Public service organizations periodically collect and disseminate performance information that enables frontline employees to act based on two aspects of performance: current performance (how is the client performing right now?) and performance progression (is the client performing better, similarly, or worse than previously?). Yet knowledge of how frontline employees use performance information about their clients’ performance progression remains limited. Building on cognitive psychology and street-level bureaucracy research, this article theorizes and tests how information on changes in client performance over time affects frontline employees’ performance information use. We develop a theoretical framework that comprises three competing hypotheses on how performance progression information (on performance improvement, performance stability, and performance deterioration) shapes purposeful performance information use at the frontlines of public services delivery. Each hypothesis relates to a distinct cognitive bias: needed-deservingness bias, negativity bias, and change-react bias. Using a pre-registered survey experiment among Danish public school teachers (n = 925), we find support for the change-react bias. Teachers are more inclined to take behavioral action when presented with performance progression information showing change (either improvement or deterioration) relative to no change (stability) in student performance. These results expand our understanding of performance information use at the frontlines by suggesting that frontline employees process performance information showing change different from performance information showing stability.