Incentives for Digital Advertising Are Misaligned with Consumers' Responses to Ads
Every month, firms serve millions of digital ad impressions--advertisements shown to
consumers. In programmatic advertising, the firms outsource bidding for ad impressions to ad platforms so they can sell their space to the best possible advertisers. Advertisers in turn want to show the ads to users who will then buy their product.
In a recent study, researchers developed a theoretical model to determine whether
incentives for ad platforms like Google are aligned with advertisers. In an ideal world,
advertisers and ad platforms should have same incentive – increase the impact of advertising. However, the way the ad platforms charge money to advertisers create misalignment in their incentives. Ad platforms generate money if a user clicks on the advertisement (or eventually buy the product). Advertisers want that but they also want to reach diverse audience. In this process, advertisers and platforms are not on the same page regarding which consumers to target. Ad platforms want to target those consumers who are very likely to click (or purchase). An example would be to give a deal on pizza to someone who is already standing in the queue to buy it. The advertisers, on the other hand want to reach out to people who are not their obvious customers.
So, whether the advertisement through ad platform generates Return on Investment (ROI)
is a critical question since ad platforms (like Google and Facebook) want to convince the
advertisers that they generate significant ROI. In a randomized experiment, the study found that ad platforms serve ads to consumers who already are predisposed to buy. Hence spending money to target them through ad platforms is not beneficial for the firms. The study, by researchers at Carnegie Mellon University (CMU), EPOS Group, NOVA University Lisbon, and Erasmus University, appears in Management Science.
"Our unique dataset allowed us to both causally identify the effectiveness of advertising
and estimate the degree of incentive misalignments between firm and ad platform," explains
Rahul Telang, professor of information systems and management at CMU’s Heinz College, who coauthored the study.
The study targeted more than 208,000 individual consumers in late spring 2016 in a display retargeting campaign. Researchers partnered with a European e-retailer that sells a range of products (e.g., home appliances, laptops, smartphones, cameras), mostly online. The study randomly allocated the consumers to receive either retargeting ads from the e-retailer (based on consumers' browsing behavior on the firm's website and products on external sites) or public service announcements (PSAs). Based on the data, the study estimated the difference in ad effectiveness for consumers who were more or less aggressively retargeted by the ad platform.
In accordance with the contracted incentives, the ad platform targeted consumers were
more likely to make purchases. Although advertising generally increases consumers’ likelihood of buying, the study did not find evidence that consumers who were more aggressively targeted by the ad platform were more receptive to the ads. Rather, the ad platform targeted consumers with a higher probability of buying, not consumers who were more receptive to ads. This rendered the ad platform’s targeting strategy unsatisfactory for the firm.
Based on further analysis, the authors suggest that the ad platform's bidding optimization
was not only misaligned but also led to a loss in profit for the firm. To remedy the incentive
misalignment, the authors propose a solution in which the firm restricts the ad platform to target only those consumers who are likely to click on the ad based on individual consumer-level estimates for baseline purchase probability and ad effectiveness.
“Our findings have significant practical implications for the digital advertising industry,"
says Thomas W. Frick, a marketing analyst with EPOS Group in Denmark, who led the study. "Although these types of contracts have become very popular, they do not adequately serve firms' interests."
The study was funded by Vereniging Trustfonds (Erasmus University), the Science and
Technology Fund POR Lisbon and POR Norte.
Summarized from an article in Management Science, Incentive Misalignments in Programmatic Advertising: Evidence from a Randomized Field Experiment by Frick, TW (EPOS Group), Belo, R (NOVA University Lisbon and Erasmus University), and Telang, R (Carnegie Mellon University). Copyright 2022 The Authors. All rights reserved.
About Heinz College of Information Systems and Public Policy
The Heinz College of Information Systems and Public Policy is home to two internationally recognized graduate-level institutions at Carnegie Mellon University: the School of Information Systems and Management and the School of Public Policy and Management. This unique colocation combined with its expertise in analytics set Heinz College apart in the areas of cybersecurity, health care, the future of work, smart cities, and arts & entertainment. In 2016, INFORMS named Heinz College the #1 academic program for Analytics Education. For more information, please visit www.heinz.cmu.edu.