Digital decision support tools for smallholder agriculture

Overview

Smallholder farms often face large yield gaps because farmers lack the information needed to allocate inputs efficiently. Expanding smartphone access creates an opportunity to close those gaps through digital decision support tools that deliver customized fertilizer and agronomic advice while simultaneously generating field data to improve future recommendations. Early field results from CIMMYT show that these tools can raise gross marginal returns on fertilizer investments by up to 20% and lift crop yields by 1.2% to 22%, depending on the crop and tool used. At scale, the approach also carries clear potential for climate change mitigation, making it attractive for partners focused on sustainable agriculture, food security, and climate adaptation.

Technical specifications

• Smartphone-based delivery of site-specific fertilizer and agronomic recommendations to smallholder farmers
• Customized advisory logic that goes beyond blanket recommendations by integrating soil test information and farmer-specific context
• Field data capture from farmer interactions, creating a feedback loop to refine future recommendations
• Validated through repeated on-farm performance trials comparing four treatments: zero control, conventional blanket recommendation, strict soil test-based nutrient recommendation, and digitally assisted customized recommendation
• Statistical evaluation using generalized linear mixed models that control for farmer-dependent covariates, repeated across locations, seasons, and socio-economic settings
• Application areas include soil fertility management, input optimization, and yield gap reduction in smallholder maize and other cropping systems

Technology readiness level

The digital decision support tools have been validated in repeated on-farm trials across multiple locations, seasons, and farmer socio-economic contexts, with consistently superior performance compared to conventional and soil-test-only approaches. Despite this proven field performance, adoption at scale remains low. Current limitations include limited coordination among technology developers, agronomists, socio-economists, behavior experts, user-experience designers, and scaling partners, as well as the fact that many existing digital tools were calibrated under data-scarce conditions. A one-year exploratory phase is planned with technology partners to address tool-, technology-, and user-related bottlenecks and prepare the solution for broader scaling.