Purpose � Retailers realize that customer churn detection is a critical success factor. However, no research study has taken into consideration that misclassifying a customer as a non-churner (i.e. predicting that (s)he will not leave the company, while in reality (s)he does) results in higher costs than predicting that a staying customer will churn. The aim of this paper is to examine the prediction performance of various cost-sensitive methodologies (direct minimum expected cost (DMECC), metacost, thresholding and weighting) that incorporate these different costs of misclassifying customers in predicting churn.
Design/methodology/approach � Cost-sensitive methodologies are benchmarked on six real-life churn datasets from the retail industry.
Findings � This article argues that total misclassification cost, as a churn prediction evaluation measure, is crucial as input for optimizing consumer decision making. The practical classification threshold of 0.5 for churn probabilities (i.e. when the churn probability is greater than 0.5, the customer is predicted as a churner, and otherwise as a non-churner) offers the worst performance. The provided managerial guidelines suggest when to use each cost-sensitive method, depending on churn levels and the cost level discrepancy between misclassifying churners versus non-churners.
Practical implications � This research emphasizes the importance of cost-sensitive learning to improve customer retention management in the retail context.
Originality/value � This article is the first to use the concept of misclassification costs in a churn prediction setting, and to offer recommendations about the circumstances in which marketing managers should use specific cost-sensitive methodologies.