New Approach for Customer Clustering by Integrating the LRFM Model and Fuzzy Inference System

Abstract

This study aimed at providing a systematic method to analyze the characteristics of customers' purchasing behavior in order to improve the performance of customer relationship management system. For this purpose, the improved model of LRFM (including Length, Recency, Frequency, and Monetary indices) was utilized which is now a more common model than the basic RFM model apt for analyzing the customer lifetime value. Since the RFM model does not take the customers' loyalty into consideration, the LRFM model has instead been applied for making amendments. Contrary to most of the past studies in which the statistical clustering techniques were used besides the RFM or LRFM model, the current study has provided the possibility of clustering analysis by importing the LRFM indices into the framework of a fuzzy inference system. The results obtained for a wholesale firm based on the proposed approach indicated that there was a significant difference between clusters in terms of the four indices of LRFM. Therefore, this approach can be well utilized for clustering the customers and for studying their characteristics. The strong point of this approach compared to the older ones is its high flexibility, because in which it is not needed to re-cluster the customers and to reformulate the strategies when the number of customers is increased or decreased. Finally, after analyzing the attributes of each cluster, some suggestions on marketing strategies were made to be compatible with clusters, and totally, to improve the performance of customer relationship management system.