What Cookie-Based Advertising Effectiveness Fails to Measure.

This article investigates the challenges of measuring the effectiveness of retargeted advertising when randomized experiments are conducted at the cookie level rather than the individual consumer level. Using a field experiment with a medium-sized American retailer, the study links cookie-level data to first-party customer relationship management (CRM) data to analyze advertising impact across online and offline channels. The findings reveal that cookie-level analyses are flawed because individuals often have multiple cookies assigned to both treatment and control groups, and retargeted ads can induce consumers to generate additional cookies (“cookie propagation”), breaking the assumption of random assignment at the individual level. Individual-level analyses enabled by first-party data provide a more accurate assessment of retargeted advertising’s lift, including offline sales effects that cookie-based methods miss, and highlight the moderating role of consumer distance from physical stores. The study concludes that with increasing privacy regulations and the impending deprecation of cookies, focusing on consented individual-level data and privacy-preserving technologies will improve advertising measurement more than cookie-based approaches.