Time to re-think segmentation
Segmentation is a staple of direct and database marketing. Enterprise segmentation is supposed to provide marketers with both a broad and a deep view of their customers. For the investment they require, these initiatives should create serious value through more effective marketing.
But in an age where companies need to be agile and responsive, enterprise segmentation has largely become a dinosaur: a slow, lumbering beast that consumes disproportionate resources from approaches better adapted to dynamic markets.
Facing the business issues
Most of the disagreements about segmentation concern technique: whether to go with survey-based segmentation, behavioral segmentation or a half-dozen other approaches. Given the high cost of such studies, companies should think through several business issues before embarking on such a venture.
First, companies should be prepared to re-allocate marketing resources, both budgetary and human, if the segmentation suggests that they are organized properly to meet the needs of customers in various segments.
Companies should be ready to design new and unique products, messages and offers specific for each of the segments. It's pointless to pay all that money for segmentation if there's no plan to take advantage of the differences.
Be prepared to develop other types of models to support marketing tactics. Segmentation is a strategic tool, and not well-suited for deciding which customers are most likely to respond to a campaign.
When budgeting for the segmentation, simultaneously create a roll-out plan. The plan should include communicating results to the organization, applying the segmentation to test cases, and expanding the use of the segmentation based on past successes.
Weighing the whys and why-nots
It is important to understand why you want segmentation.
Sometimes segmentation is seen as a way to rally support behind an already-selected strategy or to resolve disputes within the organization. One could probably achieve the same outcome at much less expense by simply testing the new strategy in the market place. No segmentation will ever be conclusive enough to win over warring sides. These human problems need to be solved first.
Even if there is good support for the segmentation and the strategies it supports, another difficulty is meeting the diverse operational needs of groups within the enterprise. Segmentations are often designed to provide a broad set of directions, but such versatility comes with a price: it is seldom truly optimal for any of the tactics that the company must implement.
A major blind spot of most segmentation schemes is that they do not anticipate major shifts in the general population. We're not speaking of migration between segments, but fundamental market changes that introduce new segments and diminish others. Traditional segmentation assumes that the environment is stable - but think of how quickly markets can change. One must redo, or at least recalibrate, segmentation regularly.
Reviving segmentation's potential
Is segmentation dead as a concept? In our view, segmentation is still valid, but needs to evolve to be less monolithic and to promote more agile marketing. It should be used segmentation in smaller doses and around specific business challenges.
Here's an example of a technique we have seen used successfully: A typical marketing project would build a model that predicts response, which we can use to score our customers. The typical technology: the logistic regression model.
But we can get a better model by simultaneously segmenting our customers - using a technique called latent class regression. It enables us to discover "latent" (or hidden) segments in our data. When combined with regression, we get a model that builds segments and a better predictive model at the same time. It's a more precise targeting tool and a mini-segmentation that informs us on how to tailor versions of our communication. This technique has another benefit: Since we know our new customer segments are tied to responsiveness, we can use these segments for prospecting.
Even if you have an enterprise segmentation that is creating value for your company, you can still benefit by incorporating additional segmentation into other analytical tools. With today's technologies, this approach can be a more cost-effective solution as well.