Consumers’ Perceived Privacy Violations in Online Advertising

In response to privacy concerns about personal data collection and use, the online advertising industry has developed privacy-enhancing technologies (PETs), of which Google's Privacy Sandbox is a prominent example. In this research, we apply dual-privacy theory, which postulates consumers have intrinsic and instrumental preferences for privacy, to understand perceived privacy violations (PPVs) for current practices and proposals. The key idea is that different practices and proposals differ in whether individual data leaves the consumer's machine or not and in how they track and target consumers; these affect, respectively, the intrinsic and instrumental components of privacy preferences differently, leading to different PPVs for different practices. We conducted online studies with U.S. consumers to elicit PPVs for various advertising practices. Our findings confirm the intuition that tracking and targeting consumers under the industry status quo of behavioral targeting results in high PPVs. While new technologies that keep data on users' devices reduce PPV compared to behavioral targeting, the reduction is minimal. Group-level targeting does not significantly reduce PPV compared to individual-level targeting. However, contextual targeting, which involves no tracking, significantly lowers PPV. Notably, when tracking is absent, consumers show similar preferences for seeing untargeted ads and no ads. Our results indicate that consumer perceptions of privacy violations may differ from technical definitions. A consumer-centric approach, based on, for instance, the dual-privacy theory, is essential for understanding privacy concerns. At a time of significant privacy-related developments, these insights are crucial for industry practitioners and policymakers.