the female farmer the app forgot

the female farmer the app forgot

Problem statement

Women make up a large share of the agricultural labor force in low-income countries, yet the digital advisory services meant to raise smallholder yields — SMS/IVR crop advice, market-price alerts, input recommendations, climate information — reach mostly men. In Africa only about a quarter of registered users of digital agriculture services are women, and field usage is far more skewed than even that registration figure suggests. The services exist and demonstrably help farmers; the problem is that the women who do much of the planting, weeding, and harvesting are structurally excluded from the information loop, so the productivity and resilience gains flow to men and the gender gap in farm output widens rather than closes.

Why this matters

Closing the gender gap in agricultural productivity is one of the highest-leverage interventions in development — when women farmers get the same access to inputs and information as men, yields and household nutrition rise. Digital advisory is the cheapest way to deliver that information at scale, which is exactly why its failure to reach women matters: a low-cost tool that could narrow the gap is instead reinforcing it. The exclusion is measurable and stark — in a documented mobile-agriculture service in India (IKSL), only ~13% of users and ~2% of active users were women — meaning the women doing the farming are almost entirely absent from the channel built to advise farmers.

What’s been tried and why it hasn’t worked

The first wave of digital agriculture assumed a gender-neutral "farmer" and pushed services through whoever held the household phone — usually a man. That failed women on several stacked constraints. Mobile access: women who work in agriculture are 5–40% less likely to own a phone than men (cost of ownership, plus cultural norms discouraging female ownership of productive assets). Relevance: services designed without women's specific crops, decisions, and time constraints feel irrelevant, so even women who can access them don't use them. Trust and safety: registration and input delivery run through predominantly male agents and clerks, which many women find uncomfortable or are socially discouraged from approaching. Measurement blindness: providers rarely collect gender-disaggregated usage data, so they cannot even see the gap, let alone manage it. Treating the problem as a pure connectivity issue (give her a phone) misses that relevance, trust, and the gender of the delivery channel matter as much as device access — a classic case of ignoring the social context of the intended user.

What would unlock progress

The unlock is designing the service and its delivery channel around women's actual constraints: content tied to the crops and decisions women manage, delivered through channels they can reach (shared phones, female agents, IVR for low-literacy users, women's groups), with gender-disaggregated metrics so providers can see and close the gap. GSMA's own evidence that female sales/extension agents dramatically improve women's uptake points to a concrete, replicable mechanism. Adjacent precedent exists in microfinance and in community health, where deploying female frontline workers and women's peer groups overcame the same access-and-trust barriers.

Entry points for student teams

A team could run the drop-out diagnostic without going to the field, mapping where women fall out of the advisory chain (no phone, wrong content, male agent, wrong timing) from the World Bank's LSMS-ISA agricultural household surveys — which record plot-manager gender, phone ownership, and extension contact for several African countries, downloadable from the Microdata Library on a free account — and turning that into a redesign spec, with interviews of women's-group organizers and diaspora farmers where the survey record runs out. A design team could prototype a gender-responsive advisory format (e.g., IVR voice advice on a woman-managed crop, delivered via a women's savings group) and specify the A/B test rather than run it, since splitting a live provider's user base requires a commercial agreement, in-country ethics approval, and months to read out: the deliverable is the working prototype, a comprehension-and-preference test with a reachable women's farming or community-gardening group, and a pre-registered evaluation design (arms, uptake outcome, disaggregation, power) handed to the provider that owns the user base. A data team could build the gender-disaggregated usage dashboard as a specification — metric definitions, disaggregation cuts, and alert thresholds — demonstrated on Global Findex indicators and simulated service logs, since providers' own usage records are proprietary and are exactly the data this brief says they do not collect. Relevant disciplines: human-centered design, agricultural extension, gender studies, ICT4D, data analytics.