A farmer photographs a diseased leaf and receives a diagnosis. Somewhere, that image, the identification, the location and the date are stored.
Who owns that record?
The question is usually answered in a document nobody reads, and it is treated as a legal formality. It is not. It determines who can build on the record, who benefits from what it becomes, and — most consequentially — whether farmers continue generating it once they understand where it goes.
Why this is not an abstract question
Agricultural data has a specific commercial property: individually it is nearly worthless, and in aggregate it is valuable.
One farm's disease history tells you very little. Ten thousand farms' histories, located and dated across several seasons, is a dataset that supports advisory products, input demand forecasting, insurance pricing, credit assessment and research.
That gap between individual and aggregate value is exactly where the ownership question bites. A farmer trading something of negligible individual value for a free service has made a reasonable exchange. The aggregate that results is a different asset, built from thousands of those exchanges, and the terms under which it was assembled usually said nothing about it.
The forms this takes in practice
Platform ownership. A service collects data as a condition of use, aggregates it, and builds products from it. The farmer receives the service; the platform holds the asset. This is the default in most agricultural technology.
Institutional collection. Government schemes, banks and insurers collect farm data for a stated purpose. Subsequent use for other purposes is common and rarely disclosed to the farm.
Buyer-side collection. A processor or trader records what a farm produced, when and at what quality. This is legitimate operational data and it is also a detailed picture of the farm's position — and it sits with the counterparty in the price negotiation.
Collective ownership. The FPO holds the record on behalf of members. Less common, and structurally the closest to the people who generated it.
The extraction pattern, and why it eventually fails
The pattern that recurs: data is collected from farms cheaply, aggregated, and sold back to the same farms as advisory — or sold to third parties whose interests are not the farm's.
This has an obvious fairness problem and a more interesting practical one. It is unstable.
Farmers are not indifferent to this. Where a community concludes that a data collection exercise serves someone else, participation degrades — reports stop being filed, records become approximate, the enthusiastic early users go quiet. The dataset stops being an asset because the thing that made it valuable was continuous voluntary contribution.
Extraction, in other words, destroys the resource it depends on. Which means the ownership question is not only ethical; it is a design constraint on anything intended to last.
What a defensible position requires
Four commitments, and they are testable rather than aspirational.
The record belongs to the farm and the FPO. Not as a sentiment — as the stated position, with everything else framed as permission granted rather than ownership transferred.
It is exportable. The group can take its data out, in a usable form, at any time, without asking. This is the concrete test of every ownership claim: an owner who cannot leave with their property does not own it. The same principle governs business systems generally, in security and data ownership in client systems.
Aggregation is disclosed and permissioned. If regional patterns are derived from many farms' records, the contributing groups know it, and it happens with their agreement.
Benefit returns to the contributors. The group that generated the record gets the seasonal analysis, the timing patterns and the recurrence data — not just the individual diagnoses. That return is what makes the exchange visible, and it is the practical use described in seasonal planning from diagnosis data.
Truffaire's position
ARCORA is deployed with five FPO partners and has produced over 2,000 diagnostic reports. Our stated position on those records is that they belong to the FPOs and their members, and that the aggregate knowledge built from them is theirs.
That commitment is the substance of one FPO, one plant, and it is deliberately structural rather than promotional: it is the reason we work through FPOs rather than signing individual farmers, because a collective can hold and enforce a claim that an individual smallholder cannot.
We are not claiming this is unusual generosity. It is the only arrangement under which a long-term agricultural record continues to be generated, and it is therefore also the commercially durable one.
Frequently asked questions
Does Indian law settle this?
Data protection law establishes obligations around personal data, and agricultural records frequently sit in a less-defined space — particularly aggregated, de-identified data. The contractual position between the parties matters more than most people assume, which is a reason to read it.
Can data be shared with researchers?
Reasonably, with the contributing group's agreement and appropriate handling. The failure is doing it silently.
What about de-identified aggregates?
De-identification reduces individual exposure and does not settle ownership. The aggregate was built from contributions, and the question of who benefits from it remains.
How can an FPO tell if a platform is extractive?
Ask two questions: can we export everything, and what do you do with the aggregate. Evasiveness on either is the answer.
Does this slow down building useful systems?
Modestly, and it makes them last. A dataset assembled without the contributors' understanding is one disclosure away from collapsing.
Where to start
If your FPO uses any digital agricultural service, find the clause covering data and check one thing: can you export everything, unconditionally.
If the answer is no or unclear, you do not own the record your members have been generating — and that is worth knowing before another season is added to it.
What the individual records in question actually contain is described in reading a crop diagnosis report.
To discuss deployment on terms where the record stays with the FPO, get in touch.