Segmentation Sanity

Segmenting your market is like assembling a 1,000 piece jigsaw puzzle … where some of the pieces are missing, none fit together very well, and a few might even be from a different puzzle.

I was reminded of this madness thanks to two separate clients. One is in the energy trading risk management (ETRM) software space and the other is in language services. Very different industries, very different customer verticals, hugely different buyer genotypes, and many of the exact same segmentation difficulties.  Their mutual difficulty was not in selecting relevant vectors on which to segment their markets, or divining the needs of any segment. The problem was the conflict between the theory of a segment and the reality. For example, one of these two clients had a subsegment that was largely untapped and, on paper, looked highly profitable. The reality was that most of this untapped segment contained companies that didn’t need my client’s product because of a fundamental difference in the way they did business from all the other segments.

Another client had the problem of too many subsegments. Even looking at the comparatively simple matrix created by industry verticals and regions, my client had 27 different primary segments. Deeper evaluations into each showed a myriad of variations, enough to keep their marketing intern busy until retirement.

Marketing people can squander copious quantities of money and time attending to too many segments, and for two reasons. First, many segments simply don’t matter. Take for example the client who had a segment populated by companies with odd business models. Once those odd ducks were eliminated, the remaining dollar value of that segment was near the bottom of the list and not worth direct promotions (casual or indirect promotions perhaps).  Elimination of needless work segments should be high on any marketing managers to-do list, though it takes some guts to look at your CMO and say “We should not sell to this group.”

More importantly though is that marketing people tend to be a little too controlling. The Internet and cheap web hosting provides great ways to allow people to self-segment. Simple web forms that gently allow visitors to pigeonhole themselves are affordable to anyone, and used by very few. After all, you are going to ask these people to give you money, asking nicely for a little data isn’t going to detour them … much.

Occasionally you see ham-handed attempts to get people to file themselves into one or another classification. Often these implementation create such end-user friction that the prospect vanishes instead of helping you sell to them. Take for example any landing page that has a form, and in that form are either too many questions or some questions with too many options. People, being members of an over-worked and under-attentive species, will default values or abandon the page entirely.

Several keys exist to customer self-segmentation. First, keep your forms simple and short. Second, give visitors reason to interact again and ask for a little more information each time. Third, use a variety of landing pages and honey pot text to attract people from all your target segments. Forth, do a little cluster analysis to see if your theorized segmentation model is working or not (if you see visitors clustering in too few segments, than either your model is wrong, your keywords are incomplete, or some of those segments are losers).

The marketing lesson is to make computers and customers do more of your work. This allows for three martini lunches that making afternoon productivity optional.


Speak up! What are your thoughts?

Your email address will not be published.