Most recent research on teacher evaluation examines evaluation's measurement properties and accountability uses. Less research studies how evaluation data can improve teaching and student learning. In other contexts, researchers have examined how teachers use data to improve their practice. From general research on teachers' data use, we apply the data-driven decision-making (DDDM) framework to synthesize research on teacher evaluation since 2009. We illustrate how evaluation data are collected, analyzed, and synthesized to inform instruction and improve student learning. Most research focused on teachers' use of observation data, not their use of student data. We find that the teachers' use of evaluation data involves more social learning than the DDDM model implies. The effects of observation on instruction and student learning are often weak, apparently because observers lack the time and knowledge to support teachers' thorough analysis and synthesis of evaluation data. Implications for policy and further research are offered.