Incorporating Direct Marketing Activity into Latent Attrition Models

Abstract

When defection is unobserved, latent attrition models provide useful insights about customer behavior and accurate forecasts of customer value. Yet extant models ignore direct marketing efforts. Response models incorporate the effects of direct marketing, but because they ignore latent attrition, they may lead firms to waste resources on inactive customers. We propose a parsimonious model that allows direct marketing to impact three relevant behaviors in latent attrition models—the frequency with which customers conduct transactions, the size of the transactions, and the duration for which customers remain active. Our model also accounts for how the organization targets its direct marketing across individuals and over time. Using donation data from a nonprofit organization, we find that direct marketing increases donation incidence for active donors. However, our analysis also shows that direct marketing has the potential to shorten the length of a donor's relationship. We find that our proposed model offers superior predictive performance compared with models that ignore the impact of direct marketing activity or latent attrition. We demonstrate the managerial applicability of our modeling approach by estimating the impact of direct marketing on donation behavior and identifying those donors most likely to conduct transactions in the future.

Some Customers Would Rather Leave Without Saying Goodbye

“…Because identifying churn is already a challenge (as it is unobserved), the literature on noncontractual settings has mostly focused on identifying which customers are most likely to have "died" (i.e., silently churned), rather than identifying which factors precede such latent attrition. Notable exceptions are Braun et al (2015), Knox and Van Oest (2014), Schweidel and Knox (2013), and Schweidel et al (2014), which build on the aforementioned latent attrition models and allow for covariates such as direct marketing activity and characteristics of customer transactions to impact the probability of (silently) churning.…”

In Pursuit of Enhanced Customer Retention Management

“…Much less work has been devoted to predicting attrition in non-contractual settings, where attrition is latent. More work is needed to evaluate the accuracy of latent attrition measures and its predictors (e.g., [80]).…”

A Multiactivity Latent Attrition Model for Customer Base Analysis

“…Depending on the value of kj , customers may become more or less likely to remain active on activity k. 4 If kj = 0, customer h's tendency to become inactive on activity k would be time invariant (e.g., Fader et al 2010). If other timevarying covariates were available and believed to affect the transaction incidence or attrition processes, they could be incorporated into Equations (1) and (6) (e.g., Schweidel and Knox 2013). Although Equation (6) allows for heterogeneity across customers and the effect of recent activity on the attrition processes, it omits two important factors that may affect the activity-specific attrition processes in a multiactivity setting.…”

“…To illustrate the model's ability to identify those customers who are likely to conduct transactions during the forecasting period, we rank customers in descending order using the expectations of purchasing and posting incidences during the forecasting period. In Figure 4, we present the proportion of customers conducting a transaction during the forecasting period based on these rankings (Schweidel and Knox 2013).…”

“…As we demonstrate, these relationships affect inferences about the likelihood with which customers may conduct activities in the future. By incorporating touchpoints beyond purchasing into our analysis, we can better distinguish those customers who are likely inactive from those who may make purchases in the future and thereby provide guidance for a firm's resource allocation across customers (e.g., Schweidel and Knox 2013). Our research also suggests the potential value associated with encouraging the production and posting of user-generated content (e.g., Trusov 2011, Moe andSchweidel 2012), as it may contribute to revenue by reducing the likelihood with which customers cease conducting purchases.…”