Online Program

Thank You, Come Again: Modeling Repeat Purchase Behavior for Business Travelers (303607)

 View Presentation

Keywords: stochastic processes, consumer behavior, revealed preferences

Across many industries, new businesses have a vested interest in understanding fluctuations in a customer’s propensity to buy over the course of that customer’s life cycle. This presentation will discuss building Spark- and Python-based models to estimate such fluctuations and develop predictions for the length of the life cycle. The patterns in customer purchase behavior can be modeled as a stochastic process where the probabilities of moving on to the next purchase can be inferred from logistic regressions of customer attributes and attributes of the previous purchases. This presentation will discuss transforming the raw operational data into a model-ready format (exploring some of the practical concerns), statistical specification of the model, and generalizability to any repeat customer behavior.