Abstract

The computation of customer lifetime value is an important metric for strategic decision making. However, there has been limited exploration of the complexities of computing customer lifetime value in a contractual or continuous payment relationship. In this dissertation, three specific challenges are addressed in the CLV area. In the first chapter, a simple duration model is used to explore the impact of billing frequency on customer duration using a non-parametric hazard model. In the second chapter, a customer migration model is proposed with is capable of handling both duration dependence and purchase decision periodicity that is often present in continuous payment service relationships. In the final chapter, a model for multiple services is proposed which is capable of modeling the complex nature of customer lifetime value when more than one service relationship is present. In all three chapters, it is shown that each of these issues have significant effect on model fit statistics. While modeling periodicity improves a model’s predictive validity, the multiple service dependency model’s improvement in predictive validity is inconclusive. However, the proposed multiple service lifetime value model (MSCLV) allows managers the ability to quantify the risks of cancellation and purchase for both single and dual service subscribers. Using data from an online content provider, it is shown that customers are more likely to purchase a service when they own another service from the focal firm, when compared to inactive customers. It is also shown that customers with dual subscriptions are less likely to cancel their subscriptions than customers with a single service. These models provide a valuable first step in the development of models capable of handling some of the complex issues existent in contractual continuous payment relationships.