Modelling Customer Churn Using Segmentation and Data Mining

Customer churn management has drawn much attention from many researchers and practitioners to improve customer retention. The term churn is related to predictions on when a customer abandons his relationship with a company; therefore it has become mandatory for most organizations seeking sustainable and profitable growth. Also increasing in churn rates make companies confront the inevitable heavy marketing campaigns to retain or acquiring new customers. Current churn literature reveals the fact that acquiring new customers costs more than keeping existing ones. However, studies related to churn management mainly focused on methodological improvements regarding the predictive ability, which failed to illustrate a dynamic process in the change of customers' churning behaviour. This paper proposes a model with multi-dimensions of customer churning level via combining segmentation concept within data mining framework to expand the prediction of customer churn. Additionally, comparison to other prediction models, proposed model provides more accurate predictions on customer behaviour and better understanding of relationship between customer and company, mostly applicable in service providing sectors. The potential implications of the model for managers and practitioners are also provided within the paper.