OBJECTIVES: Diarrhoea is a leading cause of infant mortality with the main transmission pathways being unsafe water and contaminated food, surfaces and hands. The 'Safe Start' trial evaluated a food hygiene intervention implemented in a peri-urban settlement of Kisumu, Kenya, with the aim of reducing diarrhoeagenic enteric infections among infants. Four food hygiene behaviours were targeted: handwashing with soap before preparation and feeding, boiling infant food before feeding, storing infant food in sealed containers, and exclusive use of designated utensils during feeding. METHODS: A process evaluation of the intervention was guided by a theory of change describing the hypothesised implementation and receipt of the intervention, mechanisms of change, and the context. These were assessed by qualitative and quantitative data that included debriefing sessions with the delivery teams and Community Health Volunteers (CHVs), and structured observations during food preparation. RESULTS: The intervention achieved high coverage and fidelity with over 90% of 814 eligible caregivers participating in the intervention. Caregivers in the intervention arm demonstrated an understanding of the intervention messages, and had 1.38 (95% CI: 1.02-1.87) times the odds of washing hands before food preparation and 3.5 (95% CI: 1.91-6.56) times the odds of using a feeding utensil compared to caregivers in the control group. Contextual factors, especially the movement of caregivers within and outside the study area and time constraints faced by caregivers influenced uptake of some intervention behaviours. CONCLUSION: Future interventions should seek to explicitly target contextual factors such as secondary caregivers and promote food hygiene interventions as independent of each other.