Leaders of nonprofit organizations are increasingly under pressure to demonstrate the impacts of their work in addressing pressing societal problems such as global poverty. This paper examines the growing demands for impact and their implications for performance measurement in nonprofits. We begin with a review of the debates around performance and impact, drawing on three literatures: strategic philanthropy, nonprofit management, and international development. We then articulate two key dilemmas facing nonprofit leaders in measuring their social impact: where to position their measurement on a logic chain, and how to address multiple and competing accountability demands. Drawing on this analysis, we develop a contingency framework for measuring results, arguing that it is neither feasible nor desirable for all organizations to measure impact. Some may be better off measuring their activities and outputs rather than outcomes or impacts. The contingency framework may be helpful to professionals in the social sector in critically differentiating among the types of results they can and should measure. In closing, we discuss the implications of our analysis for future research on performance management and control.