Technical Reports


Customer repeat purchase modeling: A Bayesian Hierarchical Framework

Pal, Jayanta Kumar; Misra. Subhasish
HP Global Business Services : Decision Support and Analytic Services


Abstract: Two of the major queries in a database marketing/CRM are of predicting the churn of a customer; the frequency of his repeat purchases. Data mining techniques have often been used to approach these, albeit with limited success. In this paper we develop a methodology, using a Bayesian analysis framework to answer these questions. Using the answer to these questions as inputs we predict the likelihood of customers to make a transaction within a time span (in the next six months/one year) in the future. This likelihood/propensity to buy can in turn can be used to rank customers. This leads to more efficient targeting for marketers. The proposed model when tested using EMEA store and HHO data yielded very satisfactory results.

7 Pages

External Posting Date: July 21, 2010 [Abstract Only]. Approved for External Publication - External Copyright Consideration
Internal Posting Date: July 21, 2010 [Fulltext]

Back to Index