Customer Analytics: Using Machine Learning to Predict the Customer Life Time Value (CLTV) Based on Purchase History and Ratings

Customer Life Time Value (CLTV) is a vital customer analytics metric that is intended to inform companies on the right investment in customer acquisition, retention, and engagement. Conventional CLTV models are heavily dependent on purchase frequency and money-value transaction information, but not customer satisfaction measurements like ratings. This paper proposes a general framework for incorporating ratings into machine learning models for CLTV prediction. Taking advantage of the merits of ensemble learning methods, the paper uses both transactional and satisfaction data to uncover the complex customer behavior dynamics. Our experimental finding indicates the better predictive performance of satisfaction-based models, providing companies with a sound basis to enhance customer analytics activities.