Improving customer loyalty evaluation methods in the grocery retail industry: a data mining approach

Abstract

Evaluating customer loyalty is an issue, which has gained a lot of attention in recent years due to modern facilities and tools for gathering and analysing data. These evaluations have had great and significant effects on improving business processes. Accordingly, data mining methods present significant capabilities. On the other hand, common methods for evaluating customer loyalty have been developed only based on three components, including recency (R), frequency (F) and monetary (M). In this study, it has been tried to add some other effective factors including number of bought products, number of returned products, amount of discount and delivery delay to the analysis in order to measure the impact of each one of them on the quality of the evaluation. The ideas and opinions of experts and the current available literature on the subject have been used as criteria for assessing quality. While implementing the methods, machine-learning tools such as artificial neural networks and support vector machine have been utilised. The results show that the method where the four factors are simultaneously fed into the RFM presents the highest possible accuracy in evaluating customer loyalty and among the learning models, the MLP-boosting method provides the highest accuracy.