Does agricultural input aid always lead to favourable food security outcomes? This paper describes Zimbabwe's agricultural recovery program for the 2003/2004 farming season and draws some lessons that can be used in the designing and implementation of future programs. Input aid was found to be most beneficial if it is packaged together with other backup services such as training in soil fertility and water management, development of lucrative markets, and provision of basic infrastructure. Complementarity among the components of the input package itself was also found to affect the productivity of the inputs. Poor targeting of inputs both to areas and to farmers was found to affect the effectiveness of input aid in promoting food security. The study concluded that donor organisations should work closely with the recipient communities in the design and implementation of input aid programs. (Contains 5 tables.)