A stochastic RFM model

A central problem in database marketing is how to choose which customers in the firm's database to target with an offer. This paper presents a simple stochastic RFM model to carry out such a task. By making a few straightforward assumptions about the customers in the database, the stochastic model provides a means of (1) ranking customers in terms of their expected contribution (or lifetime value) and (2) deciding how deep to go in the customer file to make the offer. The model provides a more insightful alternative to scoring or segmenting customers than regression or cross-tabulation methods. The approach is demonstrated with an example from the market research industry.