Unpacking Perceptions of Data-Driven Inferences Underlying Online Targeting and Personalization

Abstract

Much of what a user sees browsing the internet, from ads to search results, is targeted or personalized by algorithms that have made inferences about that user. Prior work has documented that users find such targeting simultaneously useful and creepy. We begin unpacking these conflicted feelings through two online studies. In the first study, 306 participants saw one of ten explanations for why they received an ad, reflecting prevalent methods of targeting based on demographics, interests, and other factors. The type of interest-based targeting described in the explanation affected participants' comfort with the targeting and perceptions of its usefulness. We conducted a follow-up study in which 237 participants saw ten interests companies might infer. Both the sensitivity of the interest category and participants' actual interest in that topic significantly impacted their attitudes toward inferencing. Our results inform the design of transparency tools.

Oh, the Places You've Been! User Reactions to Longitudinal Transparency About Third-Party Web Tracking and Inferencing

“…Users have a wide variety of reactions to web tracking related to targeted ads. These reactions range from positive to negative [79,90] and from comfortable to creepy [23,58,91]. Users often evince a tension between a desire to see relevant ads and their concerns about invasiveness [23,91].…”

“…These reactions range from positive to negative [79,90] and from comfortable to creepy [23,58,91]. Users often evince a tension between a desire to see relevant ads and their concerns about invasiveness [23,91]. Further, studies disagree on when users are willing to share information with advertisers [25,52] or pay for privacy [17,26,49].…”

“…We quantified sensitivity using an IRB-approved MTurk study in which we asked participants about their comfort with a specific topic being inferred and used for personalization. We based our study on Dolin et al [23], abbreviating their script and expanding the scope of the study to cover all 1,124 categories used in Google AdWords (excluding world localities).…”

Quantity vs. Quality: Evaluating User Interest Profiles Using Ad Preference Managers

“…Dolin et al surveyed people to understand how ad explanations (small disclosures near advertisements that provide insight into how the ad was targeted) impact peoples' opinions of targeted advertising. They found that peoples' comfort level varied based on the explanation they were given for how the targeted interest was inferred [24]. They also report that the accuracy of inferred interests was strongly, positively correlated with user comfort, regardless of the sensitivity of the interest.…”

“…However, if user interest profiles are incorrect, this would suggest that (1) online ads are being mistargeted and the money spent on them may be wasted, and (2) that greater privacy for users would have less of an impact on engagement with ads than is feared by industry. Further, surveys have found that users are deeply concerned when advertisers draw incorrect inferences about them [24], a situation that investigative journalists have anecdotally found to be true in practice [10], [48].…”