Two common incentives for participating in exercise programs are cash rewards for meeting goals and the loss of deposited money when goals are missed. Direct cash rewards lead to higher enrollment, but the risk of losing money is a stronger motivator for sticking with a program. We conducted an experiment using loss protection to leverage the power of both approaches. Participants were offered two exercise classes a week for 12 weeks. Anyone who attended the first weekly class received a chance to play a lottery that was very likely to pay a cash reward, but they also faced a low risk of not winning any money. Participants in the loss-protection group could insure against the loss by also attending the second class of the week. Participants in the control group could earn the equivalent money by likewise attending the second class, but the incentive was a straight reward for class participation (a flat payment), not as loss protection. For any weekly pattern of attendance, expected earnings were the same in both groups. We randomly assigned 153 participants to either the loss-protection or the control group. The loss-protection framing resulted in greater exercise class attendance, suggesting that the approach could enhance the outcomes of reward-based programs without increasing program costs.