Despite a rapidly growing enthusiasm around applications of information and communications technologies (ICTs) to smallholder agriculture in developing countries, there are still many questions on the effectiveness of ICT-based approaches. This study assesses the effects of videomediated agricultural extension service provision on farmers’ knowledge and adoption of improved agricultural technologies and practices in Ethiopia. The study focuses on a program piloted by the Government of Ethiopia and Digital Green and poses three questions. First, to what extent does video-mediated extension lead to increased uptake of improved agricultural technologies and practices by smallholder farmers? Second, is video-mediated extension targeted at both spouses of the household more effective than when only targeted at the (typically male) household head? Third, how cost-effective is a video-mediated approach to extension provision? The study explores these questions with a randomized controlled trial designed to evaluate the video-mediated approach as applied to three priority crops (teff, wheat, maize) and three technologies (row planting, precise seeding rates, and urea dressing). The trial was implemented in 347 kebeles (village clusters) during the 2017 meher (rainy) season in Ethiopia’s four most agriculturally important regional states. Analysis of data from our surveys of 2,422 households and 896 extension agents indicates that the video-mediated approach is more effective than the conventional approach in achieving several key outcomes. Specifically, we find that videomediated extension reaches a wider audience than the conventional approach and leads to higher levels of agricultural knowledge and uptake of technologies in those kebeles randomly assigned to the program. While our results do point to greater participation and greater knowledge of female spouses in kebeles where both male and female spouses were targeted by the program, we do not find clear evidence that the more inclusive approach translated into higher uptake of the subject technologies and practices. Finally, we find that the video-mediated approach becomes less costly as the scale of operation increases.< Réduire