We investigate the causal effect of conversations about performance and performance pay implementing a 2x2 field experiment in a retail chain. In the performance pay treatments, managers receive a bonus for profit increases. In the performance review treatments, managers have regular meetings with their supervisors discussing their activities to increase profits. We find that review conversations raise profits by 7%-8%. However, when additionally receiving performance pay this effect vanishes. Analyzing an extension of Bénabou and Tirole (2006), we rationalize this effect formally and provide empirical evidence that the use of performance pay changes the nature of conversations undermining their value.