Making Use of Derived Personality: The Case of Social Media Ad Targeting

Abstract

People’s social media text has been shown to have limited prediction power of their personality. In this work, we conducted a survey study and a field study to explore the feasibility of using predicted personality traits derived from socialmedia text for the purpose of ad targeting. In the survey study, we measured people's personalities and their responses to an advertisement tweet. We found that people withhigh openness and low neuroticism responded more favorably to a targeted advertisement, thus demonstrating the effectsof the personality traits themselves. In the field study,we sent the advertisement tweets to real-world Twitter users,and found the same effects on users' responses using personality traits derived from users' tweet text. These results suggest that the derived personality traits had the same effects as the personality traits measured by traditional personality questionnaires, and can indeed improve ad targeting in real-world settings.