Advice doesn’t reach them
Agricultural extension officers are few and far between. Many farmers never see one in a season.
For smallholder farmers, on any phone
Know what to plant before you plant it.
Snow reads the soil under a farmer’s field and the weather coming this season, then sends one clear recommendation by text message. No sensors. No smartphone. No guesswork.
SnowSMS 3434
PLANT Machakos
The weather is expected to be very dry this season. We recommend planting sorghum instead of maize because it is more drought-resistant.
Every harvest is decided months earlier, by one choice: what to plant.
One decision a season. Get it wrong and the whole harvest is gone.
A smallholder farmer chooses what to plant once, before the rains. Most choose what they planted last year, or what a neighbour plants. But the climate is no longer behaving like last year, and there is no second chance until next season.
- Before plantingFarmer picks a crop, usually the same as last year.
- Right crop for the seasonHarvest comes in. Food on the table and some to sell.
- Wrong crop for the seasonA dry spell hits maize at flowering. Seed, fertiliser and months of work are lost.
Precision farming is priced for big farms
Soil sensors, drones and farm-management apps assume capital, a smartphone and a data plan.
Weather apps don’t make the decision
A rain forecast for the next week doesn’t tell anyone whether sorghum or maize will survive the season.
From a place name to a planting decision in seconds.
-
1
Farmer sends their location
A text with the village or district name, or a pin dropped on the web app. That’s all the farmer has to do.
-
2
Snow looks up soil and weather
Soil acidity, clay and sand content come from global open soil maps. Rainfall and temperature come from years of weather records and the latest forecast.
-
3
A model scores every local crop
Each crop is checked against what it needs: water over the season, temperature range, soil type. Drought-tolerant crops score higher when the season is forecast dry.
-
4
The best crop and a backup are chosen
Plus a rotation tip for next season, so a field of cereals is followed by beans or groundnuts that put nitrogen back in the soil.
-
5
The farmer gets one clear message
Plain words, local language, under 160 characters. It works on a basic phone with no internet.
Try it on a real place.
This working prototype pulls live data: weather history and the 16-day forecast from Open-Meteo, and topsoil properties from ISRIC SoilGrids. The crop rules are simplified for the demo.
SnowSMS 3434
Choose a place and the farmer’s text message will appear here.
Four Sustainable Development Goals, one better decision.
The impact comes from changing a choice made before planting, not from fixing a failed harvest afterwards.
No poverty
For a family farm, one lost harvest can mean selling livestock or taking on debt. Choosing a crop that survives the season protects income and keeps families out of that cycle.
Zero hunger
Smallholders grow much of the food eaten in their own regions. More reliable harvests mean more food at home and in local markets.
Climate action
Adaptation that a farmer can act on this season: when droughts and shifting rains arrive, the advice shifts with them, toward crops that can cope.
Life on land
Rotation tips after every season put nitrogen back with legumes and avoid draining the same nutrients year after year, keeping soil productive.
Cheap to run, so it can be cheap to use.
The expensive parts of precision farming are hardware and field visits. Snow has neither. The data is open, the model runs in the cloud for fractions of a cent, and the main running cost per farmer is a few text messages a season.
Who pays
- Farmer cooperatives and NGOs
- Pay per member per season to give every farmer recommendations. Cheaper than sending agronomists to each field.
- Seed and input companies
- Sponsor the service in their regions. Farmers who plant the right crop buy the right seed, and come back next season.
- Crop insurers and microfinance
- Lower-risk planting decisions mean fewer defaulted loans and claims, so they have a reason to pay for farmers’ access.
- Farmers, for extras
- The first recommendation each season stays free. Mid-season alerts (“dry spell coming, weed now to save water”) are a small mobile-money subscription.
Main costs
- SMS gateway fees, paid per message
- Cloud hosting for the model and database
- Local crop and language tables for each new country
- Field partners and marketing to reach first users
How farmers find it
- Through cooperatives and agro-dealers they already trust
- Short code on posters at markets and seed shops
- Community radio, which reaches farmers without phones online
- Extension officers sharing it, multiplying their reach
Why it scales
Adding a farmer costs a few messages. Adding a country means adding its crops and languages, not building new technology. The same open soil and weather data covers the whole planet.
How it compares
| Option | No equipment to buy | Works on a basic phone | Says what to plant | Reaches every farmer |
|---|---|---|---|---|
| Snow | Yes | Yes | Yes | Yes |
| Soil sensors and precision kits | No | No | Yes | No |
| Weather apps | Yes | No | No | Partly |
| Extension officer visits | Yes | Yes | Yes | No |
Built by students, with tools that already exist.
Free weather and soil data, a simple scoring model and an SMS gateway. Nothing has to be invented from scratch, which is why the demo above already works.
-
Final weekend, day 1
Data and model
Connect weather and soil APIs, build crop requirement tables, score and rank crops. Train a simple model on historical yields where data exists.
-
Final weekend, day 2
SMS and web
Hook the model to an SMS gateway so a text in returns a recommendation out. Polish this web app for partners.
-
First season
Pilot with one cooperative
Compare harvests of farmers who followed the advice with those who didn’t, and tune the model with what we learn.
-
After that
More crops, languages and channels
USSD menus and voice messages for farmers who can’t read, and new regions added one crop table at a time.
An affordable AI tool that helps farmers choose what to plant, based on their local soil and predicted weather.
Try the demo