Diagnosing and Predicting Individual Customer Defection

• Main Authors: Bonfrer, Andre, Knox, George, Eliashberg, Jehoshua, Chiang, Jeongwen
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2007
• Subjects: Marketing
• Online Access: https://ink.library.smu.edu.sg/lkcsb_research_smu/1; https://ink.library.smu.edu.sg/cgi/viewcontent.cgi?article=1000&context=lkcsb_research_smu
• Institution: Singapore Management University

In a contractual setting, the firm observes when the customer defects. When the service is continuous, the firm also observes each customer’s usage path over time. We model this usage rate over time as a degradation process. Defection occurs when this process reaches an absorbing boundary. We develop two models to study this process, estimable at the individual level and estimable before the defection event occurs. We demonstrate that these models can provide useful diagnostic information about the service provider’s customer base through their coefficients. They are also able to discriminate between customers who defect and those who do not, before any defection has occurred. Thus, these models are useful as an early warning system for the firm to identify likely defectors in their customer base. Based on the model parameters, we also develop customer lifetime value calculations. A key managerial insight stemming from this research is that there is a premium for usage volatility.