How to Incorporate RFM Segmentation With Predictive Models
Consider these best practices when applying RFM to your segmentation process:
• Use common sense. With durable goods and large ticket items (i.e., cars, furniture), RFM may actually work in reverse (where less recent is better than more recent). This also may be true with seasonal business. Be flexible in how you operationalize the values within an RFM model. For example, Web marketing and e-commerce have their own RFM. Recency may be measured in days, not months. Frequency may turn into how frequently visitors return to your site, what they purchase or view, and where they click. Monetary may be extended to include what’s in the shopping cart (before abandons).
• Seasonality makes defining the recency ranges challenging. Does your zero- to 12-month range include “back to school,” Valentine’s Day” and “holiday shopping”?
• Augment RFM where you can if it helps to understand your business dynamics. Direct marketing innovator Bob Kestnbaum pioneered the concept of RFM, adding a “product” dimension to RFM. Kestnbaum called it “FRAC,” where “F” was frequency (the first and most predictive variable in his mind), “R” was recency, “A” was average dollars and “C” was category of purchase. The best predictor of future product purchases is past product purchases. This addresses the challenge of seasonality. Your key business levers also are viable candidates to extend the simple RFM summary. For B-to-C direct marketing, this might be RFMI (income); for B-to-B, this might be RFMI (industry). Both sectors might see benefits in an RFMC (channel) segmentation scheme.
• RFM also can be put to work to acquire new customers. There is some current thinking on calculating RFM for each ZIP code on file to target prospects based on observed RFM for customers that live in the same neighborhoods.
Roy Wollen is CEO and managing consultant for Database Insight Inc., a Chicago-based direct and interactive marketing firm. Contact him at email@example.com or call (312) 629-5043.