Firms use subjective performance evaluations to provide employees with both incentives and feedback. This article shows that if an objective measure of performance, however imperfect, is available, subjective evaluations with incentive effects can be sustained even without repeated interaction. Although full efficiency cannot be achieved in general, it is achievable if the firm can commit to a forced distribution of evaluations and employs a continuum of workers. When the number of workers is small, a forced distribution is useful only if the objective measure is poor. The model also shows that a leniency bias in evaluations can improve incentives.