Applying decision trees for value-based customer relations management: Predicting airline customers' future values

Abstract

This paper outlines a general methodology for the segmentation of customers according to lifetime value and prediction of their future value segments based on demographic and behavioural features. Not only have few previous scientific studies dealt with this question using nonaggregate single-customer data, but also we address the problem using decision trees, which has rarely been done before. Applying our method to customer data from a major European airline produces predictions not only for long-term, middle-term, and short-term customers but also for prospects whose value is difficult to predict because of limited information. Overall, our findings provide marketing managers a means to optimise company customer equity and maximise returns on marketing in the long run.