Exploring Understandable Algorithms to Suggest Fitness Tracker Goals that Foster Commitment

Abstract

While fitness trackers are gaining popularity, they struggle to offer long-term health benefits, largely due to their inability to offer engaging goals. Understanding how trackers can suggest and update fitness goals can lead to building improved systems that support wellbeing. We investigate how to suggest fitness tracker goals to users and ways to help them commit to those goals. We compared algorithms for step goal setting in a pre-study. Next, we conducted two surveys (a vignette study and a survey using the users' Fitbit data) that compared the users' attitudes to suggested goals, with and without disclosing the algorithm to them. We found that explaining how a step goal was computed increased goal commitment and, in one study, contributed to building trust in the goal. Our work shows that explaining how a tracker works can help build engaging fitness tracking experiences. We contribute insights on designing transparent personal informatics systems.