I'm Here, But Maybe Not for Long: Using Complaints to Predict and Value Customers

Abstract

The extant literature has related Customer Lifetime Value [CLV] to the customer's purchase history, but has paid little attention to other non-transactional customer-firm interactions like complaints. This paper makes three contributions. First, we build an easy-to-implement model that incorporates complaints into customer base analysis. Second, applying this model to more than two years of individual data from a major U.S. catalog retailer, we demonstrate that the occurrence and timing of customer complaints offer predictive value beyond the occurrence and timing of purchases. Lastly, we assess the link between complaints, purchases and CLV, which allows us to investigate the conditions under which observing a customer complaint increases or decreases CLV. We argue that a customer complaint may provide two opposing signals. The negative signal concerns potential defection. The positive signal is that the customer is still active and involved with the firm.

Keywords: Complaints, Customer Lifetime Value, Probability Models, Forecasting, Customer Base Analysis