2. Design around the objectives. The key to effective design is working back from the business decisions that need to be made. Once the objectives have been determined, the segmentation research itself must be rigorously designed to reflect them, and to ensure that the results will be insightful (they will tell us things we do not already know about customer behavior and needs), actionable (they will identify levers that will move behavior), and identifiable (they will be able to tag individual customers in the database with reliable segment membership). This implies that multiple dimensions — behaviors, attitudes, demographics, channel use and preferences, and profitability — must be incorporated to develop a full picture.
In the past, difficult choices had to be made about which one or two dimensions drove segmentations, leading to, for example, heavily attitudinal segments that could not be identified or heavily behavioral segments that provided no insight into the causes of behaviors or the ways to influence them. New methodologies, such as the statistical modeling technique of latent class analysis, allow the synthesis of these different types of data — for example, transactional data, survey data, continuous variables, and discrete variables — into the same segmentation model. If designed properly, these methodologies yield identifiable segments that can be used to pursue new opportunities. Common examples include fast-food chains identifying their frequent users and wireless companies looking for “data hogs” who consume inordinate amounts of bandwidth. More sophisticated companies, such as some high-end retailers, add aspirational desires to the behavioral mix to identify and attract consumers who lust for the latest brand as soon as (or before) it hits the shelves.
One detail to consider is how to approach customer profitability. Some companies use profitability-based segmentation, others incorporate profitability into the segmentation as a dimension, and still others use it as a criterion for choosing among potential segmentations based on their ability to discriminate among profitable customers. Each has its benefits and costs, but it is important to recognize that a pure profit segmentation may not be very actionable, because in some industries, such as financial services, there are multiple routes to profitability.
3. Prepare a blueprint to operationalize the segmentation. Begin to map out the decision processes by considering these questions: How is segmentation going to be used to influence the major value levers in your business? Will it underpin a redesign of the brand and the value proposition or of frontline sales and service? Will it ultimately result in more tailored and dynamic online and direct marketing? How are the business and functional units responsible going to access the information and use it on a day-to-day basis?
As soon as the outline of the segmentation permits, begin to define these process changes, share them with affected business partners, and formulate and discuss revised metrics that reflect the new capabilities. For instance, it is essential to work with the front line to define how it will modify today’s sales and service protocols to incorporate greater customer knowledge and targeted pitches. One large bank linked its strategic segments to targeted lead lists focused on explicit behaviors — cross-sell leads, at-risk/retention leads, and acquisition leads — and instituted a new and more powerful sales pipeline management discipline.
In another example, a major retailer redesigned the location, configuration, and staffing of its fitting rooms based on a target segment view of the customer’s journey through the store. It is necessary to craft practical methods of delivering the required segmentation results that effectively conceal their analytic complexity from customer-facing staff, thus enabling productive face-to-face interactions with customers in real time.
Subtle timing must be managed: Business partners need to be engaged and prepared in advance, but not so far in advance that requisite changes are too theoretical or abstract. For executives at brokerages, for example, communicating the message that “We need to talk to segment X about 529s and Roths in the context of their families” is much more meaningful than “We need to be more segmented in how we are going to market.” Building on what’s tangible helps in configuring informed decision processes.