A Multi-indicator Customer Segmentation Method Based on Consuming Behaviors Analysis

Abstract

Customer consuming data mining reveals many liable to be neglected information. Instead of classifying customers just by common demographic characteristics, behavioral patterns based customer segmentation can divide customers into homogeneous groups by consuming behaviors analysis, thus provides a dynamic and deep insight of customer. This paper puts forward a new multi-indicator RFM method for customer segmentation by consuming data mining. Firstly we construct a multi-indicator RFM model containing 10indicators in 3 groups, and extract inherent segment variables from these indicators by factor analysis method, then apply partitioning based clustering algorithm for customer segmentation. Through analysis of the sales data of a large Chinese home textile company with about 1 million data records, we demonstrate that our method can provide reliable detection rate and rich interpretation ability on customer behavior.