Utilizing customers’ purchase and contract renewal details to predict defection in the cloud software industry

Abstract

This study aims to predict customer defection in the growing market of the cloud software industry. Using the original unstructured data of a company, we propose a procedure to identify the actual defection condition (i.e., whether the customer is defecting from the company or merely stopped using a current product to up/downgrade it) and to produce a measure of customer loyalty by compiling the number of customers’ purchases and renewals. Based on the results, we investigated important variables for classifying defecting customers using a random forest and built a prediction model using a decision tree. The final results indicate that defecting customers are mainly characterized by their loyalty and their number of total payments.