Abstract Many policies – on the delivery of inputs or on marketing systems, credit, or extension – influence the potential utilization of new technologies. Through ‘farm-based policy analysis’ it is possible to use data generated in on-farm research (OFR) to identify policy constraints to the use of new technologies, and to effectively communicate that information to policy makers. This paper describes a tentative framework for farm-based policy analysis and suggests a sequence of five steps for the analysis: (1) identify the policy-induced constraints; (2) determine the rationale behind the policy; (3) identify the decision makers to whom the results should be communicated; (4) identify solutions or policy options; and (5) communicate results to decision makers. A case of farm-based policy analysis from Haiti illustrates the concepts and methods described in the first part of the paper. On-farm experiments in Les Cayes, Haiti, confirmed a response to nitrogen in maize, but adoption of the recommended practice and consequent gains in productivity and income were constrained by the scarcity of urea in the local market. An analysis of local supply and potential local demand for urea and the potential benefits of urea application was conducted, and results communicated to two target groups of decision makers: representatives of the Ministry of Agriculture responsible for fertilizer distribution policy, and representatives of private fertilizer enterprises. These groups responded by making larger supplies of urea available to local farmers. Adoption of the fertilizer recommendation and urea sales increased. The case demonstrates the potential value of farm-based policy analysis building upon data from OFR.