A crop diagnosis is used once. A problem appears, a report identifies it, a treatment follows, and the report is finished with.
Individually that is the correct use. Collectively it discards the more valuable output. A season's worth of diagnoses across a group of farms is a record of what actually occurred, where, and when — and that record answers a question no individual report can: what is likely to happen next season, and what should be done before it does.
Almost nobody keeps it, which is why most agricultural advisory remains reactive.
What a season of records contains
Set side by side across a region and a season, diagnosis records show three things that are invisible one at a time.
Timing. When a condition first appeared, and how quickly it spread across farms. A disease that reliably surfaces in the third week after a particular stage is a disease you can prepare for.
Geography. Whether cases clustered — one village, one water source, one side of a valley. Clustering points at a cause that a per-farm view cannot see.
Recurrence. Whether the same plots had the same problem last year. Repeat occurrence in the same location usually indicates something persistent in the soil or the practice rather than a seasonal arrival.
None of these require sophisticated analysis. They require that the records exist, are located, and are dated — which is the entire argument for keeping them.
From history to a plan
The practical output is a shift in three decisions.
What to plant, and where. Persistent recurrence in specific plots argues for rotation or variety change rather than for treating the same problem annually. This is the highest-value use of the record and the slowest to act on, because it operates across seasons.
When to inspect. If a condition historically appears in a particular window, that window is when to look — deliberately, before symptoms are obvious. Early detection changes both the treatment cost and the outcome, which is the point made in why speed matters in crop disease diagnosis.
What to hold in stock. An FPO that knows which inputs were needed, when, and in what quantity last season can procure ahead rather than during a shortage at peak demand. This is frequently the most immediately monetisable use of the data.
Reading the record honestly
Three cautions, because a season of records is easy to over-read.
Absence is not absence. The record shows what was reported. A condition that farmers recognised and treated without reporting does not appear, and a region that adopted diagnosis midway through a season has a partial record. Coverage has to be understood before conclusions are drawn.
Correlation in a field is not causation. Cases clustering near a water source is a hypothesis worth investigating, not a finding.
One season is not a pattern. A single year distinguishes very little from weather. Two or three years of consistent records is where the recurrence signal becomes usable — which is an argument for starting the record now rather than for waiting until it is complete.
The general principle is the same as in any operational dataset: an attributed record is actionable, an aggregate one is not. That framing is developed in precision agriculture and the gap in India.
Why the FPO is the right unit
An individual farm's diagnosis history is thin. Ten cases across two years cannot distinguish a pattern from chance.
An FPO covering several hundred farms in a contiguous area generates enough cases in one season for timing and clustering to be legible — and it has a reason to act on them, because it procures inputs, advises members and bears some of the consequence of a bad season.
This is why aggregation at the FPO rather than the farm is the structural choice, and it is the same reasoning that determines what an FPO should digitise first: the member register comes before services precisely because the register is what makes every subsequent record locatable. That sequence is in what an FPO should digitise first.
Who the record belongs to
Aggregated diagnosis data is valuable, and value attracts claims of ownership. The position worth stating before any of this is built: the record is generated by farmers, about their land, and it belongs to them and their FPO.
That is not a courtesy. It determines whether the record continues to be generated, because a group that discovers its data was collected for someone else's benefit stops contributing. The full argument is in who owns agricultural data, and why it matters.
What Truffaire has actually built
ARCORA is deployed with five FPO partners and has produced over 2,000 diagnostic reports. That is the verified figure, and it is a foundation rather than a finished dataset.
What we have not done is publish yield outcomes or economic results attributable to seasonal planning, because we do not have verified figures for them and are not going to estimate. What the record currently supports is timing and recurrence observation within participating groups — which is where the useful work starts, not where it concludes.
The long-term structure of the commitment is described in one FPO, one plant.
Frequently asked questions
How many seasons before this is useful?
Timing observations are legible within one season if case volume is adequate. Recurrence — the more valuable signal — needs at least two, preferably three.
What if a region only recently started recording?
Start anyway. A partial record understood as partial is more useful than none, and the value compounds from the first season onward.
Does this replace agricultural extension advice?
No. It provides the local, dated, located evidence that extension advice can be applied to. Judgement about what to do with it remains agronomic.
Can weather data be combined with it?
Usefully, yes — many conditions track temperature and humidity closely. The prerequisite is that the diagnosis record is dated and located precisely enough to align with it.
What about farms outside the FPO?
They are outside the record, which means the picture covers members rather than the region. Worth stating explicitly when conclusions are drawn.
Where to start
If you run an FPO already using diagnosis, take last season's reports and sort them by date and village.
The first thing most groups see is a timing pattern nobody had noticed — a fortnight in which most of a season's cases appeared. That fortnight is next year's inspection window, and identifying it costs an afternoon.
What an individual report contains, and the order to read it in, is in reading a crop diagnosis report.
To discuss deployment through an FPO, get in touch.