RFM analysis optimized

Abstract

In the classic recency, frequency and monetary approach to market segmentation, i.e. RFM analysis, given a time frame, customers are clustered together into an arbitrary number of segments according to their most recent day of purchase, the number of purchases and the monetary value of their purchases. In this work we show how the choice of the number of segments and the time frame used in the RFM segmentation process can be optimized to maximize the result of direct marketing campaigns. We also indicate how RFM analysis can be extended to accommodate new dimensions of customer behavior and how the extended RFM analysis can be optimized. Furthermore, we discuss the implications of the optimized and extended RFM approach to market segmentation for direct marketing and business strategies.