Eva Ascarza has been selected to receive the 2023 Weitz-Winer-O’Dell Award for her article, “Retention Futility: Targeting High-Risk Customers Might be Ineffective,” which appeared in the February 2018 issue (Volume 55, Issue 1) of Journal of Marketing Research (JMR).
The Weitz-Winer-O’Dell Award honors the JMR article published five years earlier that has made the most significant, long-term contribution to marketing theory, methodology, and/or practice. The award committee this year consisted of Tulin Erdem (New York University), Maureen Morrin (Rutgers University), and Michel Wedel (University of Maryland). The committee provided the following statement about their choice of Ascarza’s article for this award:
The “Retention Futility” paper by Ascarza seeks to improve marketers’ customer retention efforts by helping them determine which customers should receive promotions that are designed to boost repurchase and resubscription. Retaining customers has long been known to have a positive effect on the long-term profitability of the firm (Gupta et al. 2004). Traditionally, firms’ retention efforts have focused on identifying those customers most likely to defect via churn models, so that, once identified, such customers can be sent a promotion or other marketing intervention that will reduce their likelihood of defection. Ascarza shows that customer churn is reduced more effectively not by targeting such interventions to customers most likely to defect (i.e., high risk), but rather to those most likely to respond to the offer (i.e., high sensitivity). More broadly, the paper shows how leveraging information about customer heterogeneity in responsiveness to A/B testing can significantly improve the effectiveness of such efforts.
The author’s approach involves running targeted counterfactuals based on customer heterogeneity. It applies the potential outcomes framework for causal inference (Rubin 2011) to estimate the conditional average treatment effect (CATE) using random forests (Guelman et al. 2015). A firm utilizing this approach would first conduct a pilot study on a random sample of their customers to assess the effectiveness a planned marketing intervention designed to reduce churn in an A/B test format. The pilot results would then be used to model intervention sensitivity as a function of historic customer demographic and behavioral information. The results would be used to more effectively target the promotion to a subset of the remaining (i.e., non-pilot) customers in the firm’s database.
Two field studies were conducted to illustrate the effectiveness of this approach. One field study was conducted at a subscription-based membership organization in North America, which added a gift to a renewal communication in an A/B test format. Another field study was conducted at a wireless telecommunications firm in the Middle East which texted an offering of additional credit for recharging. Response to the offerings was tracked and customer sensitivity modeled as a function of historical customer data residing in the firms’ databases.
The results of the field studies suggest that marketers may be better off targeting customers most likely to respond to the intervention than to those most likely to churn. Targeting by customer sensitivity to the offer rather than churn risk reduced churn rates 4.1 to by 8.7 percentage points. In some instances, it was found that targeting customers most likely to defect actually increased churn rates.
Focusing retention efforts on customer sensitivity rather than customer risk runs contrary to traditional assumptions associated with strategies focused on minimizing churn. The results of the Ascarza paper have generated considerable interest among practitioners. The paper’s impact on the practice of marketing is reflected in its having been presented at numerous corporate venues such as Facebook, Google, Microsoft, T-Mobile, and Zillow. The paper was also selected as a “must read” by the Marketing Science Institute (MSI).
As the retailer John Wanamaker is famously said to have quipped, “Half of the money I spend on advertising is wasted, the trouble is I don’t know which half.” Part of the appeal of the Ascarza paper’s approach is that it allows marketers to begin to know which half of their marketing spend may be wasted, allowing for a redirection of marketing expenditures toward more effective targets. The committee believed that this paper has had a long-term impact on research in academia and practice as well as on teaching, as evidenced amongst others by its high citation count (it is the paper with the highest number of citations among all papers published in JMR in the year it was published). It challenged extant assumptions on the effectiveness of targeting in academia and practice, collaborated with two companies to conduct field experiments, applied state-of-the-art methods of causal inference, and was the first to conduct targeting counterfactuals based on the heterogeneous response of marketing interventions to move beyond assessing average treatment effects.
Four other excellent papers were named finalists for the 2023 Weitz-Winer-O’Dell Award:
- Yiting Deng and Carl F. Mela (2018), “TV Viewing and Advertising Targeting,” Journal of Marketing Research, 55 (1), 99–118.
- Brett Hollenbeck (2018), “Online Reputation Mechanisms and the Decreasing Value of Chain Affiliation,” Journal of Marketing Research, 55 (5), 636–54.
- Eugina Leung, Gabriele Paolacci, and Stefano Puntoni (2018), “Man Versus Machine: Resisting Automation in Identity-Based Consumer Behavior,” Journal of Marketing Research, 55 (6), 818–31.
- Daniel M. McCarthy and Peter S. Fader (2018), “Customer-Based Corporate Valuation for Publicly Traded Noncontractual Firms,” Journal of Marketing Research, 55 (5), 617–35.
Congratulations to the winner and the authors of the four other finalist papers! Eva will be recognized during the awards luncheon at the AMA Summer Academic Conference in San Francisco (August 4–6, 2023). A special session will be devoted to the winning article and the other finalists at the conference as well.
Guelman, Leo, Montserrat Guillén, and Ana M. Pérez-Marín (2015), “Uplift Random Forests,” Cybernetics and Systems, 46 (3/4), 230–48.
Gupta, Sunil, Donald R. Lehmann, and Jennifer Ames Stuart (2004), “Valuing Customers,” Journal of Marketing Research, 41 (1), 7–18.
Rubin, Donald B. (2011), “Causal Inference Using Potential Outcomes,” Journal of the American Statistical Association, 100 (469), 322–31.
Go to the Journal of Marketing Research