The objective of the study was to identify a viable trade-off between low data requirements and useful household-specific prioritizations of advisory messages. At three sites in Ethiopia, Kenya, and Tanzania independently, we collected experimental preference rankings from smallholder farmers for receiving information about different agricultural and livelihood practices. At each site, was identified socio-economic household variables that improved model-based predictions of individual farmers’ information preferences.

This policy brief shows how digital tools can help to ensure that public money for agricul-tural extension is spent wisely. Governments often fund offices, training centers, and the salaries of extension officers, but cannot eas-ily review the impacts of these expenditures. This is because the activities of extension agents are not monitored systematically. Ex-ension services rarely generate quantitative data on the effects of their work.