Teacher evaluation systems introduced by states and school systems in the past several years have focused attention on improving the performance of public school teachers, but they have been cost- and time-intensive, placing a significant burden on states' and districts' resources. In Tennessee, for example, trained evaluators conducted nearly 300,000 classroom observations during the 2011-2012 school year, prompting administrators to complain that "the amount of time spent to implement TEAM [the state's new system] was unmanageable." Even among school leaders who did not feel overburdened by the number of hours spent observing classrooms, many felt the time had not been used efficiently, as the system treated all teachers with the same intensity, despite the fact that teachers' skills and needs for support varied widely. To address these concerns, Tennessee and other systems have replaced their one-size-fits-all evaluation approaches with more differentiated models, using past performance data to determine which teachers should be evaluated with more or less intensity in subsequent evaluation cycles and, in some cases, what that attention should include. This brief explores differentiation strategies in nine districts, two charter management organizations, and three states, Tennessee, Delaware, and Ohio. Interviews with system leaders and analyses of teacher evaluation policies reveal that these systems now vary the format or frequency of formal evaluation cycles, the format or frequency of classroom observations, or the type of observer conducting classroom observations, based on what is known about teachers' needs, strengths, and goals. A list of sources is included.