Sei sulla pagina 1di 5

an inference about an unidentifiable attitude and still find the cus-

tomer. For example, Internet usage could be a proxy, or owner-


ship of a PDA, or both taken together. Both of those variables
would be more actionable because direct information is available
for sale from ISPs or direct marketing companies.
In developing such proxy variables, two critical developments
take place. First, the segmentation takes a step toward being practi-
cal, or being actionable. Second, the team assembled begins to build
a consensus around the market; they share their individual experi-
ences and perspectives, and in doing so not only surface tacit
knowledge that has never been systematically captured, but also
build bridges between different parts of the organization.
Create a frame. Once the team has identified variables that
score high on both meaningfulness and actionability, it per-
of purchase drivers––again, a somewhat “obvious” segmentation
variable that turns out to be richer and more predictive in the con-
text of an array of meaningful and actionable variables.
Collapse the cells. In most cases, as in this example, the first
frame will generally be treated as a strong “working hypothesis”
and then tested with all existing data. This is an iterative process
shared by the team that relies on hard customer data. With each
step, the analysis goes into increasing detail. Preliminary quanti-
tative work tests the effectiveness of the frame variables to arrive
at a final frame. Then, key behavioral data within each “cell”
defined by the frame (e.g., brand choice, channel choice, volume
consumed, price paid) are compared to determine which “cells”
will collapse into segments. For most markets it will be impracti-
cal to manage offers to more than eight or 10 segments, and in
EXHIBIT 2
Portrait of a “labor intensive” consumer in a manufacturing job

Purchase and
Usage Environment

• Predominantly men working in suburban


or rural settings
• Find work physically demanding and
repetitive
–They are usually standing or moving
around while on their feet
• Work environment likely to be unpleasant
– Least likely to work in an environment
with heating or air conditioning
• Break room without kitchen is the primary
facility where they can relax, socialize, and
consume snacks
Desired
Experience

• Want a snack that tastes good during both


meal and non-meal occasions
• More likely during non-meal occasions to
want a snack that provides energy
• More likely during non-meal occasions to
use a snack that to help them cope with their
work environment
• Want a snack that is fun
Product/Service
Beliefs and
Associations
• More likely than other segments to believe
that snacks satisfy physical needs (taste and
refreshment) rather than emotional needs
(personal reward, escape)
• More likely than other segments to enjoy the
taste of chips and pretzels
• More likely than other segments to believe that
"healthy" snacks will improve their work
performance
• Only segment to prefer competitor product over
client product
M
MARKETING MANAGEMENT
Shaping the Profession of Marketing • January/February 2002

forms quantitative analysis to identify which short list of vari- practice, you’re likely to find that cells are relatively easy to com- • The most commonly available snacks are
ables is likely to best predict customer behavior. This refined pare and collapse around real customer behaviors. chips, pretzels, and sweets
• Although they purchase most snacks at Purchase and
short list of variables is then put into a segmentation “frame,” Profile the segments. Finally, all available data from the cells work, they are more likely than any other
Usage Behavior
segment to bring snacks from home
as shown in Exhibit 1. that comprise the segments can be placed into an integrating • Least likely to consume snacks outside of
In this (disguised) example, taken from work with a client in framework. This data can and should come from multiple scheduled breaks or mealtimes • Most likely to use vending machines • Medium bag is the package of choice
• Have to walk the farthest of all segments as their source of snacks • Most likely segment to use a single serving
the food and beverage sector, the team was able to identify a sources, ranging from existing quantitative and qualitative cus- to get their snack source • Chips and pretzels are their top choice of bag during meal times
number of proxy variables that were both actionable and power- tomer research to the experiences and latent knowledge of the snack for both meal and non-meal occasions • More likely to consume a snack in social settings
• Client brands consumed more often during than other segments
fully meaningful. The original segmentation, which was not pro- team and the broader organization. The effort may––or may non-meal occasions, but at the same rate as
ducing strong results in terms of penetration or growth, was built not––require new primary research. This is obviously an empiri- competitor brands during meal occasions
around very broad demographic categories, such as “teenagers” cally driven process, but one that incorporates significant
or “blue collar workers.” amounts of judgment along the way. The fact that a team of peo-
Consistent with the steps outlined above, this team began by ple makes those judgments helps ensure the results will be both that overcomes the three classic failures. The segments you’ve you to find and target customers, you’ll find it’s much easier to
defining the scope of the segmentation around a marketing objec- rich and practical. identified will exhibit significantly distinctive behavior. You’ll be align the entire organization as it goes to market and to grow
tive—selling more snacks in the setting of the workplace. Once Our preferred “integrating framework” is the Customer able to identify the customers who lie within them. And crucial- your top-line revenues and bottom-line earnings.
that refined scope was established, the team brainstormed a long Portrait®, as shown in Exhibit 2. Rooted in cognitive psychology, ly, your organization will be well down the path toward agree- Like many clichés, this one turns out to be true––successful
list of variables, which were scored and then tested. One of the it allows us to integrate a variety of data describing customer ment about who your top-priority customers are, how to find marketing does begin with effective customer segmentation. The
more interesting results here is how powerful relatively simple behavior in a way that allows us to more effectively interpret and them, and what to do with them and for them. This approach trick is simply using an approach to segmentation that works
demographic variables turned out to be. On the left axis age and ultimately influence it. In this way, the Customer Portrait® acts produces more than segmentation, it produces a shared view of both analytically, and also practically, with your customers and
gender were used as proxies for more meaningful attitudinal vari- as bridge to the development of segment-specific marketing your market, a shared language, and a shared set of goals. If this across your organization. ■
ables around health and snacks. However, demographics alone
don’t tell the whole story. The team knew there would be a num-
ber of constraints and drivers around the choice of snack, the
accessibility of it, and the reason or occasion for buying it. After
brainstorming and quantitative testing, we determined that the
plans. The segmentation approach enables in-depth segment pro-
filing, like this, and the creation of targeted marketing plans. You
can both understand the segments and find them in the market-
place precisely because you’ve designed in the capability to easi-
ly identify them from the beginning of the process.
seems like common sense, it probably ought to. All customers
are people. Effective marketing is about finding them and under-
standing their behavior. A segmentation that works shouldn’t be
as hard, or as easy, as a complicated statistical model. Your
organization probably already has all the data and experience it
About the Authors
Jennifer Barron is a partner at Monitor Group based in Cambridge
and a co-founder of Market2Customer, a Monitor business unit that
focuses on growth-accelerating marketing strategies. She may be
Marketing in
nature of the occupation could act as a strong proxy for a number

EXHIBIT 1
Segmentation frame for selling “salty snacks” in the workplace
In the end you will have created a customer segmentation needs to build an effective customer segmentation. All that’s
required is a systematic approach to capturing and systematizing
that knowledge.
The point of all of this, of course, is more than just creating
analytical insight into customers; it’s to generate results for your
reached at jennifer_barron@monitor.com.

Jim Hollingshead is a partner at Monitor Group, London. He has 10


years of experience in marketing and strategy, having worked with
clients in consumer goods, telecommunications, and several other
Changing Times
Labor Standing and Sitting and Stationed in On-the go Up and about business. After developing a segmentation that actually allows industries. He may be reached at jim_hollingshead@monitor.com.
intensive interacting interacting office (e.g., (e.g., nurse, professional
(e.g., factory (e.g., retail (e.g., operator, secretary, securities (e.g., executive,
Age Gender worker) assistant) telemarketer) software engineer) trader ) consulting)

Male
18-25
Female Nature of occupation is a proxy for:

Age and gender are


26-35
Male
• Nature of formal work breaks and ability
to eat snacks on the job
• Type of meal consumed during the work day
Does Online Privacy Really Matter?
Female
proxies for: • Physical/emotional need for certain types of snacks
• Attitudes about
health and nutrition
• Role of the snack
36+
Male
• Availability of snacks onsite, likelihood of "home"
provisioning, availability of alternative sources SEGMENTATION THAT WORKS
Female
Reprints with permission from the American Marketing Association

Reprinted with permission from MARKETING MANAGEMENT™, January/February 2002. 220610


Legal obstacles to e-commerce
Copyright 2002 by American Marketing Association. All rights reserved. For reorders, call 651-582-3800.
MakIng SegmentatIon Work EXECUTIVE b r i e f i n g
In today’s marketing world, segmentation is often treated as old hat. And it’s no wonder. With so many misconceptions and seg-
mentation failures, marketing professionals are left with little more than survey statistics, textbook rhetoric, and wasted market
Successful marketing really does begin with effective segmentation. research dollars. But all hope is not lost. Building segmentation that works requires a new approach to an old concept. And doing
it well can revolutionize a market and create explosive top-line growth.

T oday it’s a cliché to suggest that successful marketing begins


with good customer segmentation. In fact, ever since the late
’60s marketers have sought to first divide consumers into dis-
tinct groups and then tailor products, service offerings, brands,
keting mix, or their ad spend, or the training of their sales force.
And they’ve missed real market opportunities in the quest for
that simplicity.
Failure No. 2: We can identify “clusters,” but can’t find the
of the same coin of Failure No. 1. While this seems counterintu-
itive to most experienced market researchers, we have proven
and re-proven this finding hundreds of times, even in our own
work. During the late ’80s and early ’90s, we developed extreme-
brand management, and sometimes channel partners or general
management. By including people with different perspectives,
the process will not only incorporate a richer and more varied set
of knowledge, but it will also ease the process of buy-in and com-
and channel strategies to suit. Beginning in the world of pack- customers. The analytical antidote to the first failure––or so many ly sophisticated tools for conducting this type of needs-based munication across the company.
aged consumer goods and spreading across virtually all indus- academics and consultants would claim––is finding a sophisticat- segmentation. Again and again, despite a continual increase in Define an appropriate scope. Launch the analytical task by
tries, customer segmentation is now broadly practiced by mar- ed mathematical solution to the demographic problem. This the statistical and mathematical sophistication of our analytical asking practical questions about what you’re trying to accom-
keters almost everywhere. And countless approaches to develop- approach recognizes that simple traditional segmentation tools, we ran into the same brick wall––needs-based clusters plish in the market, not in the analysis. This means getting the
ing “the right” segmentation, from needs-based cluster analysis approaches have worn thin, and it proposes a solution that seems almost never identify groups of customers you can find in the scope of the exercise right. First, segmentations should explain
to database mining to the Internet-driven myth of the “segment to be simply common sense: Ask customers what they want, and real world. Where do “the free spirits” shop? Where do they live? and predict customer behavior in the market because this is
of one” are springing up almost everywhere. The trouble is these then target their needs. In practice, this usually means a large, How old are they? How much money do they make? What what you’re trying to manage. You want to know why cus-
approaches rarely work. single customer survey that presents a sample of prospective cus- media do they read? These are the simple questions that would tomers buy what they buy so you can sell them your products
tomers with a battery of questions about what they want and lead to marketing actions, and cluster analysis rarely if ever or services. This point may seem almost too basic to mention,
Segmentation Pitfalls why they buy––a reasonable enough thing to do. Then the statis- answers them. but we’re continually surprised by how often we find segmen-
There are three fundamental failures associated with most tical magic begins. Usually this involves a series of linear regres- Failure No. 3: Sales, marketing, and management can’t tations that are too broad or too abstract to be useful. If, for
approaches to segmentation, and if you’re a marketing profes- sions or cluster-factor analysis to identify groups of customers agree on what the segmentation looks like. In light of the first example, you’re trying to grow your share of the mobile tele-
sional, chances are you’re familiar with all of them. with similar needs, who can then be targeted with products or two classic failures, the third is not surprising. Most marketing phone market, you want a segmentation designed to under-
Failure No. 1: We can target the customers, but they don’t services designed to meet those needs. These are the segmenta- organizations have war stories about the giant segmentation stand customers’ choices about mobile telephone service, not a
buy different things. This is traditionally the classic, and still the tions that result in segments with names like “the free spirits” study that went nowhere or “the guys in market research” doing segmentation built to understand “lifestyles” or “attitudes
most frequent, reason segmentations fail. Most often this hap- and “the outgoing socializers,” which are intended to be memo- yet another segmentation. Often sales, marketing, and market toward technology.” In many cases, you’ll want to be even more
pens when companies try to segment the markets they’re in rably indicative of their needs or attitudes. In theory, this is an research can’t agree on how to segment their markets, which specific. For example, in consumer goods the relevant customer
according to simple demographic (or in B2B markets, “firmo- eminently rational approach. In practice, it almost never works. leads to untold amounts of waste, not only in additional but yet- behavior may be brand choice (Coke vs. Pepsi), but it may alter-
graphic”) data. This approach is appealing because it makes the The reason the needs-based cluster approach fails is simple: to-be-unused market research, but in poorly targeted product natively be a prior category choice (soft drink vs. water), so
customer groups you identify relatively easy to target. In practice, “needs” almost never correlate neatly to demograph- development, poorly conceived channel strategies, wide-of-the- your segmentation should explain that. In any case, you need to
Segmentation built around gender and income can easily gener- ics or other metrics that would and theoretically should help you mark ad campaigns, and organizational friction. Often this fric- scope the effort around a practical set of customer behaviors
ate a direct mail list, for example. find and target customers, tion becomes so pronounced that different parts of the organiza- you want to understand and influence.
The trouble with this approach is that simple no matter how many tion stop talking to each other and a general skepticism develops Brainstorm and test new variables. Once you have appro-
demographics almost never strongly correlate regressions you regarding the usefulness of customer segmentation, or even mar- priately scoped the segmentation effort, then create a short list of
to actual behaviors in the market. While this run. This is ket research broadly speaking. This is a dangerous situation segmentation variables to explain customer behavior and help
has always been the case, it has become the oppo- because the organization becomes paralyzed, unable to generate target customers. The key in this step is to find variables that are
even more apparent as markets have site side new insights into constantly changing markets. A marketing “meaningful” (they help explain and predict customer activity in
grown more competitive over the last death spiral can result as you lose share, and different parts of the marketplace) and “actionable” (they help identify the cus-
several years. Not all American the organization argue with each other rather than focus on tomers and reach them with practical and proven marketing
men age 35 to 44 making $100,000 working together to solve the problem. actions). Variables that are actionable, but not meaningful, lead
per year buy a Mercedes. In fact, to Failure No. 1; variables that are meaningful, but not action-
not all of them buy a luxury car. Steps to Success able, lead to Failure No. 2. When we work with clients we brain-
Even worse, not all of them shop Despite these classic failures, segmentation remains a vital storm variables and then systematically score them for action-
for cars at new car dealers. Not all of strategy for any successful business and is as much an art as it is ability and meaningfulness according to a set of criteria we cre-
them buy new cars. Not all of them a science. Taking certain steps can help prevent these common ate beforehand.
buy cars for the same reasons or pitfalls and put your company on the path to success. The right In many cases the variables normally regarded as the most
use their cars in the same way. team armed with clear objectives and a systematic approach can meaningful will simply not be actionable. So, to return to the fore-
And so on. Time and again we develop a segmentation that’s both insightful and practical, form- going example, “attitudes toward technology” may well be a
have seen companies use this ing a solid foundation for a marketing strategy. meaningful variable driving customer choices of mobile phone
simple demographic-centered Assemble a team. The first step in building an effective cus- service providers. But an attitude is very difficult to identify in the
approach to seg- tomer segmentation is making sure the right people are involved. market without interviewing each customer. In these cases, the
mentation to try This means putting together a team from multiple parts of the team works together to develop hypotheses around which vari-
to reduce the com- organization. The best results are produced when the working ables could “proxy” for attitudes toward technology. The idea is
plexity of their mar- By Jennifer Barron and Jim Hollingshead team includes people from marketing, sales, market research, to find a variable or combination of variables that together draw
MakIng SegmentatIon Work EXECUTIVE b r i e f i n g
In today’s marketing world, segmentation is often treated as old hat. And it’s no wonder. With so many misconceptions and seg-
mentation failures, marketing professionals are left with little more than survey statistics, textbook rhetoric, and wasted market
Successful marketing really does begin with effective segmentation. research dollars. But all hope is not lost. Building segmentation that works requires a new approach to an old concept. And doing
it well can revolutionize a market and create explosive top-line growth.

T oday it’s a cliché to suggest that successful marketing begins


with good customer segmentation. In fact, ever since the late
’60s marketers have sought to first divide consumers into dis-
tinct groups and then tailor products, service offerings, brands,
keting mix, or their ad spend, or the training of their sales force.
And they’ve missed real market opportunities in the quest for
that simplicity.
Failure No. 2: We can identify “clusters,” but can’t find the
of the same coin of Failure No. 1. While this seems counterintu-
itive to most experienced market researchers, we have proven
and re-proven this finding hundreds of times, even in our own
work. During the late ’80s and early ’90s, we developed extreme-
brand management, and sometimes channel partners or general
management. By including people with different perspectives,
the process will not only incorporate a richer and more varied set
of knowledge, but it will also ease the process of buy-in and com-
and channel strategies to suit. Beginning in the world of pack- customers. The analytical antidote to the first failure––or so many ly sophisticated tools for conducting this type of needs-based munication across the company.
aged consumer goods and spreading across virtually all indus- academics and consultants would claim––is finding a sophisticat- segmentation. Again and again, despite a continual increase in Define an appropriate scope. Launch the analytical task by
tries, customer segmentation is now broadly practiced by mar- ed mathematical solution to the demographic problem. This the statistical and mathematical sophistication of our analytical asking practical questions about what you’re trying to accom-
keters almost everywhere. And countless approaches to develop- approach recognizes that simple traditional segmentation tools, we ran into the same brick wall––needs-based clusters plish in the market, not in the analysis. This means getting the
ing “the right” segmentation, from needs-based cluster analysis approaches have worn thin, and it proposes a solution that seems almost never identify groups of customers you can find in the scope of the exercise right. First, segmentations should explain
to database mining to the Internet-driven myth of the “segment to be simply common sense: Ask customers what they want, and real world. Where do “the free spirits” shop? Where do they live? and predict customer behavior in the market because this is
of one” are springing up almost everywhere. The trouble is these then target their needs. In practice, this usually means a large, How old are they? How much money do they make? What what you’re trying to manage. You want to know why cus-
approaches rarely work. single customer survey that presents a sample of prospective cus- media do they read? These are the simple questions that would tomers buy what they buy so you can sell them your products
tomers with a battery of questions about what they want and lead to marketing actions, and cluster analysis rarely if ever or services. This point may seem almost too basic to mention,
Segmentation Pitfalls why they buy––a reasonable enough thing to do. Then the statis- answers them. but we’re continually surprised by how often we find segmen-
There are three fundamental failures associated with most tical magic begins. Usually this involves a series of linear regres- Failure No. 3: Sales, marketing, and management can’t tations that are too broad or too abstract to be useful. If, for
approaches to segmentation, and if you’re a marketing profes- sions or cluster-factor analysis to identify groups of customers agree on what the segmentation looks like. In light of the first example, you’re trying to grow your share of the mobile tele-
sional, chances are you’re familiar with all of them. with similar needs, who can then be targeted with products or two classic failures, the third is not surprising. Most marketing phone market, you want a segmentation designed to under-
Failure No. 1: We can target the customers, but they don’t services designed to meet those needs. These are the segmenta- organizations have war stories about the giant segmentation stand customers’ choices about mobile telephone service, not a
buy different things. This is traditionally the classic, and still the tions that result in segments with names like “the free spirits” study that went nowhere or “the guys in market research” doing segmentation built to understand “lifestyles” or “attitudes
most frequent, reason segmentations fail. Most often this hap- and “the outgoing socializers,” which are intended to be memo- yet another segmentation. Often sales, marketing, and market toward technology.” In many cases, you’ll want to be even more
pens when companies try to segment the markets they’re in rably indicative of their needs or attitudes. In theory, this is an research can’t agree on how to segment their markets, which specific. For example, in consumer goods the relevant customer
according to simple demographic (or in B2B markets, “firmo- eminently rational approach. In practice, it almost never works. leads to untold amounts of waste, not only in additional but yet- behavior may be brand choice (Coke vs. Pepsi), but it may alter-
graphic”) data. This approach is appealing because it makes the The reason the needs-based cluster approach fails is simple: to-be-unused market research, but in poorly targeted product natively be a prior category choice (soft drink vs. water), so
customer groups you identify relatively easy to target. In practice, “needs” almost never correlate neatly to demograph- development, poorly conceived channel strategies, wide-of-the- your segmentation should explain that. In any case, you need to
Segmentation built around gender and income can easily gener- ics or other metrics that would and theoretically should help you mark ad campaigns, and organizational friction. Often this fric- scope the effort around a practical set of customer behaviors
ate a direct mail list, for example. find and target customers, tion becomes so pronounced that different parts of the organiza- you want to understand and influence.
The trouble with this approach is that simple no matter how many tion stop talking to each other and a general skepticism develops Brainstorm and test new variables. Once you have appro-
demographics almost never strongly correlate regressions you regarding the usefulness of customer segmentation, or even mar- priately scoped the segmentation effort, then create a short list of
to actual behaviors in the market. While this run. This is ket research broadly speaking. This is a dangerous situation segmentation variables to explain customer behavior and help
has always been the case, it has become the oppo- because the organization becomes paralyzed, unable to generate target customers. The key in this step is to find variables that are
even more apparent as markets have site side new insights into constantly changing markets. A marketing “meaningful” (they help explain and predict customer activity in
grown more competitive over the last death spiral can result as you lose share, and different parts of the marketplace) and “actionable” (they help identify the cus-
several years. Not all American the organization argue with each other rather than focus on tomers and reach them with practical and proven marketing
men age 35 to 44 making $100,000 working together to solve the problem. actions). Variables that are actionable, but not meaningful, lead
per year buy a Mercedes. In fact, to Failure No. 1; variables that are meaningful, but not action-
not all of them buy a luxury car. Steps to Success able, lead to Failure No. 2. When we work with clients we brain-
Even worse, not all of them shop Despite these classic failures, segmentation remains a vital storm variables and then systematically score them for action-
for cars at new car dealers. Not all of strategy for any successful business and is as much an art as it is ability and meaningfulness according to a set of criteria we cre-
them buy new cars. Not all of them a science. Taking certain steps can help prevent these common ate beforehand.
buy cars for the same reasons or pitfalls and put your company on the path to success. The right In many cases the variables normally regarded as the most
use their cars in the same way. team armed with clear objectives and a systematic approach can meaningful will simply not be actionable. So, to return to the fore-
And so on. Time and again we develop a segmentation that’s both insightful and practical, form- going example, “attitudes toward technology” may well be a
have seen companies use this ing a solid foundation for a marketing strategy. meaningful variable driving customer choices of mobile phone
simple demographic-centered Assemble a team. The first step in building an effective cus- service providers. But an attitude is very difficult to identify in the
approach to seg- tomer segmentation is making sure the right people are involved. market without interviewing each customer. In these cases, the
mentation to try This means putting together a team from multiple parts of the team works together to develop hypotheses around which vari-
to reduce the com- organization. The best results are produced when the working ables could “proxy” for attitudes toward technology. The idea is
plexity of their mar- By Jennifer Barron and Jim Hollingshead team includes people from marketing, sales, market research, to find a variable or combination of variables that together draw
an inference about an unidentifiable attitude and still find the cus-
tomer. For example, Internet usage could be a proxy, or owner-
ship of a PDA, or both taken together. Both of those variables
would be more actionable because direct information is available
for sale from ISPs or direct marketing companies.
In developing such proxy variables, two critical developments
take place. First, the segmentation takes a step toward being practi-
cal, or being actionable. Second, the team assembled begins to build
a consensus around the market; they share their individual experi-
ences and perspectives, and in doing so not only surface tacit
knowledge that has never been systematically captured, but also
build bridges between different parts of the organization.
Create a frame. Once the team has identified variables that
score high on both meaningfulness and actionability, it per-
of purchase drivers––again, a somewhat “obvious” segmentation
variable that turns out to be richer and more predictive in the con-
text of an array of meaningful and actionable variables.
Collapse the cells. In most cases, as in this example, the first
frame will generally be treated as a strong “working hypothesis”
and then tested with all existing data. This is an iterative process
shared by the team that relies on hard customer data. With each
step, the analysis goes into increasing detail. Preliminary quanti-
tative work tests the effectiveness of the frame variables to arrive
at a final frame. Then, key behavioral data within each “cell”
defined by the frame (e.g., brand choice, channel choice, volume
consumed, price paid) are compared to determine which “cells”
will collapse into segments. For most markets it will be impracti-
cal to manage offers to more than eight or 10 segments, and in
EXHIBIT 2
Portrait of a “labor intensive” consumer in a manufacturing job

Purchase and
Usage Environment

• Predominantly men working in suburban


or rural settings
• Find work physically demanding and
repetitive
–They are usually standing or moving
around while on their feet
• Work environment likely to be unpleasant
– Least likely to work in an environment
with heating or air conditioning
• Break room without kitchen is the primary
facility where they can relax, socialize, and
consume snacks
Desired
Experience

• Want a snack that tastes good during both


meal and non-meal occasions
• More likely during non-meal occasions to
want a snack that provides energy
• More likely during non-meal occasions to
use a snack that to help them cope with their
work environment
• Want a snack that is fun
Product/Service
Beliefs and
Associations
• More likely than other segments to believe
that snacks satisfy physical needs (taste and
refreshment) rather than emotional needs
(personal reward, escape)
• More likely than other segments to enjoy the
taste of chips and pretzels
• More likely than other segments to believe that
"healthy" snacks will improve their work
performance
• Only segment to prefer competitor product over
client product
M
MARKETING MANAGEMENT
Shaping the Profession of Marketing • January/February 2002

forms quantitative analysis to identify which short list of vari- practice, you’re likely to find that cells are relatively easy to com- • The most commonly available snacks are
ables is likely to best predict customer behavior. This refined pare and collapse around real customer behaviors. chips, pretzels, and sweets
• Although they purchase most snacks at Purchase and
short list of variables is then put into a segmentation “frame,” Profile the segments. Finally, all available data from the cells work, they are more likely than any other
Usage Behavior
segment to bring snacks from home
as shown in Exhibit 1. that comprise the segments can be placed into an integrating • Least likely to consume snacks outside of
In this (disguised) example, taken from work with a client in framework. This data can and should come from multiple scheduled breaks or mealtimes • Most likely to use vending machines • Medium bag is the package of choice
• Have to walk the farthest of all segments as their source of snacks • Most likely segment to use a single serving
the food and beverage sector, the team was able to identify a sources, ranging from existing quantitative and qualitative cus- to get their snack source • Chips and pretzels are their top choice of bag during meal times
number of proxy variables that were both actionable and power- tomer research to the experiences and latent knowledge of the snack for both meal and non-meal occasions • More likely to consume a snack in social settings
• Client brands consumed more often during than other segments
fully meaningful. The original segmentation, which was not pro- team and the broader organization. The effort may––or may non-meal occasions, but at the same rate as
ducing strong results in terms of penetration or growth, was built not––require new primary research. This is obviously an empiri- competitor brands during meal occasions
around very broad demographic categories, such as “teenagers” cally driven process, but one that incorporates significant
or “blue collar workers.” amounts of judgment along the way. The fact that a team of peo-
Consistent with the steps outlined above, this team began by ple makes those judgments helps ensure the results will be both that overcomes the three classic failures. The segments you’ve you to find and target customers, you’ll find it’s much easier to
defining the scope of the segmentation around a marketing objec- rich and practical. identified will exhibit significantly distinctive behavior. You’ll be align the entire organization as it goes to market and to grow
tive—selling more snacks in the setting of the workplace. Once Our preferred “integrating framework” is the Customer able to identify the customers who lie within them. And crucial- your top-line revenues and bottom-line earnings.
that refined scope was established, the team brainstormed a long Portrait®, as shown in Exhibit 2. Rooted in cognitive psychology, ly, your organization will be well down the path toward agree- Like many clichés, this one turns out to be true––successful
list of variables, which were scored and then tested. One of the it allows us to integrate a variety of data describing customer ment about who your top-priority customers are, how to find marketing does begin with effective customer segmentation. The
more interesting results here is how powerful relatively simple behavior in a way that allows us to more effectively interpret and them, and what to do with them and for them. This approach trick is simply using an approach to segmentation that works
demographic variables turned out to be. On the left axis age and ultimately influence it. In this way, the Customer Portrait® acts produces more than segmentation, it produces a shared view of both analytically, and also practically, with your customers and
gender were used as proxies for more meaningful attitudinal vari- as bridge to the development of segment-specific marketing your market, a shared language, and a shared set of goals. If this across your organization. ■
ables around health and snacks. However, demographics alone
don’t tell the whole story. The team knew there would be a num-
ber of constraints and drivers around the choice of snack, the
accessibility of it, and the reason or occasion for buying it. After
brainstorming and quantitative testing, we determined that the
plans. The segmentation approach enables in-depth segment pro-
filing, like this, and the creation of targeted marketing plans. You
can both understand the segments and find them in the market-
place precisely because you’ve designed in the capability to easi-
ly identify them from the beginning of the process.
seems like common sense, it probably ought to. All customers
are people. Effective marketing is about finding them and under-
standing their behavior. A segmentation that works shouldn’t be
as hard, or as easy, as a complicated statistical model. Your
organization probably already has all the data and experience it
About the Authors
Jennifer Barron is a partner at Monitor Group based in Cambridge
and a co-founder of Market2Customer, a Monitor business unit that
focuses on growth-accelerating marketing strategies. She may be
Marketing in
nature of the occupation could act as a strong proxy for a number

EXHIBIT 1
Segmentation frame for selling “salty snacks” in the workplace
In the end you will have created a customer segmentation needs to build an effective customer segmentation. All that’s
required is a systematic approach to capturing and systematizing
that knowledge.
The point of all of this, of course, is more than just creating
analytical insight into customers; it’s to generate results for your
reached at jennifer_barron@monitor.com.

Jim Hollingshead is a partner at Monitor Group, London. He has 10


years of experience in marketing and strategy, having worked with
clients in consumer goods, telecommunications, and several other
Changing Times
Labor Standing and Sitting and Stationed in On-the go Up and about business. After developing a segmentation that actually allows industries. He may be reached at jim_hollingshead@monitor.com.
intensive interacting interacting office (e.g., (e.g., nurse, professional
(e.g., factory (e.g., retail (e.g., operator, secretary, securities (e.g., executive,
Age Gender worker) assistant) telemarketer) software engineer) trader ) consulting)

Male
18-25
Female Nature of occupation is a proxy for:

Age and gender are


26-35
Male
• Nature of formal work breaks and ability
to eat snacks on the job
• Type of meal consumed during the work day
Does Online Privacy Really Matter?
Female
proxies for: • Physical/emotional need for certain types of snacks
• Attitudes about
health and nutrition
• Role of the snack
36+
Male
• Availability of snacks onsite, likelihood of "home"
provisioning, availability of alternative sources SEGMENTATION THAT WORKS
Female
Reprints with permission from the American Marketing Association

Reprinted with permission from MARKETING MANAGEMENT™, January/February 2002. 220610


Legal obstacles to e-commerce
Copyright 2002 by American Marketing Association. All rights reserved. For reorders, call 651-582-3800.
an inference about an unidentifiable attitude and still find the cus-
tomer. For example, Internet usage could be a proxy, or owner-
ship of a PDA, or both taken together. Both of those variables
would be more actionable because direct information is available
for sale from ISPs or direct marketing companies.
In developing such proxy variables, two critical developments
take place. First, the segmentation takes a step toward being practi-
cal, or being actionable. Second, the team assembled begins to build
a consensus around the market; they share their individual experi-
ences and perspectives, and in doing so not only surface tacit
knowledge that has never been systematically captured, but also
build bridges between different parts of the organization.
Create a frame. Once the team has identified variables that
score high on both meaningfulness and actionability, it per-
of purchase drivers––again, a somewhat “obvious” segmentation
variable that turns out to be richer and more predictive in the con-
text of an array of meaningful and actionable variables.
Collapse the cells. In most cases, as in this example, the first
frame will generally be treated as a strong “working hypothesis”
and then tested with all existing data. This is an iterative process
shared by the team that relies on hard customer data. With each
step, the analysis goes into increasing detail. Preliminary quanti-
tative work tests the effectiveness of the frame variables to arrive
at a final frame. Then, key behavioral data within each “cell”
defined by the frame (e.g., brand choice, channel choice, volume
consumed, price paid) are compared to determine which “cells”
will collapse into segments. For most markets it will be impracti-
cal to manage offers to more than eight or 10 segments, and in
EXHIBIT 2
Portrait of a “labor intensive” consumer in a manufacturing job

Purchase and
Usage Environment

• Predominantly men working in suburban


or rural settings
• Find work physically demanding and
repetitive
–They are usually standing or moving
around while on their feet
• Work environment likely to be unpleasant
– Least likely to work in an environment
with heating or air conditioning
• Break room without kitchen is the primary
facility where they can relax, socialize, and
consume snacks
Desired
Experience

• Want a snack that tastes good during both


meal and non-meal occasions
• More likely during non-meal occasions to
want a snack that provides energy
• More likely during non-meal occasions to
use a snack that to help them cope with their
work environment
• Want a snack that is fun
Product/Service
Beliefs and
Associations
• More likely than other segments to believe
that snacks satisfy physical needs (taste and
refreshment) rather than emotional needs
(personal reward, escape)
• More likely than other segments to enjoy the
taste of chips and pretzels
• More likely than other segments to believe that
"healthy" snacks will improve their work
performance
• Only segment to prefer competitor product over
client product
M
MARKETING MANAGEMENT
Shaping the Profession of Marketing • January/February 2002

forms quantitative analysis to identify which short list of vari- practice, you’re likely to find that cells are relatively easy to com- • The most commonly available snacks are
ables is likely to best predict customer behavior. This refined pare and collapse around real customer behaviors. chips, pretzels, and sweets
• Although they purchase most snacks at Purchase and
short list of variables is then put into a segmentation “frame,” Profile the segments. Finally, all available data from the cells work, they are more likely than any other
Usage Behavior
segment to bring snacks from home
as shown in Exhibit 1. that comprise the segments can be placed into an integrating • Least likely to consume snacks outside of
In this (disguised) example, taken from work with a client in framework. This data can and should come from multiple scheduled breaks or mealtimes • Most likely to use vending machines • Medium bag is the package of choice
• Have to walk the farthest of all segments as their source of snacks • Most likely segment to use a single serving
the food and beverage sector, the team was able to identify a sources, ranging from existing quantitative and qualitative cus- to get their snack source • Chips and pretzels are their top choice of bag during meal times
number of proxy variables that were both actionable and power- tomer research to the experiences and latent knowledge of the snack for both meal and non-meal occasions • More likely to consume a snack in social settings
• Client brands consumed more often during than other segments
fully meaningful. The original segmentation, which was not pro- team and the broader organization. The effort may––or may non-meal occasions, but at the same rate as
ducing strong results in terms of penetration or growth, was built not––require new primary research. This is obviously an empiri- competitor brands during meal occasions
around very broad demographic categories, such as “teenagers” cally driven process, but one that incorporates significant
or “blue collar workers.” amounts of judgment along the way. The fact that a team of peo-
Consistent with the steps outlined above, this team began by ple makes those judgments helps ensure the results will be both that overcomes the three classic failures. The segments you’ve you to find and target customers, you’ll find it’s much easier to
defining the scope of the segmentation around a marketing objec- rich and practical. identified will exhibit significantly distinctive behavior. You’ll be align the entire organization as it goes to market and to grow
tive—selling more snacks in the setting of the workplace. Once Our preferred “integrating framework” is the Customer able to identify the customers who lie within them. And crucial- your top-line revenues and bottom-line earnings.
that refined scope was established, the team brainstormed a long Portrait®, as shown in Exhibit 2. Rooted in cognitive psychology, ly, your organization will be well down the path toward agree- Like many clichés, this one turns out to be true––successful
list of variables, which were scored and then tested. One of the it allows us to integrate a variety of data describing customer ment about who your top-priority customers are, how to find marketing does begin with effective customer segmentation. The
more interesting results here is how powerful relatively simple behavior in a way that allows us to more effectively interpret and them, and what to do with them and for them. This approach trick is simply using an approach to segmentation that works
demographic variables turned out to be. On the left axis age and ultimately influence it. In this way, the Customer Portrait® acts produces more than segmentation, it produces a shared view of both analytically, and also practically, with your customers and
gender were used as proxies for more meaningful attitudinal vari- as bridge to the development of segment-specific marketing your market, a shared language, and a shared set of goals. If this across your organization. ■
ables around health and snacks. However, demographics alone
don’t tell the whole story. The team knew there would be a num-
ber of constraints and drivers around the choice of snack, the
accessibility of it, and the reason or occasion for buying it. After
brainstorming and quantitative testing, we determined that the
plans. The segmentation approach enables in-depth segment pro-
filing, like this, and the creation of targeted marketing plans. You
can both understand the segments and find them in the market-
place precisely because you’ve designed in the capability to easi-
ly identify them from the beginning of the process.
seems like common sense, it probably ought to. All customers
are people. Effective marketing is about finding them and under-
standing their behavior. A segmentation that works shouldn’t be
as hard, or as easy, as a complicated statistical model. Your
organization probably already has all the data and experience it
About the Authors
Jennifer Barron is a partner at Monitor Group based in Cambridge
and a co-founder of Market2Customer, a Monitor business unit that
focuses on growth-accelerating marketing strategies. She may be
Marketing in
nature of the occupation could act as a strong proxy for a number

EXHIBIT 1
Segmentation frame for selling “salty snacks” in the workplace
In the end you will have created a customer segmentation needs to build an effective customer segmentation. All that’s
required is a systematic approach to capturing and systematizing
that knowledge.
The point of all of this, of course, is more than just creating
analytical insight into customers; it’s to generate results for your
reached at jennifer_barron@monitor.com.

Jim Hollingshead is a partner at Monitor Group, London. He has 10


years of experience in marketing and strategy, having worked with
clients in consumer goods, telecommunications, and several other
Changing Times
Labor Standing and Sitting and Stationed in On-the go Up and about business. After developing a segmentation that actually allows industries. He may be reached at jim_hollingshead@monitor.com.
intensive interacting interacting office (e.g., (e.g., nurse, professional
(e.g., factory (e.g., retail (e.g., operator, secretary, securities (e.g., executive,
Age Gender worker) assistant) telemarketer) software engineer) trader ) consulting)

Male
18-25
Female Nature of occupation is a proxy for:

Age and gender are


26-35
Male
• Nature of formal work breaks and ability
to eat snacks on the job
• Type of meal consumed during the work day
Does Online Privacy Really Matter?
Female
proxies for: • Physical/emotional need for certain types of snacks
• Attitudes about
health and nutrition
• Role of the snack
36+
Male
• Availability of snacks onsite, likelihood of "home"
provisioning, availability of alternative sources SEGMENTATION THAT WORKS
Female
Reprints with permission from the American Marketing Association

Reprinted with permission from MARKETING MANAGEMENT™, January/February 2002. 220610


Legal obstacles to e-commerce
Copyright 2002 by American Marketing Association. All rights reserved. For reorders, call 651-582-3800.

Potrebbero piacerti anche