Study on Ensemble Algorithm to Increase Customer Lifetime Value

Abstract

In retail and service industries, it is strategically important to identify a customer segment whose lifetime value (LTV) is higher, and to continue trading with them. However, it is not so easy because it is difficult to identify factors that are important to continue trading. In this study, we identify a customer segment that has higher LTV using ID-POS data of a hair salon chain, and find the factors to continue trading with them using a classification approach. We propose an ensemble algorithm consisting of multiple stages using several white box classification algorithms. From computational experiments using the data, we show that our results outperform our previous results.