In the last decade, efforts to systematically study the effectiveness of programs in developing countries have expanded dramatically. In this paper, Miguel Székely, director of the Institute for Innovation in Education at Tecnológico de Monterrey, shows how Mexico has improved the evidence base for public policy in a number of ways. He explains the difficulties of conducting good impact evaluations and assesses the interests of key stakeholders in promoting or opposing the creation and use of evidence. He draws out lessons from the government’s effort to evaluate a major antipoverty program (PROGRESA-Oportunidades), publish politically sensitive poverty data, introduce performance measurement in education, and institutionalize learning. He concludes with a proposal for how developing countries could systematically incorporate evidence in policymaking.