Carnegie Mellon University

Accounting for Seasonal and COVID in CLV Model Predictions for a Retail Chain

Capstone Team: Andrea Levy, Michael Zakin, Vivek Jayaprakasan

Customer Lifetime Value (“CLV”) modeling focuses on measuring and understanding the value of a customer based on past customer transaction data. Reliable models that accurately predict CLV is a critical tool that helps a company understand customer purchasing behavior, enabling managers to make focused marketing decisions.

Two common “buy-’til-you-die” probability models are used to predict CLV in non-contractual customer retail environments: the Pareto-NBD model (Reference 1), and the BG-NBD (Reference 2) model. One of the primary benefits of these types of models is that they work with aggregate data and are easy to implement.

Capstone Presentation