Proof · the record of what happened on farms

Write it down once. Prove it for years.

A farmer writes down what they did, when they did it, and what happened afterwards. That single record then answers questions for nine different industries who are currently estimating the answer. None of them see the farm unless the farmer says so.

Written at the timeCompared against itself firstFewer than five records, nothing publishedNo scores, no rankings, no verdicts
The whole idea, in one interaction

One record. Nine people who need it, for nine different reasons.

This is a single application, written on a spring morning on one field. Choose who is asking. The record does not change. What they read from it, the question it settles for them, and the part they still cannot see, all do.

The taxonomy

Not a free text box. A shape that means the same thing on every farm.

Records are structured, so a figure from one farm can stand next to a figure from another without anybody having to interpret it. Here is what is captured, by type of operation.

Why farm records and not more trials

A trial answers the question under the conditions it was run in.

Trials are careful, controlled and necessary. They are also few, expensive and run somewhere that is not your farm. Farm records are the opposite on every axis, which is why the two belong together rather than in competition.

A replicated trial

Precise, and somewhere else.

  • Four to eight site years, often on one soil type
  • Conditions controlled, which is the point and also the limit
  • Ends when the funding ends, so the year after is rarely seen
  • Answers whether it can work
  • Costs six figures and takes three years
Farm records at scale

Messier, and where you actually farm.

  • Thousands of fields across every soil, season and system
  • Conditions recorded rather than controlled, so the variation is the finding
  • Keeps running, because it is a by-product of farming rather than a project
  • Answers whether it does work, on ground like yours
  • Costs a farmer forty seconds at the time

Proof does not replace trials. It tells you which of their conclusions survived contact with a real farm, and names the conditions where they did not.

Public evidence, private farm

The industry gets the picture. Nobody gets your farm.

This is the trade that makes the whole thing possible, and it only works if it is honest in both directions.

Never

Identified farm records are not sold.

Organisations pay for the aggregated, anonymised picture, and for records a farm has chosen to share with them by name. Every one of those views is logged, visible to the farm, and can be withdrawn.

Never

Below five records, nothing is published.

Narrow a question far enough and the individual farm becomes obvious again, so Proof stops answering. It applies to paying customers, and it applies to Proof. Nobody can buy their way past it.

Never

No scores, no rankings, no verdicts.

There is no traffic light on any farm, no percentile badge and no supplier league table. Proof holds the record. Anyone who wants to draw a conclusion draws it themselves, with their own name against it.

Never

Absence is not filled in.

Where a record does not exist, Proof says so rather than substituting a regional average to make a page look complete. A gap reported as a gap is worth more than a number nobody can defend.

Who it’s for

Nine industries, one underlying record.

Each of these has a different question and a different reason to distrust the answer they currently get. Each page below is written for that reader and nobody else.

Bring one question you currently answer with an estimate.

Twenty minutes. We will take your question, find the records that bear on it, and show you the answer, the denominator behind it, and the part of it nobody can currently evidence. Including when that last part is most of it.

The public record of arable evidence
PROOF AG LTD · Company no. 17211914Grosvenor House, 11 St Pauls Square, Birmingham B3 1RB© 2026 Proof