A machine learning framework for customer purchase prediction in the non-contractual setting

Predicting future customer behavior provides key information for efficiently directing resources at sales and marketing departments. Such information supports planning the inventory at the warehouse and point of sales, as well strategic decisions during manufacturing processes. In this paper, we develop advanced analytics tools that predict future customer behavior in the non-contractual setting. We establish a dynamic and data driven framework for predicting whether a customer is going to make purchase at the company within a certain time frame in the near future. For that purpose, we propose a new set of customer relevant features that derives from times and values of previous purchases. These customer features are updated every month, and state of the art machine learning algorithms are applied for purchase prediction. In our studies, the gradient tree boosting method turns out to be the best performing method. Using a data set containing more than 10 000 customers and a total number of 200 000 purchases we obtain an accuracy score of 89% and an AUC value of 0.95 for predicting next moth purchases on the test data set.