People often set goals for events that contain uncertainty which will be (partly) resolved when time passes. When part of the uncertainty is resolved, initial goals may have turned out to be unrealistic, and hence these goals no longer work as a motivation device. We study goal setting of university students in courses where uncertainty about students’ ability and several aspects of the course is reduced over time. We examine whether students should be explicitly invited to revise their goals over time, or whether goals should be set rigid. We test this by means of a field experiment involving almost 2200 students. Students receive a series of three surveys. Surveys differ in i) whether students are motivated to set goals and ii) if and when they are asked whether they want to revise any of their goals. We find that only motivating students to set goals has no effect on performance as compared to the control group, but motivating students to set goals and explicitly giving them the opportunity to revise their goals from the start has negative effects. We then explore the mechanisms that drive these results.