In their recent paper ‘On-farm trials for development impact? The organization of research and the scaling of agricultural technologies’, de Roo, Andersson and Krupnik report on three case studies, each undertaken by one of the authors, of projects conducting on-farm research. They reach conclusions on the limitations of the projects themselves and the effects of ‘donor dependency’, and propose a strategy to overcome these issues. However, the description of the philosophy, strategies and conduct of the projects reviewed in the southern African case study is incomplete and misleading, and shows that the case study author did not understand or overlooked important project components. Due to this the conclusions reached, insofar as this case study is concerned, are largely either invalid or already contemplated in the project activities. Here, we describe more fully the philosophy and strategies followed by the series of projects on which the case study was conducted, which were designed to facilitate, through the upscaling of project methodologies, the eventual outscaling and widespread adoption of more sustainable farming systems by smallholder farmers in eastern and southern Africa. We propose these methodologies as a valid comprehensive approach to the organization of agricultural research for development for the successful development, scaling-up and scaling-out of agricultural technologies.