Customer retention is a common concern for many industries and a critical issue for the survival in today’s greatly compressed marketplace. Current customer retention models with data mining techniques only focus on detection of potential defectors based on the likelihood of defection by using demographic and customer profile information. In this paper, we propose a dynamic procedure for detecting and preventing defection using past and current customer behavior by utilizing SOM and Markov chain. The basic idea originates from the observation that a customer has a tendency to change his behavior (i.e. trim-out his usage volumes) before his eventual withdrawal. This gradual pulling out process offers the company the opportunity to detect the defection signals. With this approach, we have two significant benefits compared with existing studies for detecting defection which are based on the likelihood of defection. First, this suggested procedure can predict when the potential defectors could withdraw and this feature helps to give marketing managers the information of lead-time for preparing defection prevention plans. The second benefit is that our approach can provide a procedure for not only detecting but also preventing defection, which could suggest the desirable behavior state for the next period so as to lower the likelihood of defection.
We applied our dynamic procedure for detecting and preventing defection to the online gaming industry. And we compared the prediction accuracy of the proposed procedure to that of the MLP neural network and a decision tree model. This suggested procedure could predict potential defectors and provide the important information of lead-time without deterioration of prediction accuracy compared to that of MLP neural network and a decision tree.