Comparing user behavior when targeted based on firm inferred interest vs. user stated interest

Abstract

There has been a lot of discussions on giving user more control over their private information. Our paper wants to show how practices that give user more control over privacy, could facilitate collection of self-stated interest from voluntary users. This might potentially help firms to better understand users' true interests. The paper investigates the short run and long run impact of adopting the user stated interest into ad targeting, compared with targeting using firm inferred user interest. It also explores the side effects on user's overall app engagement and the possible mechanisms that associate users' action of stating interest with impact on their interest in ads and organic contents. Results show that on average users get more interested in ads immediately after the ad targeting adopts user stated interest categories. In the long run, however, not only does the effect diminish quickly, but there is also a lasting and significantly negative effect on users' overall app engagement, possibly because both organic content and ad targeting use the same user profile.