Market Segmentation

Text-only Preview

Market Segmentation
By Jerry W. Thomas, Decision Analyst
When the term “market segmen- focus. Before discussing psychograph-
tation” is used, most of us im-
ic or lifestyle segmentation (which
mediately think of psychographics,
is what most of us mean when using
The purpose of
lifestyles, values, behaviors, and multi-
the term “segmentation”), let’s review
segmentation is
variate cluster analysis routines. Mar-
other types of market segmentation.
the concentra-
ket segmentation is a much broader
Our focus is on consumer markets
tion of market-
concept, however, and pervades the
rather than business markets.
ing energy and
practice of business throughout the
force on the
Geographic Segmentation
subdivision (or
the market seg-

What is market segmentation? At its
This is perhaps the most common
ment) to gain a
most basic level, the term “market
form of market segmentation, wherein
segmentation” refers to subdividing
companies segment the market by
a market along some
attacking a restricted
within the
commonality, similar-
geographic area. For
ity, or kinship. That
example, corpora-
is, the members of a
tions may choose to
market segment share
market their brands in
something in common.
certain countries, but
The purpose of seg-
not in others. A brand
mentation is the con-
could be sold only in
centration of market-
one market, one state,
ing energy and force
or one region of the
on the subdivision
United States. Many
(or the market seg-
restaurant chains
ment) to gain a competitive advantage
focus on a limited geographic area to
within the segment. It’s analogous to
achieve concentration of force. Re-
the military principle of “Concentra-
gional differences in consumer prefer-
tion of Force” to overwhelm an enemy.
ences exist, and this often provides a
Concentration of marketing energy (or
basis for geographic specialization.
force) is the essence of all marketing
For example, a company might choose
strategy, and market segmentation is
to market its redeye gravy only in the
the conceptual tool to help achieve this
southeastern U.S. Likewise, a picante
604 Avenue H East
Arlington, TX 76011-3100, USA
The global leader in analytical research systems
(1) 817.640.6166 or 1.800. ANALYSIS
© 2007 Decision Analyst, Inc.

sauce might concentrate its distribution tion is sometimes a possibility. It is
and advertising in the southwest. A
based on the fact that different media
If a brand pours
chainsaw company might only market
tend to reach different audiences. If a
all of its budget
its products in areas with forests. Geo-
brand pours all of its budget into one
into one media,
graphic segmentation can take many
media, it can possibly dominate the
it can possibly
forms (urban versus rural, north versus
segment of the market that listens to
dominate the
south, seacoasts versus interior, warm
that radio station or reads that maga-
segment of the
areas versus cold, high-humidity areas
zine. Media segmentation is most
market that
versus dry areas, high elevation versus
often practiced by companies that have
listens to that
low elevation areas, and so on). These
some control over the media, and can
radio station or
examples also reveal that geographic
somehow discourage competitors from
reads that
segmentation is sometimes a surrogate
using that media.
for (or a means to) other types of seg-
Price segmentation
Distribution Segmentation
Price segmentation is common and
widely practiced. Variation in house-
Different markets can be reached
hold incomes creates an opportunity
through different channels of distribu-
for segmenting some markets along a
tion. For example, a company might
price dimension. If personal incomes
segment the “tick and flea collar” mar-
range from low to high, the reason-
ket by selling the product to supermar-
ing goes, then a company should offer
kets under one brand name, to mass
some cheap products, some medium-
merchandisers under another brand, to
priced ones, and some expensive ones.
pet stores under another brand name,
This type of price segmentation is well
and to veterinarians under yet another
illustrated by the range of automo-
brand name. This type of distribution-
tive brands marketed by General Mo-
al segmentation is common, especially
tors, historically. Chevrolet, Pontiac,
among small companies that grant
Oldsmobile, Buick, and Cadillac var-
each channel a unique brand to gain
distribution within that channel. Other
examples of distributional segmenta-
tion would be an upscale line of cloth-
ing sold only in expensive department
stores, or a hair shampoo sold only
through upscale beauty salons.
Media Segmentation

While not common, media segmenta-

ied in price (and status) along a clearly
defined spectrum to appeal to succes-
sively higher income groups.
Markets can be
segmented by

Demographic segmentation
hobbies, by po-
litical affiliation,

Gender, age, income, type housing,
by religion, by
and education level are common de-
special interest
mographic variables. Some brands are
groups, by sports
targeted only to women, others only
team loyalties,
to men. Music downloads tend to be
by university at-
targeted to the young, while hearing
tended, and hun-
aids are targeted to the elderly. Edu-
dreds of other
cation levels often define market seg-
variables. You
ments. For instance, private elementary
are only limited
schools might define their target market interesting basis for segmentation. In
by your market-
as highly educated households contain-
addition to the foregoing, markets can
ing imagination.
ing women of childbearing age. Demo-
be segmented by hobbies, by political
graphic segmentation almost always
affiliation, by religion, by special inter-
plays some role in a segmentation
est groups, by sports team loyalties, by
university attended, and hundreds of
other variables. You are only limited by
Time segmentation
your marketing imagination.
Time segmentation is less common, but Psychographic or Lifestyle
can be highly effective. Some stores
stay open later than others, or stay open
on weekends. Some products are sold
Lastly, we come to psychographic (or
only at certain times of the year (e.g.,
lifestyle) segmentation, based upon
Christmas cards, turkeys, fireworks,
multivariate analyses of consumer at-
cranberry sauce). Chili is marketed
titudes, values, behaviors, emotions,
more aggressively in the fall, with the
perceptions, beliefs, and interests. Psy-
onset of cooler weather. Football is
chographic segmentation is a legitimate
played in the fall, basketball in the win- way to segment a market, if we can
ter, and baseball in the spring and sum-
identify the proper segmentation vari-
mer (or at least this used to be the pat-
ables (or lifestyle statements, words,
tern). The Olympics come along every pictures, etc.). Qualitative research
four years. Department stores some-
techniques (focus groups, depth inter-
times schedule midnight promotional
views, ethnography) become invalu-
events. The time dimension can be an
able at this stage. Qualitative research

different airlines; that is, their “brand
images.” You could go further and add
Verbatim com-
a section on media consumption, or per-
ments from con-
sonal values as well. It is at this point
sumers are used
that you realize the questionnaire is too
to build batteries
long, and you have to make some hard
of psychographic
decisions about what questions or state-
or lifestyle state-
ments to include.
ments (these two
terms are used

The method of data collection is very
provides the insight, the conceptual
important, because the question-
knowledge, and the consumer’s exact
naire is so long (often 45 to 90 min-
language necessary to design the seg-
utes in length). The telephone is not
mentation questionnaire. Typically,
recommended for segmentation stud-
verbatim comments from consumers are ies because of questionnaire length.
used to build batteries of psychographic Moreover, the various rating scales and
or lifestyle statements (these two terms attitudinal statements are difficult to
are used interchangeably). A large rep-
communicate by phone, and the result-
resentative sample of consumers (gen-
ing phone data tends to be “insensi-
erally, a 1,000 or more) is then asked
tive” and rift with “noise.” In-person
about the degree to which they agree or interviews, or Internet-based interviews,
disagree with each statement.
or even mail surveys, are much better.
Rating scales and attitudinal statements
For example, if you were designing a
can be seen and fully comprehended by
market segmentation questionnaire for
respondents. Seeing is much better than
an airline, you might conduct a series
hearing, and it produces more accurate
of depth interviews to help design the
answers. The Internet is especially
questionnaire. You probably would
valuable for segmentation studies, since
include a behaviorial section (frequency respondents can take the survey at a
of flying, how purchase tickets, who
time of their own choosing, when they
travel with, cities flown to, where sit,
can give it their full, undivided atten-
airlines flown, money spent on airline
tion. A mail survey offers some of the
tickets, etc.). You would include a
same advantages, but without the ques-
major section on attitudes toward air
tionnaire controls, checks, and safe-
travel (motivations for air travel, fears
guards built into an Internet survey.
related to air travel, positive emotions
of flying, attitudes about airline employ- Analytical Methods
ees, checking luggage, buying tickets,
and so forth). You would also want to
Most segmentation analyses are based
include a section on perceptions of the
upon various types of “cluster analy-

sis,” a set of well-defined statistical
group the statements at this point. The
procedures that group people according final step is to attach a segment code to
to the proximity of their ratings. Un-
each market segment identified and then A good psycho-
fortunately, cluster analysis (regard-
cross-tab all of the questionnaire vari-
graphic seg-
less of its many types and forms) has
ables by the segments. You must then
mentation is to
inherent limitations and seldom yields
study the segments and the attitudes/
first identify the
coherent market segments. Cluster
statements that make up each segment
statements that
analysis routines ignore the pattern of
to make sure they make sense and hang
are more impor-
respondent ratings and rely primarily
together. If the segmentation results
tant (i.e., the
upon the proximity of respondent rat-
don’t make sense, then you have to go
statements that
ings. Too often, this leads to clusters,
back, change some of your assumptions
tend to explain
or market segments, that don’t seem to or methods, rerun the analysis, and re-
or cause specific
make much sense when cross-tabulated peat the cross-tab exercise to apply the
against the original segmentation vari-
“common sense” validity check.
ables. Another limitation of clustering
approaches is that all statements are
Common Mistakes
treated as equal; whereas, in truth, some
statements might be much more impor-
Segmentation studes tend to be large
tant than others in explaining consumer and complicated, so it’s easy for errors
behavior in a particular product cat-
and mistakes to be made. Some of the
most common mistakes:
A better way to achieve a good psy- 1. Segmenting a segment.
chographic segmentation is to first For example, someone might want to
identify the statements that are more
segment the market for widgets
important (i.e., the statements that tend among 18 to 24-year-olds, who live
to explain or cause specific consumer
in Vermont and buy brand XYZ. As
behaviors). Correlation analysis and
is evident, the client is asking that a
regression can be used for this purpose.
Factor analysis is also a powerful tech-
nique to identify the statements and
groups of statements that account for
much of the variance in the attitudinal
dataset. Directly, and indirectly, these
techniques can help you identify the
most important statements (i.e., atti-
tudes, perceptions, values). Then, these
statements become the inputs to the fi-
nal segmentation analysis. Many differ-
ent methods can be used to “cluster” or

tiny sliver of the market be segment- you run the risk that the resulting
ed. True, this tiny sliver can be seg-
segments will be too small to target,
The market
mented, but rarely are the resulting
at least by mass media. This is not
should be broad-
segments of any value, because they always true, but it is a good rule of
ly defined for a
are just too small. General rule: seg- thumb.
ment the whole market, including all
analysis to be
age groups. The market should be
4. Targeting all segments.
most effective.
broadly defined for a segmentation
So you have carefully subdivided
In other words,
analysis to be most effective. In
your target market into five mutually
don’t preordain
other words, don’t preordain the re-
exclusive psychographic segments,
the results by
sults by sampling restrictions.
and your boss tells you to develop a
marketing plan to attack each seg-
2. Overlooking the “universals.”
ment. If all of your marketing is
Many attitudinal statements in the
direct mail, and you can identify the
questionnaire will not show up in
addresses that belong to each seg-
the final segments, because they tend ment, then you can attack all seg-
to be the same across all segments.
ments (assuming your product is rel-
Statements that everyone agrees with, evant to all sements). But, if you use
or everyone disagrees with (we call
broadcast media in marketing your
them “universals”), cannot explain
product, it is very difficult to target
much in the multivariate analyses.
multiple segments, because of
Variables have to move up and down media “spillover.” What you say to
or the multivariate analysis to work. one segment will be muddled and
The highest rated variables, and the
confused by the different messages
lowest rated, are likely to fall out of
targeted to other segments.
the multivariate analyses. However,
you should always look at these
5. Confusing the results.
“universal” statements. Anyone of
Segmentation studies are large and
them might be the basis for a posi-
complicated, with enormous amounts
tioning or a strategy that would ap-
peal to everyone. If you find some-
thing unique that appeals to every-
one, the heck with segmentation. Go
for the whole hog.
3. Creating too many segments.
There is a practical limit to the size
of segments that companies can
effectively target. If you create more
than four or five market segments,

of data. It is easy to get lost in this
treasure trove of answers and come
up with confused and baffling results.
Often, hidden
in plain view

6. Overlooking the basics.
in the plain old
The dazzle and glitter of the ad-
cross-tabs, are
vanced, rocket-science multivariate
analyses attract everyone’s attention.
findings that
No one ever opens up the crosstabs
could form the
and looks at the answers to the hun-
basis for new or
dreds of questions asked. Often,
hidden in plain view in the plain old
marketing strate-
cross-tabs, are tremendous findings
gies, advertising
that could form the basis for new or
campaigns, or
improved marketing strategies, ad-
new products.
vertising campaigns, or new prod-
ucts. Rarely does anyone analyze this exclusive segments and one based upon
basic data, however.
overlapping segments. Both of these
segmentation “solutions” should be
7. Targeting people instead of dollars. cross-tabulated by the original question-
A market segment might represent a
naire variables to identify which type
large percentage of the population,
of solution yields the most meaningful
but a small part of the market.
(and actionable) market segments.
Always look at the dollar potential of
market segments, not just the number Final Thoughts
of people in the segments.
Nonmutually Exclusive Segments
The concept of market segmentation
is sound. It’s a way to apply great-
er marketing energy or force to a subset

of the market. A great deal of money
Virtually all segmentation work, his-
torically, has been based upon the is wasted on psychographic segmenta-
assumption of mutually exclusive mar-
tions that never lead to any marketing
ket segments. The mutually exclusive
model, however, does not always apply
to psychographic or lifestyle segmenta- If you segment the market by psycho-
tion (since most of us hold many over-
graphics, there are several essential
lapping and/or conflicting beliefs and
uses of the segmentation: first, target
attitudes). Therefore, it is wise to de-
your brand to the largest segment with
velop two distinctly different segmenta- relevant brand fit (or even target two
tion solutions: one based upon mutually closely related segments) by media

advertising and message. That is, the advertis-
chographic segment. That is, the market segments
ing message is the way to reach the psychographic can be a template for new product development.
segment (rarely can a psychographic segment be
For example, if you find that 15% of the U.S. popu-
defined by demographics or geography). Second, lation belongs to a “safety first” segment when it
segmentation can provide the guiderails for brand comes to buying cars, then you can design and build
positioning. That is, positioning assumes, or takes the safest car in the world to target this segment. So
place in relation to, a target market segment; you
psychographic segmentation’s greatest value lies in
are positioning your brand in relation to a market
positioning, targeting via advertising message, and
segment. Third, the segmentation can define op-
defining new product opportunities. Go forth and
portunites for new products targeted to each psy-
The author, Jerry W. Thomas, is President/CEO at Dallas-Fort Worth-based Decision Analyst
(, a leading international marketing research and marketing con-
sulting firm. The company specializes in advertising testing, strategy research, new product
development, and advanced modeling for marketing decision optimization. Decision Analyst
also operates the American Consumer Opinion® online panel, one of the world’s largest
Internet consumer opinion panels, with more than six million participants.
To contact Jerry W. Thomas, please call 1.800.262.5974 or email him at
[email protected]
604 Avenue H East
Arlington, TX 76011-3100, USA
The global leader in analytical research systems
(1) 817.640.6166 or 1.800. ANALYSIS
© 2007 Decision Analyst, Inc.