Mining customer dynamics in designing customer segmentation using data mining techniques

One of the main problems in dynamic customer segmentation is finding the dominant patterns of customer movements between different segments via time. Accordingly، we concentrate on the customer dynamics in this paper and try to find different groups of customers in transmissions between segments via time. The dominant characteristics of these groups are also investigated. To obtain this objective، a new hybrid technique based on the K-means algorithm، hierarchical clustering and association rule mining is presented and implemented on the data of one of the main telecommunication corporations in Iran. The results show that there are seven different groups of customers. Furthermore، the impact of customer dynamics on segments’ changes via time is investigated. In this regard، a new approach of categorizing customers is proposed according to their impact on the structure and the content of segments’ changes. These new groups include “the customers who preserve the structure”، “the ones who are consistent with the structure” and “the customers who destroy the structure”.