The method of finding potentially churning customers based on social networks

Abstract

Customer churn analysis has become an important focus of corporate marketing. It will be a great help to profitability if there is a method can find losing customers in time. In the paper, a method based on RFM and Cross-correlation model is proposed. Firstly, the customer’s value is calculated by RFM. Secondly, the typical losing curves of customer value are matched via cross-correlation. And finally, integrated with social network analysis (SNA) and community detection, the group of potential losing customers are revealed. The effectiveness of the presented method has been proven in a dataset of retail sales records.