Process evaluation was used to examine the implementation of a randomized, controlled trial of an education intervention that improved infant growth in Trujillo, Peru. Health personnel delivered the multi-component intervention as part of usual care in the government health centers. Quantitative and qualitative methods were used to examine process indicators, which included the extent of delivery (dose), fidelity to intervention protocol, barriers to implementation and context. Results demonstrated that most intervention components were delivered at a level of 50-90% of expectations. Fidelity to intervention protocol, where measured, was lower (28-70% of expectations). However, when compared with existing nutrition education, as represented by the control centers, significant improvements were demonstrated. This included both improved delivery of existing educational activities as well as delivery of new intervention components to strengthen overall nutrition education. Barriers to, and facilitators of, implementation were explored with health personnel and helped to explain results. This study demonstrates the importance of examining actual versus planned implementation in order to improve our understanding of how interventions succeed. The information gained from this study will inform future evaluation designs, and lead to the development and implementation of more effective intervention programs for child health. © The Author 2006. Published by Oxford University Press. All rights reserved.