1. Review our service design analysis framework (the Francis Frei article); how to analyze cases; and the primer on hownotto get lost in the numbers. In other words, get your analysis mindset and tool

M ARKETING

W h a t Y o u N e e d t o K n o w

A b o u t S e g m e n t a t i o n

b y G re tch e n G a ve tt

JULY 09, 2014

The m arketers of C learblue A dvanced Pregnancy Test, a product that can tell you if you’re one-w eek, tw o-w eeks,

or three-plus w eeks pregnant, asked a couple of D -list celebrities to tw eet out their positive tests back in 2013. A s

Businessweek’s Jessica G rose reported, the m aker of the test, Sw iss Precision D iagnostics, has a 25% share of the at-

hom e pregnancy-testing industry and is targeting its m arketing efforts at M illennials. G rose quotes IbisW orld

researcher Jocelyn Phillips as pointing to the high-tech aspects of C learblue’s test, also noting that young w om en

m ight be m ore w illing to shell out m ore m oney for such technology — the digital version costs about $5 m ore

than the boring old blue and pink line version.

There is nothing new about this kind of segm enting in the pregnancy test m arket, how ever. A nd it’s actually a

really useful (if not slightly unsettling) exam ple of how you m ight segm ent potential custom ers w ith very different

needs and behaviors.

For exam ple, you could segm ent the m arket for early pregnancy tests based on dem ographics such as age and

incom e, or you could segm ent the m arket based on consum ers’ price sensitivity. But in this situation, it is useful to

ask w hy: W hy w ould a w om an w ant to take a pregnancy test? A nd are these reasons the sam e for everyone? A

little bit of thought w ould suggest that there are tw o groups of w om en: hopefuls, those w ho w ant to be pregnant,

and fearfuls, those w ho are afraid that they m ight be pregnant. H ow w ould you identify these tw o segm ents and m arket to them differently? O ften com panies offer m ultiple

products that appeal to different m arket segm ents and let custom ers self-select. That is, the firm does not identify

custom ers in various m arket segm ents; instead, the custom ers reveal their m arket segm ent identity by choosing

different products. Q uidol, a com pany based in San D iego, C alifornia, created tw o different products to appeal to

tw o segm ents in the m arket for early pregnancy tests: the hopefuls and the fearfuls. The actual test products w ere

alm ost identical, but the tw o products w ere given different nam es and package designs, w ere placed in different

aisles of a drugstore, and w ere priced differently.

The Hopefuls The Fearfuls

Source: John Forsyth, Sunil Gupta, Sudeep Haldar, Anil Kaul, and Keith Kettle, "A Segmentation You Can Act On," McKinsey Quarterly,

August 1999, McKinsey & Company. Reprinted by permission.

Segm enting, at its m ost basic, is the separation of a group of custom ers w ith different needs into subgroups of

custom ers w ith sim ilar needs and preferences. By doing this, a com pany can better tailor and target its products

and services to m eet each segm ent’s needs. This isn’t, as M cK insey’s John Forsyth says, sim ply for m arketing or

retail firm s. “W e see m any, m any com panies saying, ‘I w ant to get m ore consum er-driven and custom er-facing.

But som etim es the organizations don’t know how to start. I’d say you really start w ith a basic understanding of

your consum ers or custom ers, right? A nd that’s segm entation.” It sounds straightforw ard but often it isn’t. H ere are a few pitfalls that m any com panies fall into w hen they start

thinking about segm entation. O ne, com panies rarely create a segm ent — m ore often they uncover one. Tw o,

segm entation and dem ographics are very different things. “Y ou have tw o people, w e know they’re the sam e age,

w e know they’re British citizens, and w e know they’re of royal blood,” explains Forsyth. “O ne of them is Prince

C harles. The other is O zzy O sbourne, the Prince of D arkness. They’re in the sam e dem ographic segm ent, but I

can’t im agine m arketing to them the sam e w ay.”

A nd three: you have to ask yourself w hy you w ant to segm ent and w hat decisions you’ll m ake based on the

inform ation. “M any com panies say, w ell, I think I just need a segm entation,” says Forsyth. “But before you even

start the segm entation, you need to really understand w hy you’re doing it and w hat som e of the actions are that

you’re planning to take, based on w hat you think you m ight see. It helps you understand w hat’s actionable in term s

of driving a com pany’s business.”

O nce you’ve answ ered these questions, you have to decide w hether you w ant to start segm enting by needs or

behaviors. “If you’re doing som ething strategic and you’re trying to figure out if you have the right brands, the

right value proposition, the right product line, then I w ould say you should start w ith needs or attitude

segm entation,” explains Forsyth. This is basically trying to identify w hat needs your product or service is or could

m eet.

“But if you think you’ve got that pretty m uch under control,” he continues, “and you need to understand how to

go to m arket or target your digital and TV spending, then I w ould start w ith behavior.” This involves trying to

identify differences in custom er groups based on their buying and lifestyle patterns, for exam ple.

Regardless of your approach, a useful segm entation should include these six characteristics:

1) Identifiable. Y ou should be able to identify custom ers in each segm ent and m easure their characteristics, like

dem ographics or usage behavior.

2) Substantial. It’s usually not cost-effective to target sm all segm ents — a segm ent, therefore, m ust be large

enough to be potentially profitable. 3) A ccessible. It sounds obvious, but your com pany should be able to reach its segm ents via com m unication and

distribution channels. W hen it com es to young people, for exam ple, your com pany should have access to Tw itter

and Tum blr and know how to use them authentically — or, as C learblue sm artly did, reach out to celebrities w ith

active Tw itter presences to do som e of your m arketing for you.

4) Stable. In order for a m arketing effort to be successful, a segm ent should be stable enough for a long enough

period of tim e to be m arketed to strategically. For exam ple, lifestyle is often used as a w ay to segm ent. But

research has found that, internationally, lifestyle is dynam ic and constantly evolving. Thus, segm enting based on

that variable globally m ight not be w ise.

5) D ifferentiable. The people (or organizations, in B2B m arketing) in a segm ent should have sim ilar needs that

are clearly different from the needs of other people in other segm ents.

6) A ctionable. Y ou have to be able to provide products or services to your segm ents. O ne U .S. insurance

com pany, for exam ple, spent a lot of tim e and m oney identifying a segm ent, only to discover that it couldn’t find

any custom ers for its insurance product in that segm ent, nor w as the organization able to design any actions to

target them .

N ow you can start breaking dow n segm ents by who buys, what they buy, and why they buy (or use or view , etc.) it.

The pregnancy test interactive above is a great exam ple of how this w orks.

There are also prom inent failures that com panies should heed. O ne of the m ost infam ous is w hen Bic decided to

segm ent its young fem ale consum ers. The “Bic C ristal for H er” w riting utensils w ere thinner, designed w ith m ore

pastel colors, and priced higher than other pens. W om en, in general, w ere offended, taking to A m azon to w rite

som e very creative review s. The pen m arket, in other w ords, w as not as heterogeneous along gender lines as Bic

had thought.

W hen thinking about how you segm ent, John Forsyth has several suggestions. For one, he notes, “focus groups

are dead. If you’re still using focus groups, you’re using 30-year-old technology.” A m uch better w ay to understand

custom er needs and behaviors is to spend tim e w ith people in their hom es, stores, or health clubs. “Y ou w atch

them , you talk to them w hile they’re doing the kinds of things w e w ant to be observing.” This type of qualitative research is all the m ore im portant because it show cases real stories that are key to

convincing stakeholders. “W hen w e illustrate things w ith qualitative research, w e get C E O s going, ‘W ow , you’re

really telling m e m y m arketing strategy is all w rong and I need to change it,'” says Forsyth. “It’s very pow erful, and

it’s really exploded in the last 10 years.”

Big D ata and technology have changed how com panies approach segm enting. “The old m odel, particularly in the

m arket research w orld w as, ‘I understand people’s needs and attitudes, and behaviors w ill com e from that,'”

Forsyth explains. “Today, in m any situations, [m arketers] have flipped it to say, ‘I’m going to do segm entation

based on their behaviors, and then I’m going to try to understand the needs that drive behavioral differences.”

H e w arns, how ever, that this type of segm entation is “a lot harder to do than people think, and I don’t think w e’re

anyw here near being good at it yet.”

Forsyth’s also seeing a lot of m ovem ent in the area of segm enting em erging m arkets w orldw ide, w hich poses a

num ber of challenges. For one, scales m arketers use to m easure needs or behaviors in one country m ay be w ay off

in another due to different cultural norm s.

H e also notes that affordability is still a huge factor in developing countries, too, w hereas it m ay not be elsew here

— as the $20 pack of digital pregnancy tests dem onstrates nicely.

This post includes material adapted and reprinted from Core Reading: Segmentation and Targeting, H BP. N o. 8219, by Sunil

Keller, M arketing M anagement, 14th ed. (U pper Saddle River, N J: Prentice H all, 2012).

G re tc h e n G a v e tt is an associate editor at the H arvard B usiness R eview .

Follow her on Tw itter @ gretchenm arg. R elated Topics: CUSTO M ERS | INTERNATIO NAL BUSINESS

T h is a rtic le is a b o u t M A R K E T IN G

 FO LLO W TH IS TO PIC

L o a d in g ...