Modeling Sources of Heterogeneity in Customer Lifetime Purchase Behavior

In cases where a firm does not observe customer defections, two key issues in analyzing customer lifetime purchase behavior are how often will a customer purchase from the firm (purchase frequency) and how long will the customer continue purchasing from the firm (customer lifetime). Furthermore, it is well documented in the marketing literature that customers are heterogenous in terms of their purchase behavior. These are also important because they have a direct impact on the quality of decisions made by firms such as those related to customer retention, new customer acquisition, and customer promotions. Existing models dealing with such issues have incorporated heterogeneity (both observed and unobserved) in a limited manner. In addition, these models make assumptions that are restrictive in nature. We propose a new model called the Normal-NBD model that addresses the issues outlined above in a comprehensive manner. We compare the performance of the proposed model using data by two key existing models, i.e., the Pareto-NBD (the seminal model) and the GGompertz-NBD (the latest model). We also show how the availability of richer time series data can be used to better understand and predict customer lifetime purchase behavior.