Commentary

Overlooking Segmentation

As the debate rages over which measurement methodology provides the best perspective on digital (or analog, for that matter) advertising effectiveness, I think there's a tendency to overlook something important: segmentation.

The behavior-driven trackability of the online customer creates a belief that we're dealing with the ultimate "segment of one," where any customer's clickstream and cookies gives us enough clues as to who they are and what they need. This is a far safer assumption in some categories (e.g., hotel reservations) than in others (automotive or healthcare). And it affects how we should think about measurement.

"Segmentation" seems to be on its way to becoming one of those co-opted words that is fast losing its meaning. As I've learned painfully with the word "dashboard," once a concept catches on, it quickly becomes perverted beyond recognition -- until, on some level, everyone believes they are doing it, and all consultants and technology providers are experts in it (CRM, anyone?).

advertisement

advertisement

Lately, I've seen "segmentation" used to describe everything from finding the most likely prospects for an existing product/service set to explaining the similarities and differences between competitors in a given market space.

Now, I've never been too hung up on definitions. But I do place a high value on clarity of meaning. To me, "segmentation" is the process of defining naturally occurring groups of homogenous prospects or customers who share a common need-set, determining the relative size of each group (from the perspective of profit potential), prioritizing their attractiveness, and then developing go-to-market plans best suited to appeal to each. This requires some investment in quantitative research of some sort -- but not just about perceptions. Segmentation based on attitudes or beliefs is fine for defining positioning statements or writing copy. But it often overlooks the trade-offs in constructing the real value proposition of the product/service by under-weighting the importance of making your product easier to spot on the shelf, or modifying your distribution-channel structure, or even extending credit to achieve competitive advantage.

And of course, segmentation without sizing is irrelevant. Whether you size on revenue or contribution margin (preferred), you need to understand the relative opportunity of allocating your resources one way versus another. Good segmentation forces you to make hard decisions because it shows you multiple viable pathways. Great segmentation helps you quantify the risk/reward propositions and leads you to the best choice.

How, you may be asking, does this relate to online marketing measurement? Segmentation is the basis of all effective resource allocation. Segment your market properly, and your most meaningful metrics will emerge from your understanding of how to create customer value. The measurement vehicles that are "right" for you will be those which best lend themselves to understanding the unique behavior of your priority segments, not just the overall efficiency of eyeball delivery or rate of response.

There's huge opportunity in appealing to that "segment of one" when you can spot them. But you'd better have measurement systems that guide you in increasingly targeted and relevant increments as you seek to draw them toward interacting with you.

Next story loading loading..