A Joint Model of Usage and Churn in Contractual Settings

What is it about?

Any serious attempt to quantify customer lifetime value requires modeling techniques that can provide accurate multi-period forecasts of customer behavior.  Several researchers have explored the problem of modeling customer retention. However, usage (or consumption) under contract influences the value of a customer as well. In contrast to the work on retention, researchers have been surprisingly silent on how to forecast customers’ usage in contractual settings (Blattberg et al. 2008). This research contributes to the literature by developing an integrated model of usage and retention. To do so, we develop a dynamic latent‐trait model in which usage and renewal behaviors are modeled simultaneously as dependent variables by assuming that both behaviors are driven by a common underlying (individual‐level) factor that evolves over time.

Featured Image

Why is it important?

In addition to providing accurate multi-period predictions of customer behaviors (churn and usage), a novel aspect of the model was that it allows researchers to model two behaviors observed at different time intervals (e.g., annual and quarterly, respectively), as is typically the case in most contractual/subscription-based business settings. Using data from a performing arts organization, we demonstrate how our proposed model not only outperforms a set of benchmark models (widely used in practice to predict churn) on several important dimensions, but also provides managerially relevant insights. For example, we show how the model can be used to dynamically segment the customer base and to identify the most common “paths to death” (i.e., stages that customers go through before churn). Interestingly, we find a segment of “walking dead” customers who, despite having made the decision of leaving the company, are unknown to the firm because the contract has yet to end.