There has long been a tendency among policy researchers in food and agriculture to view ministries of finance and agriculture as their main, if not only, constituencies. This paper argues that greater insight that links the outcomes of research to real policy problems and processes makes for better scientific inquiry. The author contends that policy research needs to recognize multiple constituencies from the onset of the research process. This paper discusses requirements for effective policy research incorporating broad participation and ownership of the research process and presents this approach in the context of IFPRI's research on urban farming in Uganda and food insecurity awareness generation in Ghana. The benefits and implications for using this approach are highlighted.