Enhancing e-commerce customer churn management with a profit- and AUC-focused prescriptive analytics approach

Abstract

This study introduces a profit- and AUC-focused prescriptive analytics method (PAM) grounded in big data analytics capability, as supported by the dynamic capabilities theory, to manage customer churn in the e-commerce sector. This method accounts for the diversity in customer lifetime value and the associated costs of incentives to accurately evaluate the expected maximum profit (EMPB). PAM not only balances EMPB and AUC effectively but also prescribes optimal actions to align with various decision-makers’ preferences, enhancing both business and predictive outcomes. Our experiments, validated by a real-world case study, demonstrate PAM’s adaptability and superior performance in managing customer churn. Moreover, optimal action is explained by leveraging interpretable data science methods to provide clear insights into decision-making processes, further emphasizing its role as a big data analytics capability in a changing business environment.