiring’
a brand to do a job can be
a useful metaphor to help
marketers think creatively
about their category, and
identify opportunities
to innovate.
Key Performance
Indictators (KPIs)
Now I love a KPI as much as the next
marketer, but I think the desire for
one magic number is holding back
marketers and marketing thinking.
In the red corner we have the ‘Emperor’ metrics,
whereby apparently one metric reigns supreme and
will tell you all you need to know about your brands
health. The NPS is the most recent in a long line of
such KPIs. I remember when the Conversion model was
all the rage, turning how people convert to religions
into a marketing KPI - OK might have been the age of
‘cult brands’ but talk about stretching the metaphor.
Often marketers subscribe to this approach by default,
and lean on a favourite metric as an oracle, such as
Spontaneous awareness, without really knowing why.
It’s risky to rely on one KPI out of one measure, as this
one measure needs to capture everything important.
Given the variety of possible memory changes, this
seems improbable and indeed every sliver bullet metric
proposed so far has been quickly unmasked.
In the blue corner we have ‘Of the masses’ metrics that
take say a bit of awareness and a bit of image, and
a touch of attitude, combine them to create a single
magic number. However, if the components all move in
the same direction, why do you need them all? In this
scenario, most measures are superfluous.
However if the components move independently, how
do you benefit from combining them? The number is of
little value over time if a decrease in one component
could counteract for an increase in another component.
That destroys the simplicity the one number was
supposed to provide.
Instead, how about a middle ground.
Let’s aim for more than one component part, but not all
the measures all mixed up: A dashboard where every
measure has earned its place, a wise council, where
each member has its own area of expertise, and we
have the knowledge to know when to ask.
K
That suggests considerable value lay KPIs that document
changes in the memories of these future potential
buyers. For most memory metrics we can only observe
non-buyer memories if we analyse the brand’s very
light/non-buyers separately. This is why in Better Brand
Health’ mantra for good brand health measurement, the
middle is ‘analyse for the buyer’. Buyers’ memories are
also important, but perhaps we need different metrics
to fully understand them.
To create useful KPI’s we need a greater understanding
of how marketing activities change brand memories
and how brand memories buttress, or change, brand
buying. So I encourage you to review the KPIs you
prefer and ask yourself about the evidence as to why
and when this metric matters? If you don’t really know
then perhaps it’s time to learn more.
As a start, if brand growth comes predominantly from
expanding the customer base, this means a major shift
in buyer behaviour will be the cohort of category buyers
who did not buy the brand this time period, but who
will end up buy it next time period.
Ehrenberg-Bass Institute for Marketing Science
17
Love (Brand Love)
L
Memory Building
M
I just could not resist this one, given we are so close to Valentine’s day and in homage to the great
songwriter Burt Bacharach who passed recently, here is a song for all marketers to sing.
(to be song to the tune of ‘What do you get when you fall in love?’)
What do you get when you build brand love?
Your buyers will think, you live in a bubble
And won’t buy you once, let alone double
Marketers never build brand love again
I’ll never track brand love again
What do you get when you build brand love?
You get a media plan that’s way to narrow
You won’t reach non-buyers today or tomorrow
I’ll never build brand love again
No, no, I’ll never track brand love again
What do you get when you build brand love?
You waste enough money to annoy the CFO
And of course, your brand won’t grow
Marketers never build brand love again
Don’t you know, I’ll never track brand love again
I’m out of that dogma, that love is what binds sales
I need to remember, that brand love doesn’t scale
Don’t tell me I need buyers devout
‘Cause I’ve seen the evidence and I’m glad I’m out
Out of that dogma, that love is what binds sales
I need to remember, brand love doesn’t scale
What do you get when you build brand love?
You get advertising that doesn’t appeal
To the normal buyers, you need to steal
I’ll never build brand love again
Don’t you know that I’ll never build brand love again
I’ll never track brand love again
In brand health tracking we often test
how brand memories have changed
over time.
For example, has the link between the brand and a
key attribute* improved? But it often feels like there
is a disconnect between the marketing activity design
process and the measurement of its effect on memory.
The more we can close this gap, the more useful our
brand health tracker will be.
In Better Brand Health, I talk about two different types
of memory effects we can see in brand attribute data:
(a) messaging effects, which is a change in a specific
brand-attribute link and
(b) mental availability effects, which is a change in the
freshness of the total network.
Here I am going to focus on the ‘messaging effect’.
To build and/or refresh specific brand-attribute links
means paying attention to what we say and how we
say it.
18
A-Z for (Better) Brand Health Tracking
What we say
So much effort goes into deciding on the advertised
message. We isolate important messages we think will
help the brand get bought. These messages become
the specific memories you are building/refreshing, such
as ’value for money’, ‘a special treat when out with your
partner’ or ‘will make it easier to pay the bills’?
To see a messaging effect, the crux of the message
needs to be reflected in the attributes in your tracking.
If your attribute list is really full of useful, relevant
memories and your communications is trying to build
useful, relevant memories then there should be a
match. If there is not a match, then at least one side is
not working for you.
How we say it
We want to build brand memories, not just advertising
memories. This means we need both the brand and the
message. First, we need to assess how well does the
message translate into memories. The more easily and
universally processed the message, the more likely it
will result in a change observable in brand tracking.
Second, we need to assess if we have excellent quality
branding alongside the clear message, so the two are
co-presented. The brand anchors the message in the
right part of memory.
BTW I am writing this while watching Super Bowl ads,
and thinking about message usefulness and clarity,
as well as branding quality. Here are some interesting
contrasts from my viewing so far (you should be able to
guess which is good and which is a poor example):
Branding quality - T-mobile with Bradley Cooper versus
Remy Martin with Serena Williams
Message usefulness - Google’s FixedonPixel versus
Workaday’s ‘Rockstar’
Message clarity - Hellman’s with John Hamm & Brie
Larsen versus Michelob Ultra’s Serena William’s ad
Remember having memory changes that show up in
brand health tracking starts with having marketing
activities capable of building brand memories.
To build and/or refresh specific brandattribute links means paying attention
to what we say and how we say it.
Ehrenberg-Bass Institute for Marketing Science
19
Non-Buyers
This post is about questionnaire wording
(please don’t stop reading, it will be worth
it I promise!). Linked to the theme of a
brand’s Non-buyers, lets talk about how we
can easily (and unintentionally) depress the
responses from a brand’s non-buyers with
just the addition of a few words to our tracker
attributes.
Do you have any attributes that are worded as a comparison
with other brands? For example, ‘is more innovative than other
brands’ or ‘has better service than competitors’. If so, you will
get on average 30% fewer brand linkages from a brand’s nonbuyers that if you just used the general form of the attribute
such as ‘innovative’ or ‘has good service’.
The responses are lower because comparative worded
attributes encourage category buyers to undertake a 2-step
cognitive process:
First - think of brands linked to the attribute; and
Second - select a subset of options based on which one is
‘better than others’.
20
A-Z for (Better) Brand Health Tracking
While all buyers undertake the same 2-step process, its more
difficult for non-buyers to overcome these two hurdles to be
linked to any attribute, because they are buyers of other brands
and have no direct brand experience. So their memories for
these brands are harder to retrieve.
This approach gets you an evaluation of a brand (in an
attitudinal sense), not an association with the brand (in a
memory sense). Given that we want to see if a brand’s nonbuyers have been building brand memories, it seems like an
unfortunately ‘own goal’ to track non-buyers using questions
that depress their response. We are left with an even more
incomplete view of the memory networks of the brand nonbuyers that are vitally important for growth.
Note: If you really do want buyers to evaluate your brand versus
competitors on a quality, then there are better ways to do this
than a free choice, pick any, attribute measurement approach.
More on this and the effect of other attribute wording
modifications in Better Brand Health.
BTW - I think we might need to change the name of non-buyers
to not-yet buyers or potential buyers, as non-buyers seems to
downplay their importance to the future of the brand. But not
today as otherwise I would have to find another N!
N
If you really do want buyers
to evaluate your brand
versus competitors on a
quality, then there are better
ways to do this than a free
choice, pick any, attribute
measurement approach.
Ownership
Marketers can be a possessive
bunch, always wanting to ‘own’
something. Here are a few evidencebased thoughts on when Ownership
matters (and when it doesn’t).
You can’t (and don’t need to) own a customer
Trying to get buyers to buy only your brand is a
waste of time and resources. Sole brand loyalty,
where people only buy one brand for a category, is
rare and is typically linked to light category buying.
As Professor Andrew Ehrenberg said ‘Your customers
are really other brand’s customers who buy your
brand occasionally’.
You can’t (and don’t need to) own an attribute
Trying to be the one brand known for X quality (e.g.,
‘top quality service’, ‘value for money’, ‘to treat the
kids’) is that wonderful combination of both difficult
and unnecessary. Unique brand linkages are rare (<3%.
Instead we find around half of category buyers
(46%) have multiple brands linked to an attribute
and one-third have no links (33%). A brand owning
an attribute in the eyes of its category buyers is
rare. If an attribute is important to category buyers,
they typically link it to multiple brands.
This means you can’t avoid mental competition, you
just need to get better at combatting it - this is why
building Mental Availability is so important.
But it is essential to own a Distinctive Asset
There is an owl on the cover of Better Brand
Health. This is the same owl that appeared on the
cover of Building Distinctive Brand Assets. Hopefully
by now, you see the owl and you think of the
Ehrenberg-Bass Institute, and only the EhrenbergBass Institute. If you don’t know what owl I am
talking about, look at a book cover, to build this
asset in your memory for future reference!
Empirically, owning a Distinctive Asset means
you have (close to) 100% Fame and 100%
Uniqueness. That is, pretty much every category
buyer, when they experience the asset in the
absence of the brand, thinks of your brand, and
only of your brand. You need 100% Fame because
then the Distinctive Asset always does its primary
job. You need 100% Uniqueness to avoid evoking
competitor brands and working against yourself.
Direct your marketing resources to own the right
things rather than trying to own everything.
O
A little Limerick I wrote for Max Winchester,
who was disappointed that N was not Net Promoter
There once was as Score called Net Promoter
Which to lifting, marketers became devote-r
Til research came along
And showed face validity is not strong
And the link to future brand growth is even remote-r
Prominence
A Pillar of Physical Availability
In Better Brand Health, Professor Magda
Nenycz-Thiel and I wrote a chapter to assess
when Physical Availability should be part
of a Category Buyer Memory (CBM) tracker.
This question arose because I often see
facets of Physical Availability get turned
into attributes, whereby brand links to
these attributes are tracked over time.
You will have no idea if the colour, pack shape, logo
placement or the pink cap explains why the packaging
stands out. In any event, this is a very torturous way to
gain insight that is much more easily obtained via direct
measurement of Distinctive Assets.
For example, Prominence, which is about the ability to easily
find the brand in retail settings, gets converted to attributes
such as is easy to find on shelf or has packaging that stands
out. The issue with these attributes is they tell you very little
about the brand’s actual prominence or how to improve it.
Category buyers respond to these attributes in the same way
as they do other attributes - brand buyers score higher than
non-buyers, and big brands score more than small brands.
In the Better Brand Health chapter, we talk about all three
pillars: Presence, Prominence and Portfolio. This provides
some better alternatives to measure these important
elements of brand growth.
But aha, I hear you say, ‘This means we can use the approach
you show in Chapter 7 to identify brands that score higher or
lower than expected!’ Yes you can, but you then what do you
do with this information? How do you go from deviation to
explanation and then turn it into something actionable?
22
A-Z for (Better) Brand Health Tracking
First, you should do the background work to identify the
shopping assets you have or want to build. Then, once you
know the brand’s current or desired shopping assets, you can
monitor them either in brand tracking research or as a stand
alone study (see D for Distinctive Assets).
P
Quicker Always Better?
It is fashionable in some areas to
assess performance with a speed of
response measure. This can involve
directly timing how long it takes
someone to respond, or indirectly
through focusing on top-of-mind
measures. This approach assumes
that a quicker response = better
performance.
But is quicker always better?
There are some times when yes, it is obvious. Being
quicker to find on shelf, or quicker to find in an
online market place, are easy examples when being
quickest found is likely to reap rewards. When
looking at factors that might effect the ability of the
brand to be quickly found in those environments,
timing measures could be useful dependent
(outcome) variable.
There are other issues, such as the generalisability
of the testing environment(s) to the many, varied,
real world environments (particularly in-store where
even different stores in the same chain can have
very different category plan-o-grams), but that is a
method rather than measure issue.
Q
However, when it comes to brand memories,
the advantage of being retrieved quicker is not
so obvious. If I remember three restaurants for
booking a celebratory dinner out, do you want
to be the first restaurant I remember? Or perhaps
it would be better to be the last? Or maybe it
doesn’t matter because I can (like most people)
hold multiple items in working memory and so
I will contemplate all three.
And for Distinctive Assets, does it really matter if the
asset triggers the brand in 0.8 or 1.2 seconds? Or is
the most important thing that it triggers the brand
at all? And not a competitor brand either as well or
instead? (See O for Ownership!).
In the latter two examples, focusing on timing of
responses or just on first response is of no benefit,
and can be of great detriment as it can lead you to
miss brand memories. There might be more of this
in this series if I select Top-of-mind for T, but there
is definitely more on this in Better Brand Health.
changes you then have to reverse engineer
the calculation to work out what happened.
This just wastes your time.
Ehrenberg-Bass Institute for Marketing Science
23
Rating Scales
Two of the most common approaches
for assessing brand performance on an
attribute are:
∙ Free Choice, Pick Any where respondents tick
brands linked to attributes. They can tick as many
or as few as they like, and provide a binary 1 = yes,
0 =no or don’t know score for every brand on every
attribute
∙ Rating Scales, where respondents rate every brand
on every attribute. This is typically on a 5, 7 or 11
point scale, and so they provide each brand with a
number on the scale.
Which is better?
On the surface it might seem like a simple trade off
between easy to answer (Free Choice, Pick Any) versus
sensitivity (Rating Scales), but the empirical results
tell a different story. Both approaches rank brands in
similar order, but suprisingly a Free choice, Pick Any
approach has greater discrimination between brands
than Ratings.
24
A-Z for (Better) Brand Health Tracking
R
There are two key reasons for this:
1.
Few category buyers use the negative/disagree
part of a brand attribute scale, which makes around
half the scale points redundant. More scale points
only increase sensitivity if they are needed.
2. Non-brand buyers who don’t know about the brand
default to the scale midpoint (say 3 on a 5-point
scale). Small brands have many non-buyers who
don’t know them and this bumps up their score,
which reduces the range of the scores from highest
to lowest brand for rating scales. In a Free Choice,
Pick Any approach, non-buyers who don’t know
get zero, and so don’t count to a brand’s score.
A third factor to remember is that we don’t store
memories as ratings. Therefore, the ratings a brand
gets are calculated ‘on the spot’ and only exist when
the question is asked. In contrast, a Free Choice, Pick
Any approach can mimic the Associative Network
Structure of memory (provided you word the attributes
right, see for example N is for Non Buyer), and so can
draw directly from buyer memories.
So if you are measuring brand attributes on rating
scales, you can immediately improve your tracker
by converting to a Free Choice, Pick Any approach.
This collects better quality data in a way that is
easier on respondents.
BTW Academia loves scales
because they are easier for
multivariate analyses (such as
regression) and you can get statistically
significant difference between groups
with lower sample sizes. That is why in
academic studies even brand awareness
often gets turned into a three item scale!
Sample Screening
Questions
I will be the first to admit, Screening
questions are not a Sexy topic.
Today is Sunday, this is letter 19
and am not inspired to a Song,
a Sonnet or even a good Story.
But I will draw on an old Saying
- rubbish in, rubbish out.
At the start of the questionnaire, screening
questions are really important because expertly
crafted brand health questions, or the most
sophisticated analysis, won’t save your insights
if you have a biased sample.
What should you aim for? Remember the first
part of the mantra: Design for the category.
Your tracker sample buying weight characteristics
should mirror the normal category buyer weight
distribution. You risk getting an excess of heavy
category buyers, and a deficit of light category
buyers if you have:
1.
Buying timeframes that are short, relative
to typical category buying frequency
2. Adding category buying weight
requirements
Excluding light/very infrequent category buyers is
particularly damaging to your tracker insights if a
category is growing or declining via penetration.
S
It is also useful to use externally collected
buying data and run some parallel profiles on
the buying weight metrics for your category,
to check your sample is pretty close. If there
is a big gap between normal buying and your
tracker sample buying then taking steps to fix
this can improve the quality of data you collect.
Test different screening options for the impact
on your sample. This will highlight which
questions have the biggest impact on the
composition of your sample and data quality.
How well do you know the buying patterns of
your category? To identify key category buying
knowledge, Better Brand Health includes a
four question checklist that highlights key
information to help you make smart decisions
on category buying screening questions, as well
as when you want to capture category buying
behaviour within the questionnaire.
Ehrenberg-Bass Institute for Marketing Science
25
Top-of-Mind
Brand Awareness
Top-of-Mind Brand Awareness (TOMA) is the first
brand is recalled, unprompted, with the category
as the retrieval cue.
This conjecture is supported by empirical analysis that shows:
Simple to collect and easy to understand, TOMA is a popular brand
health tracking measure as even if suppliers change, TOMA typically
remains. This means it is often tracked over a long period of time,
and provide a sense of continuity when trackers change.
∙ Relying on the single category cue under-represents the brand’s
ability to be retrieved, but TOM is particularly restricts retrieval of
smaller brands (see Better Brand Health for this).
TOMA’s perceived value is linked to the idea that retrieving a brand
quicker is an indicator of better future performance (see Q for is
Quicker always better). When we started researching brand salience/
mental availability, TOMA was one of the measures we researched
as a possible measure. We rejected it as unsuitable to measure
Mental Availability because buyers use multiple cues to enter the
category, therefore it made little sense conceptually that one single
category cue can capture retrieval for all cues used in buying situations,
particularly once you factor in how human memory works (see 2004
paper below).
26
A-Z for (Better) Brand Health Tracking
∙ When retrieval cues change, so do the brands that are evoked,
including which are TOM; and
But even as its own metric, TOMA is lacks evidence as a measure
of brand growth because:
∙ TOM brand awareness disproportionately biases against retrieval of
brands from non-buyers and for small brands from all buyers.
∙ Over time, changes in TOM awareness are largely concentrated in
brand buyers, rather than the brands’ non-buyers needed for growth.
Therefore, it fails the ‘key audience for growth’ test on at least
two counts.
In Better Brand Health, I present the evidence for each of these
points, plus a few other areas of concern around TOMA and top-ofmind approaches in general. If you do rely on TOMA as a key Brand
Performance Indicator, this evidence might be useful to put these
readings at the Top of your reading list.
T
U-Shaped Distributions
One of the most useful (nerdy)
things to do with any data on a
scale is to look at the distribution
of responses. This can tell you a lot.
Most people are familiar with normal distribution,
which looks like a mountain and has the
beautiful property of the mean being usually
representative of the typical buyer. But much
of the data we deal with in brand health is not
distributed normally, so if we don’t understand
the distribution, we risk making unfounded
assumptions about the representativeness of the
mean and how much variance there is around it.
Those familiar with the Laws of Growth will know
category and brand buying frequencies follow
the Negative Binomial Distribution, which usually
looks like a reverse J (see How Brands Grow 1 & 2
for more on this). There are many buyers buying
rarely, if at all, and a long tail of a few people
buying very frequently. Because of the long tail,
the average (mean) buying rate is higher than
the typical buying rate, which can lead you to
assume normal buyers buy more often than
they really do.
Looking at the distribution can help you judge
the usefulness of data from different sources.
For example, the underlying distributions help
us understand whether these ubiquitous online
brand ratings can give a brand manager an
accurate depiction of category buyers’ attitude
to the brand.
A U shaped distribution, with lots of people
either really happy or really unhappy, and very
few people in the middle, often pops up in
ratings or review data online. This occurs largely
because we only see responses from people
motivated to leave a review, due to a very
good or very bad experience.
U
In Better Brand Health, Professor Anne Sharp and
I have a chapter that looks at whether online
data scraping can replace brand health surveys.
In it we compare the data from a survey and
from yelp reviews for the same restaurant brands
to illustrate how far online ratings data differs
from normal. This highlights that online review
data has to be interpreted with care, due to its
biased approach to sample recruitment.
So take care with the U-shaped distribution
data such as from online review data - its
mean is meaningless and its meaning is often
unrepresentative.
In contrast, brand attitude distributions collected
on a normal sample of category buyers, follow
a slightly positive skewed distribution, with
the peak usually around the mildly positive/
somewhat agree point in the scale.
Therefore online review data can give you the
impression category buyers feel more strongly
about the brand than they do.
Ehrenberg-Bass Institute for Marketing Science
27
Valence in
Word-of-Mouth Effects
When choosing what to write for each
letter, sometimes I feel like I am robbing
Peter to pay Paul.
For example today when thinking of V options, I thought
Valence in word-of-mouth. Yes that is a V, but it could also
be a W, which is tomorrows’ letter. Then, with a heavy
sigh, I decided that tomorrow’s letter would be future
Jenni’s problem and that V for Valence in word-of-mouth
effects it is….
Back to Valence in word-of-mouth effects, I am sure we
all logically intuit that positively valence WOM (PWOM)
is good, and negatively valanced WOM (NWOM) is
bad. And for the most part we would be right. We have
benchmarked that about 3% of WOM has counterintuitive
effects, where PWOM decreases your chance of buying
and NWOM increases your chance of buying. This happens
when there is a mismatch in preferences with the giver of
the WOM - if they like it, I probably won’t, kind of thing.
What we often fail to grasp is that valence also has an
impact on WOM effects as it interacts with the probability
of buying the brand being talked about. This means
who’s responses we prioritise for WOM metrics, differs
by Valence.
28
A-Z for (Better) Brand Health Tracking
Most people have close to zero probability of buying
most brands. Mostly because they are not in the market,
and (close to) zero probability of buying the category
means (close to) zero probability of buying brands within
the category.
For example, NWOM is most influential when it reaches
current buyers, with sufficient headway in their future
buying probability to decline in the face of bad news
about a brand. However, someone with car insurance
due for renewal in November, right now (in February)
has a zero probability of buying most car insurance
brands. Therefore, if that person receives NWOM about
one of the zero (or very close) probability brands, nothing
can happen as their buying probability can’t be lowered.
In contrast, PWOM is most influential on buying when
it reaches category buyers with low probability of buying,
as these people have the greatest room to improve their
chance of buying. PWOM received just before renewal for
a brand, has only a small as the buyer was already going
to buy.
V
Professor Robert East’s modelling (showcased in How
Brands Grow Part 2) shows the influence of PWOM
declining and NWOM increasing as a category buyers’
initial probability of buying a brand gets higher. This
means if you are tracking WOM, you want metrics that
measure:
1.
the reach of NWOM amongst the brand’s buyers; and
2. the reach of PWOM amongst the brands non-buyers.
Word-of-Mouthish Attributes W
Are a Waste of Time
Brand attributes list can get very
long. Here is one tip to shorten them.
Check if your list includes WOMish attributes,
such as:
∙ A brand lots of people are talking about
∙ A brand you would recommend
∙ Is recommended by family and friends
∙ Has a buzz about it
See the Word-of-mouth chapter in
Better Brand Health for more on this
and other WOM measurement issues.
∙ Heard people say positive things
Just ask yourself, what are these attributes really
measuring? There is no link to a time frame or even
a specific WOM event, and therefore it is difficult
to understand how a category buyer comes up
with a response. If you don’t know what drives the
response, how do you interpret or act upon the
results if they change? Therefore these attributes
add clutter not value. If you really want to measure
WOM, then do this properly, don’t use a vaguely
worded abstract as a proxy.
An exception is the attribute ‘Recommended by
<insert expert relevant to category, such as Dentists
or Vets>’ which can be a CEP or baseline brand
competency in high risk categories (e.g., infant
pain relief).
In this case you are not measuring actual WOM,
but the category buyers’ perception that the
brand is endorsed by experts. This perception of
endorsement can provide the extra confidence to
make a risky purchase. When a brand gets unusually
high responses, this can usually be traced to the use
of experts in advertising or social media.
Ehrenberg-Bass Institute for Marketing Science
29
eXcellence in
X (and Y) Axes
I was fortunate enough to work
on a couple of research projects
with Andrew Ehrenberg.
One of the most frustrating things about
working with Andrew is he would first edit all
the charts and tables of any research paper,
and you would have to do those changes
before he would even look at the text. His edits
typically focused on improving the clarity of data
communication. So I quickly learnt if I wanted
him to work on any text, I had to first get the
charts and tables right.
This was one of the most valuable things I learnt
from him!
Charts are both working devices (to see the
data) and communication devices (to show the
data). Smart use of your X (and Y) axes ensures
you can do both well. When preparing charts
for other people, it should always be easier for
them to come to your conclusions than it was for
you. Well designed and labelled charts help you
30
A-Z for (Better) Brand Health Tracking
achieve this. I know there are lots of software
programs that claim to help with this, but often
what looks good can be poor communication.
For example, we sometimes take off axis
numbers to improve chart aesthetics - but pretty
does not beat useful when it comes to data
presentation (ideally you want both, but if you
have to pick one……!)
It’s quite easy to inflate or suppress perceived
data variance in a chart by judiciously choosing
the highest and lowest X or Y axis figures. This
risks giving you and any reader a misleading view
of the data. Sometimes this is done intentionally,
but often it is because the default settings on
charting programs is often designed to maximise
variance - don’t be a slave to the default.
X
Before you review any chart data from someone
else, look at the X (and Y) axes and check
they are not designed to magnify or minimise
variance. Often I see something that looks like it
varies a lot over time, only to discover the X axis
is very short or very long and/or the Y axis is
say between 3.0 and 3.5. A simple illustration of
this is share price charts for say a company like
Nestle - look at how the X and Y axes change as
you change time and how your perception of the
variance in the data also changes.
And also remember, sometimes the best data
communication device is a well-constructed
table. It does not alway have to be a chart.
Yesterday to Over a
Year Ago - Selecting Timeframes
In repeat buying categories such
as Consumer Packaged Goods,
or B2B consumables, to identify/
classify category buyers you need
to set a timeframe, such as having
bought <insert category> in the
past X months, to identify who to
survey or include in analysis. To
decide on the timeframe we turn,
not to Yoda, but to Goldilocks, for
inspiration.
However, a timeframe can also be too long.
This is when youinadvertently include lapsed
category buyers or make it to difficult for
heavy category buyers to remember their
purchases. Therefore, you get inaccurate data.
Identify your ‘Goldilocks’ timeframe
If you are unsure, experiment with different
time frames. Check the distributions and brand
repertoire sizes (number of brands bought
in the time period). At some point you will
find that taking a longer time does not mean
substantially more brands are bought, which
means you have captured the current brand
repertoire of most category buyers.
A Goldilocks time frame is neither too short or
too long. Either extreme can be detrimental to
data quality. Your time frame is too short if your
category buyer base skews to heavier category
buyers, because many light buyers have not yet
had a chance to buy. Therefore you miss data
from really important category buyers.
A ‘Goldilocks’ timeframe is long enough for
category buyers to have bought multiple
times if they want to, but short enough for the
purchases to be memorable. This means most
useful time frame for a category varies with the
category’s purchase frequency and memorability
of the buying event.
In Better Brand Health Chapter 9, there are some
timeframes and examples of categories where
they could be useful.
Y
BTW - the timeframe decision is why there
is no single penetration or Pareto figure for a
brand or category, it depends on the time frame
you take. Similar to my last post on how the
X and Y axes can distort the data, so can the
timeframe. If you say calculate Pareto share in
Packaged Goods over five years, it can get to
around 70% (see Kim et al, 2017), but if you do
that same calculation over one year it is more
likely to be between 40-60% (see Sharp et al
2019). Both are correct, the question you need
to ask yourself is whether one year or five years
is a more useful time frame for your calculation.
Ehrenberg-Bass Institute for Marketing Science
31
Zealotry versus Zest
For the last post in this series, let’s first turn
to Frank to create the mood:
And now the end is here
And so I face the final letter
Unfortunately, it is a Z
A D, M, or P, would be much better
I posted each day, some short, some long
I even created, an anti-brand love song
In the pursuit, of Better Brand Health
Finally the last posting day
For my last post I want to encourage you to reject the Zealotry
of fanatically believing in an idea regardless of the evidence, in
favour of a Zest for learning, wanting to know more. Marketing
has a history of embracing and championing ideas that seem
good on paper, but crumble under scrutiny.
You can have a Zeal for a new idea, but ask for evidence,
look for evidence, and if necessary, create the environment
for evidence to emerge - and be prepared to temper or even
turn off your Zeal, if the evidence doesn’t appear. That makes
you stronger as a marketing professional, and strengthens all
of us, as a discipline.
I hope these posts and Better Brand Health lead you to improve
your marketing research. As we get better quality data, we
make better decisions now, and lay the foundations to learn
even more in the future. Our R&D is ongoing but the quality
of data we have is a big contributor to how much and how
quickly we learn.
Z
How you can engage
as an individual
How Brands Grow – for Executives
Whether you are a C-Suite marketing
executive with a global team, a
business owner looking to expand
your footprint, or an aspiring young
leader – this experience is for those
who want to achieve evidence-based
growth.
How Brands Grow - for Executives equips you, the
decision-maker, with the critical evidence and
tools to implement world’s best-practice marketing
techniques to brands.
Gain insights from the best in the business through
interactive workshops and dynamic discussions.
Learn more
32
A-Z for (Better) Brand Health Tracking
Delivered by world-class Ehrenberg- Bass Institute
marketing experts, this event leverages the most
up-to-date findings to show you:
∙ The strategies that will (and won’t) lead to
sustainable brand growth
∙ The real-world value of various marketing
interventions
∙ How to incorporate evidence-based knowledge
into your strategic decision-making and planning
processes while bringing this knowledge back with
you to up- skill your department
Our impact is global.
How you can engage
as a company
The Ehrenberg-Bass Institute are considered an authority on marketing by many of the world’s biggest
brands. They provide companies with the tools and knowledge to grow brands and develop smarter,
evidence-based marketing teams.
Ehrenberg-Bass Sponsorship
Ehrenberg-Bass Sponsorship
offers exclusive access to a multimillion dollar research R&D program
in exchange for an annual financial
contribution. With these funds
our team of experts investigate
the questions that matter most to
marketers.
The result is a huge body of research translated into
practical insights that are scientifically proven to
apply to every brand, across all markets, anywhere
in the world.
Only our Sponsors have access to all of our
important research, our latest findings, as well as
the support they need to apply it to their business.
Joining as a Sponsor will make your marketing
team more efficient and effective. We teach
scientific marketing laws, and how to apply that
knowledge to grow brands, gain market share
and increase sales.
The Ehrenberg-Bass Institute
actually is the language of the
C-suite. It does lean more towards
the scientific, empirical, economic
language that lends more
credibility in the boardroom.
Vice President Marketing ANZ,
Unilever
Learn more
34
A-Z for (Better) Brand Health Tracking
Ehrenberg-Bass Institute for Marketing Science
35
References
A is Attitude
I is for Intentions-to-buy
Romaniuk, J. (2023). Brand Attitude, Chapter 8, Better Brand Health. Australia, Oxford University Press.
Juster, F. T. (1966). “Consumer buying intentions and purchase probability: An experiment in survey design.” Journal of American Statistical Association 61(315): 658-696.
B is for B2B
Wright, M. and M. MacRae (2007). “Bias and variability in purchase intention scales “ Journal of the Academy of Marketing Science 35(4): 617-624.
Romaniuk, J. (2023). The rise of the machines?, Chapter 13, Better Brand Health. Australia, Oxford University Press.
Romaniuk, J. (2023). Brand Buying, Chapter 10, Better Brand Health. Australia, Oxford University Press.
C is for Category
J is for Jobs To Be Done (JTBD)
Romaniuk, J. (2023). Applying the laws of growth to brand health tracking, Chapter 1, Better Brand Health. Australia, Oxford University Press.
Christensen, C. M., et al. (2016). Know Your Customers’ “Jobs to be Done”. United States, Harvard Business Press.
D is for Distinctive Assets
Romaniuk, J. (2021). Building Mental Availability. How Brands Grow: Part 2. J. Butler. Victoria, Australia, Oxford University Press: 61-84.
Romaniuk, J. (2023). Brand attribute selection, Chapter 3, Better Brand Health. Australia, Oxford University Press.
Romaniuk, J. (2023). Mental Availability and Category Entry Points, Chapter 5, Better Brand Health. Australia, Oxford University Press.
Romaniuk, J. (2018). Building Distinctive Brand Assets. South Melbourne, Victoria, Oxford University Press.
K is for Key Performance Indicators (KPIs)
E is for Exposure to Executions
Romaniuk, J. (2023). Applying the Laws of Growth to Brand Health Tracking, Chapter 1, Better Brand Health. Australia, Oxford University Press.
Romaniuk, J. (2023). Brand attribute selection, Chapter 3, Better Brand Health. Australia, Oxford University Press.
L is for Love (Brand Love)
Harrison, F. (2013). “Digging Deeper Down into the Empirical Generalization of Brand Recall.” Journal of Advertising Research 53(2): 181-185.
Romaniuk, J. (2023). Brand Attitude, Chapter 8, Better Brand Health. Australia, Oxford University Press.
Vaughan, K., V. Beal and J. Romaniuk (2016). “Can brand users really remember advertising more than nonusers? Testing an empirical generalization across six advertising awareness
measures.” Journal of Advertising Research 56(3): 311-320.
M is for Memory Building
G is for Good Measures
Romaniuk, J. (2023). Applying the Laws of Growth to Brand Tracking, Chapter 1, Better Brand Health. Australia, Oxford University Press.
H is for Handling the Haters
East, R., et al. (2007). “The relative incidence of positive and negative word of mouth: a multi-category study.” International Journal of Research in
Marketing 24(2): 175-184.
Romaniuk, J. (2023). Exposure to Marketing Activity, Chapter 11, Better Brand Health. Australia, Oxford University Press
N is for Non-buyers
Romaniuk, J. (2023). Brand Attribute Measurement, Chapter 4, Better Brand Health. Australia, Oxford University Press
O is for Ownership
Romaniuk, J. and R. East (2021). Word-of-Mouth Facts Worth Talking About. How Brands Grow: Part 2. J. Butler. Victoria, Australia, Oxford University Press: 119-138.
Romaniuk, J. and E. Gaillard (2007). “The relationship between unique brand associations, brand usage and brand performance: Analysis across eight categories.” Journal of
Marketing Management 23(3): 267-284.
Romaniuk, J. (2023). Word-of-mouth measurement, Chapter 12, Better Brand Health. Australia, Oxford University Press.
Romaniuk, J. (2018). Building Distinctive Brand Assets. South Melbourne, Victoria, Oxford University Press.
Ehrenberg, A. (1988). Repeat-buying: Facts, theory and applications. London, Oxford University Press.
36
A-Z for (Better) Brand Health Tracking
Ehrenberg-Bass Institute for Marketing Science
37
P is for Prominence, a Pillar of Physical Availability
U is for U-shaped distributions
Nenycz-Thiel, M. and J. Romaniuk (2021). Building Physical Availability: Prominence and Portfolio. How Brands Grow: Part 2. J. Butler. Victoria, Australia, Oxford University Press: 159-172.
Ehrenberg, A. S. C. (1959). “The pattern of consumer purchases.” Applied Statistics 8(1): 26-41.
Nenycz-Thiel, M. and J. Romaniuk (2023). What about Physical Availability, Chapter 14, Better Brand Health. Australia, Oxford University Press
Schmittlein, D. C., et al. (1985). “Why does the NBD model work? robustness in representing product purchases, brand purchases and imperfectly recorded purchases.” Marketing Science
4(No. 3, Summer): 255-266.
Q is for is Quicker always better?
Romaniuk, J. (2023). Brand attribute measurement, Chapter 4, Better Brand Health. Australia, Oxford University Press
R is for Rating Scales
Romaniuk, J. (2023). Brand attribute measurement, Chapter 4, Better Brand Health. Australia, Oxford University Press
Anderson, J. R. and G. H. Bower (2013). Human associative memory, Psychology Press.
Barnard, N. R. and A. Ehrenberg (1990). “Robust Measures of Consumer Brand Beliefs.” Journal of Marketing Research 27(4): 477-484.
Driesener, C. and J. Romaniuk (2006). “Comparing methods of brand image measurement.” International Journal of Market Research 48(6): 681-698.
Romaniuk, J. (2008). “Comparing methods of measuring brand personality traits.” The Journal of Marketing Theory and Practice 16(2): 153-161.
S is for Sample Screening questions
Romaniuk, J. (2023). Category buying behaviour, Chapter 9, Better Brand Health. Australia, Oxford University Press
Schoenmueller, V., et al. (2020). “The polarity of online reviews: Prevalence, drivers and implications.” Journal of Marketing Research 57(5): 853-877.
Sharp, A and J. Romaniuk, (2023). The rise of the machines? Chapter 13, Better Brand Health. Australia, Oxford University Press
V is for Valence in word-of-mouth effects
Romaniuk, J. and R. East (2021). Word-of-Mouth Facts Worth Talking About. How Brands Grow: Part 2. J. Butler. Victoria, Australia, Oxford University Press: 119-138.
East, R., et al. (2008). “Measuring the impact of positive and negative word of mouth on brand purchase probability.” International Journal of Research in Marketing 25(3): 215-224.
Romaniuk, J. (2023). Word-of-mouth measurement. Chapter 12, Better Brand Health. Australia, Oxford University Press
W is for Word-of-mouthish attributes are a Waste of time
Romaniuk, J. (2023). Word-of-mouth measurement. Chapter 12, Better Brand Health. Australia, Oxford University Press
Romaniuk, J. (2023). Brand attribute measurement. Chapter 4, Better Brand Health. Australia, Oxford University Press
X is for eXcellence in X (and Y) axes
T is for Top-of-mind Brand Awareness
Ehrenberg, A. (2000). “Data reduction - Analysing and interpreting statistical data.” Journal of Empirical Generalisations in Marketing Science 5: 1-391.
Romaniuk, J. (2023). Brand awareness, Chapter 2, Better Brand Health. Australia, Oxford University Press
Ehrenberg, A. and J. A. Bound (2000). Turning data into knowledge. Marketing research: State of the art perspectives. C. Chakrapani. Chicago, IL, American Marketing Association: 23-46.
Romaniuk, J. and B. Sharp (2004). “Conceptualizing and measuring brand salience.” Marketing Theory 4(4): 327-342.
Romaniuk, J. (2021). Building Mental Availability. How Brands Grow: Part 2. J. Butler. Victoria, Australia, Oxford University Press: 61-84.
Y is for from Yesterday to over a Year ago - selecting timeframes
Sharp, B., et al. (2019). “Marketing’s 60/20 Pareto Law.” Social Science Research Network: 1-5.
Kim, B. J., et al. (2017). “The Pareto rule for frequently purchased packaged goods: an empirical generalization.” Marketing Letters 28(4): 1-17.
Romaniuk, J. (2023). Category buying behaviour. Chapter 9, Better Brand Health. Australia, Oxford University Press
38
A-Z for (Better) Brand Health Tracking
Ehrenberg-Bass Institute for Marketing Science
39
Ehrenberg-Bass Institute
University of South Australia
City West Campus
Level 4, Yungondi Building
70 North Terrace
Adelaide, SA 5000
Australia
www.MarketingScience.info
info@MarketingScience.info
/ehrenbergbass
@ehrenbergbass