My Little Robot: User Preferences in Game Agent Customization

Abstract

Agents are part of our daily lives. They are in the games we play and in our phones as personal assistants. As users develop relationships with these agents, one question is about how to cultivate those relationships. This study chose one design feature, customization, to understand whether it would affect user experiences and whether there are optimal number of choices users prefer. This paper describes the results of two studies. The first, a small scale (N=15), preliminary study explored user preferences when customizing an on-screen, robot-like, virtual agent partner. The data includes eye-tracking, screen capture, and a post-survey. Findings from this study suggests that a robot's clothes (both top and bottom) are the most important for users to customize. Informed by the first study, a custom application was developed to test user preferences for the number of choices and attributes in a large scale online study (N = 343). Findings from the second study suggest that participants preferred at least 8 different attributes to customize and they showed no decrease in satisfaction with the most number of choices (i.e., 24) per attribute. These findings are different from previous studies on choice overload and contribute to further understanding of choices in games and digital spaces.