Multi-site interventions can be challenging to evaluate, particularly when implementers are working toward localized outcomes. The diversity of intended outcomes makes it difficult to assess the overall effectiveness of the intervention. It also complicates stakeholders’ ability to identify drivers of implementation success. This is particularly true in the area of community development, where millions of dollars are invested annually in locally driven community development (LD-CD) interventions (i.e., national programs that are customized to community needs and implemented locally). This paper describes the challenges of evaluating these localized programs, and discusses practical methods for defining and measuring grantees’ relative success in this context. It concludes by discussing the implications of these methods for future evaluations in which funders allow grantees the freedom to implement program models in flexible ways driven by local needs.