Benefit Segmentation:

A Decision-oriented
Research Tool
segmentation has been steadily moving toward center
MARKET
stage as a topic of discussion in marketing and research

RUSSELL I. HALEY

circles. Hardly a conference passes without at least one session

devoted to it. Moreover, in March the American Management
Association held a three-day conference entirely concerned with
various aspects of the segmentation problem.
According to Wendell Smith, IlS egmentat ion is based upon developments on the demand s ide of the market and represents a
rational and more precise adjustment of product and marketing
effort to consumer or user requirements.'" The idea that all markets can be profitably segmented has now received almost as widespread acceptance as the marketing concept itself. However, problems remain. In the extreme, a marketer can divide up his market
in as many ways as he can describe his prospects. If he wishes,
he can define a left-handed segment, or a blue-eyed segment, or a
German-speaking segment. Consequently, current discussion revo lves la rgely around which of the virtually limitless alternatives
is likely to be most productive.

According to this article,

most techniques of market
segmentation rely only on
DESCRIPTIVE factors pertaining to purchasers and are
not efficient predictors of
future buyer behavior. The
author proposes an approach whereby market segments are delineated first
on the basis of factors with
a CAUSAL relationship to
future purchase behavior.
The belief underlying this
segmentation strategy is
that the benefits which people are seeking in consuming
a given product are the
basic reasons for the existence of true market segments.

Journal of MaTketiflU.
1968). 'Pp. 3035 ,

Vol.

32

(July.

Segmenlat;otl Methods
Several varieties of market segmentation have been popular in
the r ecent past. At least three kinds have achieved some degree
of prominence. Histo rically, perhaps the fU'st type to exist was
geographi c segmentation. Small manufactu r er s who wished to
limit the ir investments, or whose distr ibution channels were not
large enough to cover the entire country, segmented the U. S.
market, in effect, by selling their products on ly in certain areas.
However, as more and more brands became national, the second major system of segmentation--ciemograph ic segmentation-became popu lar. Under this philosophy targets were defined as
yo unger people, men, or fam ilies with children. Unfortunately, a
number of recent studi es have shown that demographic variables
such as age, sex, income, occupation and race are, in general, POOl'
predictors of behavio r and, consequently, less than optim um bases
for segmentation strategies.2
1

Wendell R. Smith , "Product Differentiation a nd Market Segmentation

as Alternative Product Strategies," J OURNA.L OF MARKETI NG, Vol. XXI
(Ju ly, 1956), pp. 3-8.
Rona ld E. Frank, HCorrelates of Buyi ng Behavior for Groce ry Products," JOURN AL OF MARKETING , Vol. 31 (Octobe r, 1967), pp. 48-53; Ronald E. Frank, William Massy, a nd Harper W. Boyd, Jr ., "Cor re lates
of Grocery Product Consumption Rates," Journal of Marketing R esea.rch, Vol. 4 (May, 1967), pp. 184-190; and Clark Wil son, "Home_
maker Living Patterns and Marketplace Behavior-A Psychometric
Approach," in John S . Wright and Jac L. Goldstucker, Editors, N ew
Ideas for Succe88ful Marketing, P roceedings of 1966 Wol'ld Consrrcss
(Chicago: American Marketing Association, June, 1966), pp. 305-331.

30

31

Benefit Segmell tation: A Decision-oriented Research Tool

More recently, a third type of segmentation has

come into increasing favor-volume seg mentati on.

The so-called '4heavy half" t heory, popularized by

Dik Twedt o[ t he Oscar Mayer Co mpany," points
out that in most product catego ri es one-half of the
consumers account for around 80 q. of the consumpt ion. I f this is true, the argument goes, shou ldn 't
knnwledgcable murketc)'s concent rate their eft"olts
on the.se high-volume consumers '! Ce rta inly they
are the most valuable co nsu me rs .
The trouble with this line of reason ing is that not
a ll heavy consumers il re llsually avai lab le to the same
brand-because they a re not all seeking the same
kinds of be nefits from a product. F or exa mple,
heavy coff ee drinkers cO I1::;ist of two types of COIlsumers- those who d r ink chai n sto re brands and
t hose who dr ink premium bra nd s. The chai n store
custo mers feel that all coffees are basically alike and.
because they drink .so much coffee, t hey feel it is
sensible to buy a relcltively inex])ensive brand. The
prernium orand buyers, all the othe r hal1(1. feel that
the few added pennies which coffees like YubHI1.
Martinson's, Chock Full O'Nuts, and Savarin cost
a re more than justi fi ed by their fulle)' taste. Obviously, these two g roups of people, although they are
both members of the "heavy half" segment. a re not
equally good prospects for anyone brand, nor can
they be expected to respond to the same advert ising
cla ims.
These three systems of seg mentati on have 'been
lIsed because they provide helpful guidance in t he
use of certain marketing tools. For example, geognlphic segmentat ion . because it describes the market in a di screte way, provides definite direction in
media purchases. S I)Qt TV. spot radio, Clnd newspapers CHn be bought for the geographical segment
selected fo), co ncentrated effort. S imilar ly. demographi c segmentation allows media to be bought
more efJicie nt ly since demog raphi c data on readers,
viewe rs, and listeners are readily ava ilable fo)' most
media veh icles. Also. in some product catego ri es
demographic var iab les are ext remely helpful in differentiating users fro m non-users, although they are
typiC~lll.v less helpful in distinguishing between the
usel's of var ious brands. The h eav ~t- h a lf philosophy is especia lly effective in directing dollars tow~lrd
the most important parts of the market.
Howeve r, each of these three systems of segmentatio n is handicapped by an und erlying di sadvantage
inherent in its nature. All are b~lsed on an ex-post
facto anHiysis of the kinds of people who make up
var ious segments of a nUlIket. They rely on de.'1c'r iptive factors rather than cawml factors.
For
this reason t hey are not efficient predict ors of future
:J

Oik Wurren 'fw edt, "So me Practical Applications of

the 'Heavy Half' Theo r y" (New York: Ad vert ising
Research Foundation 10th Annual Confe rence, Octobe,' 6, 1964).

buying behHvior, and it is future buying behavior

is of central inte rest to marketers.

t h ~lt

Benefit Segmentation
An approach to mnrket segmentation whereby it
is possible to identify market segments by c;Hlsal
factol's J'Hthel' than descriptive factors, might be
called "benefit segmentation." The belief underlying this segmentation strategy is that the benefits which people Hre seek in g in consuming a given
product Hre t he basic reHSOIH:\ for the existence of
true market segments. Experience with this approach has shown that benefits sought by consumers
determine their behavior mu ch more accllrately
than do demographic charncteristics or volume of
consumption.
This does not mean that the ki nds of dabl
gathered in mOre traditional types of segmentati on
are not useful. Once people ha ve been classified
into segments in accordullce with the benefits they
are seeking. eac h segment is cont rasted with all of
t he othe r segments in terms of its demography, its
vo lume of consumption, its brand perceptions, itt:;
media habits, its perso nality and life-style, a nd so
forth. In this way. a reaso nab ly deep understa ndin g
of t he people who make up eHch segment can be
obta ined. And by capitali zing on t hi s undel'sblllding, it is possible to reach them, to talk to them in
their OW11 terms, and to present a product in t he
most favo .. able ligh t possible.
The benefit segmentatio n approach is not new. 1t
has been employed by a number of Ame ri ca's largest
co rporati ons s ince it was introduced in ]961." Howeve r, case hi t ori es have been notably absent from
the Iitel'ature because most studies have been contl'acted for privately. and have been treated confidentially.
.\ RU$seli 1. Haley, HExperimental Research on Attitudes Towu rd Shampoos," an unpublished papel' (February, 196t).

ABOUT THE AUTHOR. Russell I.

Haley is Vice Presi dent and Corporate
Research Director of D'Arcy Advertising in New York City. Prior to his current position he was Vice Presid e nt and
Associate Director of the Marketing a nd
Resea rch Department at Grey Adve rtising. While there. he developed new
methods for measuring attitudes. a
unique way of segmenting markets by
attitude patterns. and improved methods
of conducting large-sca le market tests. Mr. Haley received his
M.B.A. from Columbia in marketing and statistics.
Mr. Haley is a past president of the Cleveland Chapter of
the American Statistical Associ ation. He is a member of the
American Market ing Association . the Americdn Associdt ion for
Public Opinion Research. dnd the Executive Committee of the
Copy Research Council. Mr. Haley is currently Chairma n of an
A.R.F. CommiHee dealing with attitude measurement and i~
teaching at Rutgers University.

--32
The be nefit segmentnti on approach is based upon
being able to meas ure consu mer value sy tern s in
deta il, t ogethel" wi th wh at the consum er thinks about
various brands in the prod uct catego ry of interest.
While this concept seems s imple enough, operat iona lly i t is very complex. There is no s imple straigh tforward way of handling the volumes of data that
have to be generated. Compu ters and sophi sti cated
multi vari ate at t itude measurement techni ques ar e a
necessity.
Severa l alternative stati sti cal approaches can be
employed, a mong them t he so-called " Q" technique
of f actor ana lys is, multi-d imens ional scaling, and
other distance measures.!) All of these methods relate th e rat in gs of eac h respond ent to t hose of eve ry
other r espondent and t hen seek clusters of individuals with s irni)nr rat ing pattern s. If t he items
rnted are potentia l cons umer benefi ts, the clu ster s
th at emerge will be g rou ps of people who attach
s imila r degl'ees of impOI'ta nce to the va ri ous benefits ,
Whatever the stati stical approach selected , t he end
res ult of t he ana lys is is likely to be between t h ree
and seven cons umeI' segments, each represent ing a
potentially produ ctive foca l point fo r marketing
efforts.
Each segment is identified by t he benefi ts it is
seeking, H owever , it is th e total confiU'lt'1'ution of
the benefits sought which differentiates one segment
from another, ra t her th an t he fact t hat one seg m ~ nt
is seeking one particu la r benefit and another a quite
different benefi t. Individua l benefits are likely to
have appeal for severa l segments. I n f act . th e research th at has been done t hus f ar suggests t hat
most people would like as many benefits as possible.
H owever , the '1'elative importance they at tach to ind ividual benefits can d iffer impOltant ly and, accordingl y, can be used as an effecti ve lever in segmenting ma rkets.
Of course, it is poss ible to det ermine benefit segments int ui tively as well as with compu te rs and
sophi sti cated research methods. The ki nds of brillia nt ins ig hts whi ch pl'oduced t he Mustlmg a nd the
first IOO-millimeter cigarette have a good chan ce
of succeeding whenever market er s a re able to tap
an ex istin g benefit segment.
H owever, in tu ition can be ve ry expensive when it
is mi staken. Ma rketin g hi story is rep lete with examples of produ cts which someone f elt co uld not
miss. Over the longer term, systemat ic benefi t segmentation resea rch is likely to have a hi g her proporti on of s uccesses.
But is benefit segmenta tion practical ? And is it
truly ope rat ional ? The answer to both of these
qu estions is "yes." In effect, the cru x of th e prob~

R onald E. Frank a nd Paul E . G reen, " Numeri cal

Taxonom y in Marketin g Anal ysis: A Review Atti cle."
J ournal of Marketing Research, Vol. V ( F eb ruary,
1968), pp. 83-98.

Journal o f M arket;IIg, J uly, 1968

tern of choosing t he best segmentati on system is to

determ ine wh ich has the g reatest number of practi cal marketing implicati ons. An example should
show that benefit segmentat ion has a muc h wider
range of implicati ons th an alte rn at ive fOl'ms of
segmentat ion.
A n Example o f Benefit Segm entation
While the materi a l presented here is purely ill ustrati ve to protect th e competit ive edge of companies
who have invested in stud ies of thi s kind, it is based
on actu a l segmen tat ion studi es. Consequ ently. it is
quite ty pical of the kind s of t hings which a re normally lea rn ed in the co urse of a benefi t segmentati on study.
The toothpaste market has been chosen as a n example because it is one with which everyone is
familial'. Let us ass ume th at a benefit segmentati on
study has bee n done Hnd 'fOU l' major segments have
been identified--one pa rti cularly concerned with decay prevention, one with bri g htness of teeth. one
with t he flavor and appea ran ce of t he prod uct, and
one with price. A relati vely la rge amou nt of supplementa ry informat ion has a lso bee n gat hered
(Tab le 1) about t he people in each of t hese segments.
The decay preve ntion segment. it has bee n f ound,
contains a di sproportionately lar ge number of families w ith childl'en. They are seriollsly concerned
about the possibi lity of cavit ies and show a de fin ite
preference for fl uo ride toothpaste. Thi s is r einforced by t heir pe rso na lities . T hey tend to be a
little hypoc hondr iacal and, in t heir li fe-st yles, t hey
are less socially-o ri ented t han some of t he other
g rou ps. This segment has been named The Worri ers.
The second segment. comprised of people who show
concern fo r th e bri ghtness of t heir teeth, is quite
d ifferent . It includes a r elat ively large grou p of
young marri eds. They smoke more than ave r age.
This is wher e the swin ge rs are. They are strong ly
social and their life-style pa t te rn s are very active.
This is probably th e g rou p to whi ch toothpastes such
as Macleans or Pl us White or U ltra Brite would
appeal. Th is segment has been named The Sociables.
In the third segment, t he one which is pa r t icular ly
concerned wit h the fla vor and appeara nce of t he
produ ct. a large port ion of the brand decider s are
children. Their use of spearm int tooth paste is well
above average. Stri pe has done relatively well in
this segment. They a re more ego-cente red t han
ot her segments, and t heir life-st yle is outgoing bu t
not to t he extent of the swinger s. They w ill be
called T he Sensory Segment.
The fourt h segment, t he pri ce-ori ented segment,
shows a predominance of men. It t ends to be above
average in terms of toothpaste usage. P eople in
t his segment see very few mea ningful differences between brands. They switch mo re frequ ently th an

Be"e/it Segmentation: A

Dec;sjou.~orje1Jled

33

Research Tool
TABLE 1

T OOT HPASTE MAR KET SEGMENT DESCRIPTIO N

Sel1ment: Name :
Principa l bene fit sough t::
Demograph:1c st rength s :

The Sensor

Se ment:

Fl avor , product appearance

Child r en

The Socisbles

The Worriers

Brightness of teeth

Decay prevention

Teens, young people

Large families

Special behavio r al
cha r acteristics:

Users of spearmint
flavored toothpas te

Brands disproportionatel y
favored:

Co l gate , Stripe

Hacleans, Plus White ,

Ultra Brite

Personality cha racteristics :

High self-involvement

High sociabiHty

Life-style cha ra cteristics:

Hedonistic

people in other segments and tend to bu y a brand

on sale. In terms of personality, they are cognitive
and they are independent. They like to think f or
themselves a nd make brand choi ces on the basis of
their judgment. 'fhey will be called The Independent Segment .
Marketing Im plications of
Benefit Segmentation Studies
Both copy directions and medi.a choices will show
shurp differen ces depending upon which of these
segments is chosen as the ta rget-The Worri ers, The
Sociab les, The Sensory Segment, or The Independent
Segment. F or example, th e tonality of the copy
will be light if The Sociable Segment or The Sensory Segment is to be addressed. It will be more
seri ous if the copy is aimed at The Worrier s. And
if The Independent Segment is selected, it will probab ly be desirable to use rati onal, two-s ided arguments. Of cou rse, to talk to this group at all it
will be necessary to have either a pl'ice edge or
some kind of demonstrable product superi ori ty.
The depth-of-sell refl ected by the copy will a lso
vary, depending upon the segment which is of interest, It will be fairly in tensi ve for The Worri er
Segment and f or The Independent Segment, but
much more su perficial and mood-ol'iented for The
Sociable and Sensory Segments.
Likewise, t he setting will vary. It will focus on
t he produ ct for The Sensory Group, on sociallyoriented s ituations for The Sociable Grou p, and
perhaps on dem onstrati on or on competitive CODlpn!'i sons for The rndependent Group.
Medi a environments will also be tailo red to the
segments chosen as targets. Those with seri ous envii'onments will be used for The Worrier and Independent Segments, and those with youthful, modern
and active environments for The Sociable and the
Sensory Groups. F or example, it might be logical
to use a larger pJ'oportion of television for The So-

Smokers

Active

Heavy users

Crest

High hypochond r iasis

Conservative

The Independent
SeO'ment
Pr ice

Hen
Heavy users

Bran~8

on sale

H1gh

autonomy
Value-oriented

ciable and Sensory Groups, while The \oVorriers and

fndependents might have heavier print schedules.
The depth-of-sell needed will also be reflected in
the media choices. F or The Worri er and Rational
Segments longer commerCials-perhaps 60-second
commercials-would be indicated, while for the
other two groups shorter commercials and higher
freq uency wou ld be desirable.
Of co urse, in med ia selecti on the facts that have
been gathered about the demographic characteristics of the segment chosen as the tnt'get would also
be taken into consider ati on.
The informati on in Table 1 also has packag ing
implications. F or example, it might be appropriate
to have colorful P<~ckages for The Sensory Segment,
perhaps aqua (to indicate fluoride ) for The Worrier Grou p, and gleaming white for The Sociable
Segment because of their interest in bright white
teeth .
It should be readily apparent that the kinds of
informati on normally obtained in t he cou rse of a
benefit segmentation study ha ve a wide range of
marketing implications. Sometimes they are useful in suggesting phYSical changes in a produ ct.
F or example, one manufacturer discovered that his
product was weLl su ited to the needs of his chosen
target with a single exception in the area of fl avor.
He was able to make a relatively inexpens ive modificati on in his product and thereby strengthen his
market position.
The new produ ct implications of benefit segmen~
tation studies a r e equally apparent. Once a mar~
keter understands the kinds of segments that exist
in hi s ma rket, he is often able to see new pl'oduct
oPPOl'tuniti es or particularly effective ways of positioning the produ cts emergi ng f rom his research and
development oper ati on.
Similarly, benefit segmentation information has
been found helpful in providing direction in the

-,
34
choice of compatible point-oi-purchase materials and
in the selection of the kinds of sales promotions
which are most likely to be effective for any given
ma"ket target.
Genera lizations from
Benefit Segmentation St udies
A numbel' of generalizations are possible on the
basis of the major benefit segmentntion studies
which have been conducted thus far. For example,
the following general rules of thumb have become
apparent:
jg easier to ulke advantage of market segments that already exist than to attempt to
create new ones. Some time ago the strategy
of product cUlfel'entiation was heavily emphasized in marketing textbooks. Under th is philosophy it was believed that a manufacturer
was mure or less able to create new market
segments at will by mnking his product somewhat different from those of his competitors.
Now it is generally recognized that fewe)' costly elTors will be made if money is first in vested
in consumer research aimed at determ ining the
present contours of the market, Once this
knowledge is available, it is usually most efficient to tnilor marketing strategies to existing
consumer-need patterns.
No brand can expect to appeal to all consumers.
The very act of attracting one segment may
automatica lly ali enate others. A corollary to
this principle is that any marketer who wishes
to cove r a market fully must oifel' consumers
more than a single brand. The flood oJ new
b rands which have rece ntly appeared on the
market is concrete recognition of this principle.
A company's brands can sometimes cannibali ze
each other but need not necessarily do so. It
depends on whether 01' not they al'e positioned
against the same segment of the market, Ivory
Snow sharply reduced Ivory Flakes' sha re of
market, Hnd the Ford Falcon cut deeply into
the sales of the standard size Ford becnuse, in
each case, the pl'oducts wer e competing in the
same segments. Later on, for t he same companies. the Mustang was successfully introduced
with compamtively little damage to Ford; and
the success of Crest did not have a disproportionately adverse effect on Gleem's market
position because, in these CHses, the segments to
which the products appea led were different.
New and old products alike should be designed
to fit exactly the needs of some segment of the
market. In other words, they should be aimed
at people seeki ng a speCific combination of benefits. It is a marketing truism that you sell
people one at a time-that you have to get
8,fHneOne to buy your product before you get

It

loltr",,[ of MarketillK, llt[y, 1968

anyone to buy it,

A substantial group of people must be interested in your specific set of

benefits before yOll can make progress in a
market. Yet, many products attempt to aim
at two 01' more segments Simultaneously. As
a result, they are not able to maximize their
appeal to any segment of the mal'ket. and t hey
run the risk of ending up with a dangeroLisly
fuzzy brand image.
Marketers who adopt a benefit segmentation
strategy have a distinct competiti ve edge. If
a benefit segment can be located wh ich i5 seeking exactly the kinds of satisfactions that one
marketer's brand can offer better than any
other brand, the rnarkete l" can almost certainly
dominate the purchases of that segment. Furthe rmore, if his competitors are looking nt the
market in terms of tl'aditiollHI types of segments. they may not even be aware of the existence of the benefit segment which he has chosen
as his market target. If they are ignorant in
this sense, they will be at a loss to expla in the
success of his brand. And it naturally follows
that if they do not understand the reasons fa,'
his Sllccess, the kinds of people buying his
brand. and the be nefits they are obtaining from
it, his competitors wi ll find it very diflicult to
successfully attack the marketer's position.
An understanding of the benefit segments which
exist within a market can be used to advantage
when competitors introduce new products. Once
the way in wh ich consumers <lre pOSitioning
t he new product has been determined, the likelihood that it will make major inroads into
segments of interest can be assessed, and a decision ca n be made 011 whether 01' not cou nteractions of any kind are required. 1f t he new
product appears to be assuming an amb iguous
pOSition, no money need be invested in defensive
measures. However, if it appeal's that the new
product is ideally suited to the needs of an
important segment of the market. the manuf~\c
ture l' in question can introduce a new competitive product of his own, modify the physical
properties of existing brands, change his advertising strategy, or take whatever steps appear
appropriate.

Types of Seg ments Uuco1lered Through

Benefit Seg metJlatiOtI S tudies
It is diflieult to generalize about the types of
segments ,\Chich are apt to be discovered in the
co urse of a benefit segmentation study. To a large
extent, the segments which have been found have
been unique to the product categories being analyzed. However, a few types of segments have
appeared in two 0 1' more private studies. Among
them are the following:

Benefit Segm entation: A Decision-oriented Research Tool

The Status Seeker

The Swinger

The Conservative

The Rational Man

a group which is ve ry
much concerned with the
prestige of the brands
purchased.
a group whi ch tries to
be modern and up to
date in all of its activities. Brand choices reflect thi s orientation.
. a group which prefers to
stick to large successful
companies and popular
brands.
. a group which looks for
benefits such as economy, value, durability,

etc.

The Innerdirected Man

The Hedoni st

a group which is especially concerned with

self-concept. M ern bel's
consider themselves to
have a sense of humor,
to be independent and /
or h onest.
a group which is concerned primarily with
sensory benefits.

Some of these segments appear among the customers of almost all products and servi ces. However, there is no guarantee that a majority of them

35

or, for that matter, any of them exist in any given

product category. Finding out whether they do and,
if so, what should be done about them is the purpose of benefit segmentation research.
Conclusion
The benefit segmentation approach is of particular
interest because it never fai Is to provide fresh insight into markets. As was indicated in the toothpaste example cited earUer, the marketing implications of this analytical research tool are Limi ted
only by the imagination of the person using the
informati on a segmentation study provides. In
effect, when segmentation studies are cond ucted, a
number of smaller markets emerge instead of one
large one. Moreover, each of these smaller markets
can be subjected to the same kinds of thorough
analyses to which total markets have been s ubj ected
in the past. The only difference--a cr uc ial one-is
that the total market was a heterogeneous conglomeration of sub-groups. The so-called average consumer existed only in the minds of some marketing
people. When benefit segmentation is used, a number of relatively homogeneous segments are uncovered. And. because they are homogeneous, descriptio ns of them in terms of averages are much more
appropriate and meaningful as marketing gu ides.

-------------MARKETING MEMO------------A difference between managers and scientists

A basic controversy revo lves around th e question of problem vs. technique orienta
Lion. In the extreme, managers are interested in problem solving regal'dl eS5 of technique, and scientjsls are interested in sop histicat ion of method regardless of app lica
bility. Yet the quest ion of pure vs. app lied research is really one of degree rather
than kind, and th ese desi gnati ons are rea ll y related to long- and shor t-term consi_d era ti ons. However, overemphasis on pure resea.rch and technique developmenl
ralher than problem solving has resulted in low yield from investm en ls in mpnagemenl science. The percentage of recommendations that eventually affect operations
has been mise rably low.
- James E. Rosenzweig, " Managers and
Management Scientists (Two Cui
hIres) ," Business Horizons, Vol. 10
(fall, 1967 ), pp. 79.86, at pages 79,
and 80.