them any color they want, as long as it's black" -- Henry Ford
In the early stages of an industry, when the key focus is making a reliable
product in volume and at low cost, vendors can profitably ignore customer differences. But soon enough, along
will come your own personal Sloan who uses differentiation to zap your monolithic
"sell to anyone" strategy. And those pesky customers will up and leave you.
Sloan developed a strategy that nearly destroyed Ford in the late 20's by
creating segmented product lines. Individually, GM's products wouldn't have done much damage, but collectively
they were able to under-price as well as outclass anything Ford could do.
Instead of the stable and monolithic designs from Ford, customers were attracted to
GM's segmented brands, snazzy nameplates,
and annual styling changes. Of course, 50 years later both GM and Ford
became decadent and moribund (check out Halberstam's
The Reckoning), but this marketing lesson from
the US auto industry is
Spoke at Bay Area Capital
slide deck here.
* Was syndicated in the
* Wrote on Web 2.0 and viral
Identify and execute on a small number
of segments that will really make a difference to your focus and
Don't kid yourself: segmentation is
hard. Even when you've figured out the relevant target markets to go
after, actually aligning sales, engineering, marketing, and (yes!)
professional services can be quite an exercise.
When developing segments, never say no to
a customer. If you're winning big deals and they can be serviced
profitably, make them happen no matter what segment they are in. Let
the sales guys go after whoever they can succeed with.... but provide clear
financial incentives and tools only for the segments you really want.
The reps will get the message quickly.
Be realistic about time: it can
take months of research and experimental deals to really figure out a
segmentation. Once you've invested in segment-specific sales tools,
etc., give yourself six months to measure results. It will take
even longer to get serious productivity increases, so don't make any big
decisions earlier than that.
Understand the difference between
visionary and inductive segmentation. For any one product line, you
should only use one of these techniques.
Usually, behavioral segmentation is the
most powerful. Naturally, it's the hardest to develop and verify.
Demographic segmentation is surprisingly powerful, considering its
Know that most of the time, your
competitors' segmentation is fairly simplistic and not very effective.
Segmentation for fun and profit
The first problem with understanding segmentation is the word "segment."
The word usually means more or less similar sections -- like a piece of fruit,
with fairly discernable dividing lines and patterns. But in markets, segments are
often hard to discern and measure, with boundaries that are irregular or
even invisible unless you've done some serious
Really usable market segmentation is pretty hard, so businesses
tend to go with a guess. This is incredibly wasteful, and not
just in the marketing arena. The wrong sales reps are hired, the
wrong products are built, and the company's viability may be called into
question. All because the company didn't have the right definition for the
and customer segment.
My favorite example of bad segmentation is Volkswagen, whose CEO wanted to re-brand
his company and decided to build
the Phaeton: a 12-cyllinder 5.4 liter sedan weighing nearly 3
tons and costing as much as $100,000 before taxes. Thanks to
his segment was full of contradictions, so customers didn't really
exist. The car sold only a few hundred units per year. Auf
Wiedersehen, Herr Generaldirektor.
If there's one marketing reason why the Republicans won in the 2000-4 elections, it's because they
invested millions in defining, testing, and cultivating their segments. Ask Ken Mellman,
or Frank Luntz, or even Howard Dean. Segmentation is not an idle
"marketing exercise" -- it provides leverage so you can get to and
influence the people who
will really make a difference to your results.
What kind of impact can segmentation have? For individual offers, an order
of magnitude improvement. For the overall business, you'll have to settle
for just a 3x increase in marketing and sales effectiveness. Segmentation
works because it helps you focus on the right things.
Before we get into the "how," let's do a reality check. It's just as
defocusing to have too many market segments as to not have any at all.
Unless you're a giant company, it's almost impossible to effectively execute
more than 3 segments for the whole business. If you're starting out,
seriously develop only one segment at a time, while exploring/experimenting with
one other. All too often, developing a segment is a matter of trial and
error, so you have to be able to cheaply experiment for a few months to see
segmentation works for you.
Segmentation for Fun
A segment is a group of potential customers who have similar tastes,
preferences, characteristics and
buying patterns. Segments are defined by customers and behaviors, not
by competitors. A segment can be defined along lots of different lines
(age, income, education, geography, profession, company size...). When you
have a new business with only a few customers, or you don't even have your
product yet, you can only use visionary segmentation. You use theory and external data to develop your segment definition. Nothing
wrong with this, but you'll almost certainly need to adjust your initial
segmentation after the first few months of sales. Reality trumps
hypothesis every time.
visionary segmentation, the first place to look is your competitors:
figure out what segmentation they are using, and then ask around to find out how
it's working out for them. You'll never find any real numbers on their
segments' performance (even from public companies), but if you discover that
a competitor has recently had a lot of sales reps leave, their segmentation probably
isn't worth imitating.
One reality check to do about visionary segmentation: how many competitors
are vying for customers in a segment? If there are no competitors,
you're either way early or have defined a non-existent market. If there
are more than 10 direct competitors, you've either made the segment too general,
you're way late to market, or a huge war of attrition looms.
If your product category has a "standard" segmentation used by industry analysts
(like Gartner or Wall Street), you can't ignore that... but this type of
segmentation will be far too generic to provide
business leverage. For example, "Telecom" isn't a segment: it's an
industry that is economically as large and variegated as Costa Rica.
In envisioning your segments, there are two huge mistakes to avoid. One is
overestimating the customer's interest in or willingness to adopt new
technologies: all too often, the crummy product they have now isn't
causing that much pain. The second is jumping to the conclusion
that because the customer is willing to buy, they're willing to buy from you.
The only way to avoid these mistakes is to survey prospective customers and
At the very least, you can quickly figure out what segments you aren't
in: what kind of people don't have a need, can't afford to solve
it, are too dispersed to be accessed profitably? You can lop off huge sectors of the economy so that you don't fall
into the trap of trying to sell to everyone.
Segmentation for B2B markets typically goes along these lines:
Knee-jerk: company size
(e.g., $250 M-$2 B), company location (e.g., US),
vertical industry (e.g., automotive and defense/aerospace). Get these
data from the Department of Commerce, Lexis-Nexis, Dunn and Bradstreet, and
technology base: users of BEA WegLogic, or IT operations people
using CISCO. You can get these data from vendors and their industry
functional: by business process (e.g., bid-to-cash cycle, product
development cycle), or by organizational affiliation (e.g., CFOs). Data on this can be obtained from
professional groups, industry associations, and the business press.
required qualities / attributes: for example, industrial engineers needing
gasses with parts-per-billion purity, or time-domain reflectometers needing
an 80 dB signal/noise ratio. These data can be hard to find, but start
with professional associations like IEEE or SAE.
behaviors / psychographics: for example, open source java developers,
WiFi road-warriors, fast-track executives. Getting good data on this
is hard, but you can get valuable hints from industry analysts, newsgroups,
blogs, and -- of course -- competitors.
Segmentation for B2C markets typically goes along these lines:
Knee-jerk: age, sex, family status, ethnicity, education, income,
state/province, and other demographics. Get these data from the Census
Bureau, Department of Commerce, and trade magazines for your industry.
Consumer type: first-time user of this kind of product, occasional
user, intense user, loyal user. This data is really only available
surveys and market research firms.
Consumer cohort: college students, urban twentysomethings,
soccer moms, AARP members. These are related to demographics, but are
defined more by whom they associate with than by just "who they are."
This data is mainly gotten from analyzing community membership (like, "people who play
X-box games" or "Macintosh owners.")
Values and Lifestyles: who the consumer thinks they are,
what they value, how they live, and what they aspire to. This is
really hard data to get, but is still some of the most valuable for the
speaking, the more of these axes you can use in defining a target market, the more you understand the
prospective customer. The biggest mistake you can make in
segmentation is to make the segment too general, too large.
Segmentation for Profit
Looking for segments is a whole lot more valuable and rewarding if you already
have a bunch of prospects and customers to work with. Inductive segmentation
has you look at people who are really in your market, and lets you
understand much better who will be your profitable segments. This is
particularly important for
open source and
(three links there) because you have large communities of interest that you want to
Of course, you need to have some hard data to analyze. Don't have a customer
database or CRM/SFA system?
Better get your customer support and marketing people working with surveys so
you can understand who your customers are, what they do, who they work for, why
they bought, and other basics. If you have a community site, run contests
and other promotions to get members to fill out questionnaires whenever they
visit your website.
your community is actually doing, because behavior measurement beats
surveys every time.
Once you have some good data, you'll need to organize it in a multidimensional
database (because you can't know in advance how you want to segment).
You'll probably want to get a consultant or other sharpshooter to crunch the
numbers, unless you have a large enough business to justify this specialization
on permanent staff.
While you're at it, look for repeating patterns in the people who didn't
buy your product or service. Don't have any data on that? Better
provide a monetary incentive for your sales reps to put real info into the SFA
system about the ones who got away. Doing an email survey of people who
lose interest is one of the highest value things a marketer can do.
The whole point of the analysis is to find the clusters of members, prospects,
and customers that have consistent patterns of preferences and behaviors.
So, start from the behavior you're looking for, and develop the segment
definition from that. Segmenting along behavioral lines may surprise you:
Industry and company size are often not very relevant to (or highly
correlated with) decision behaviors.
Geography (particularly country) and job title (particularly rank) may
provide much more effective bases for segmentation.
Company personality (are they leaders or laggards, risk takers or Luddites?)
and internal infrastructure (do they build vs buy, have they standardized on
Oracle or SQLserver?) may also provide highly correlated behaviors.
For open source and viral marketing strategies, the individual's degree of
participation may be the most relevant precursor to purchasing. If
someone is constantly posting to your blog or forwarding your viral offer,
they are exhibiting pre-purchase loyalty that can dramatically increase your
yield. Make sure to add measurement points into all your systems that
support community interaction and participation.
Using inductive segmentation, you may discover a dozen or more clusters of
users...and the danger here is in creating segments that are too small to profitably
pursue. You need to find a way to group them (even in unnatural ways) so
that you have a manageable number of segments to pursue (3-5, tops).
Segmentation -- visionary or inductive -- is most effective with large
communities. If you are working consumer markets, viral campaigns, or open
source projects, segmentation can be make-or-break for your business. So get
some good data on your customers, segment them, listen to what they say, and create an
offering just for them: don't let yourself be Sloaned.
business coming from? -- coming in January
copyright 2006 by DOTnet Consulting, Inc., all rights reserved.
Feel free to forward this report, but you must include this copyright notice.
All trademarks and graphics are the property of their respective owners.
This newsletter was mailed to you because our records show you asked to receive
here to opt-out of future newsletters, although that will make us feel very sad.