Each plot's own soil test every season, sold through FPOs with a dose plan
TL;DR Soilround tests the soil on each member's plot before every main season, works out a fertiliser plan in bags of named products, and rechecks a share of plots mid-season. Farmer producer organisations sell it to members alongside their input shop. The aim is better yield per rupee of fertiliser, not a promise to use less.
India's Soil Health Card scheme is huge, with more than 27 million tests run in 2017–2019 alone[1]. But it samples on a grid. The target is one sample per 2.5 hectares of irrigated land and one per 10 hectares of rainfed land, every two years[2]. So a smallholder's card can easily describe a neighbour's field, measured two seasons ago.
The data is patchy too.
In Andhra Pradesh, 2.4 million records came from only about 1.6 million distinct locations, with duplicates and coordinates sitting in the wrong district[2]. Nobody knows the quality of the labs behind the national figures[1]. So farmers fall back on a standard dose or whatever the shopkeeper says.
What it does, in regulatory terms: soil testing and fertiliser advice only. Soilround does not make or sell fertiliser. The FPO's input shop does that, under its own dealer authorisation.
Before kharif and rabi, a trained village sampler visits each enrolled plot, takes a GPS-tagged composite sample and sends it to a partner lab. The result turns into a dose plan for that plot and crop, written in bags of products the FPO shop stocks, with timing for each split. The method follows soil-test crop response equations, which researchers calibrate for each crop and region[3]. Mid-season, a tenth of plots get retested and checked against crop condition, so the plan keeps learning.
The original idea was continuous in-ground soil sensors. Moisture probes work fine, but in-field nutrient sensors still need regular calibration, and the one review I found reported an 8.47% average error against lab instruments from just four sites[4]. Soil nutrients move slowly anyway. Testing each plot every season gives farmers most of the value for a fraction of the cost.
The farmer uses it but the FPO pays, buying tests in bulk and recovering part from members as a small fee or folding it into input sales. FPOs win because members who see better yields stay loyal to the FPO shop. Seed or input companies running demonstration plots are a second buyer.
The price per test and the farmer's share are assumptions to test, and you're up against a free government card that most farmers already ignore.
The evidence for plot-level dosing keeps piling up. A 2026 field study in Karnataka found a soil-test-based dose gave 68.85 quintals of aerobic rice per hectare against 53.25 with the general recommendation, with a benefit-cost ratio of 1.61 against 1.09[3]. The national natural farming mission's dashboard also tracks soil health cards[5], which tells you soil data now sits inside newer farm schemes, even if it's still collected on a grid.
Nobody publishes a figure for paid soil testing in India. The government spends heavily on the free version - researchers put the Soil Health Card scheme's cost at about US$85 M (about ₹820 Cr)[2], and it ran 27 million tests in three years[1].
So the need is accepted. Whether farmers or FPOs will pay for a better version is the open question.
The Soil Health Card scheme is the default and it's free[2]. Farm-sensor startups sell soil moisture units, soil temperature sensors and fertigation control, mostly to horticulture growers with drip[6]. Krishi Vigyan Kendras and state labs test soil on request.
I couldn't verify a private firm that retests every plot each season through FPOs, so check before you assume the gap is open.
| Difficulty | Moderate (41/100) |
| Starting capital | ₹45–84 L |
| Time to launch | 4–6 months |
The science is published. The hard parts are logistics and trust. You need thousands of samples collected properly in a three-week window before sowing, a lab that turns them round in days and dose plans that farmers actually follow when the shopkeeper says otherwise.
I couldn't find a licence needed for a private soil-testing service, but confirm with the state agriculture department [Requires validation].
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't let AI change fertiliser doses on its own. Dose plans come from published crop-response calibrations and an agronomist signs them.
Want to work out the AI layer for Soilround? Talk to Yash →
Here's what can go south -
I like this for a founder from an FPO network, a soil lab or input retail who can run seasonal field teams. If you're hoping to sell sensors then skip it.
Prove two things first - that 500 members of two FPOs will pay something for a test and follow the plan, and that their yield or cost per quintal actually moves.
Plan for roughly ₹45–84 L for field teams, sampling kit, lab partnerships and the software that turns results into dose plans.
I couldn't find one for a private service, but use an accredited lab and confirm with your state agriculture department. Fertiliser sales need a separate dealer authorisation.
Yes, wherever small farms share inputs through cooperatives. Check local lab standards and data protection rules first.
The free Soilround validation kit has the line-by-line capital breakdown, the effort scoring, the lab and data-protection checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for FPOs and farmers, and 5 assumptions to test. Download the kit →
Thinking about building Soilround or something close to it? The make-or-break work is getting samples tested and plans to farmers before sowing starts, and I help founders plan exactly that kind of crunch. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.