What does a food safety auditor actually look at?

The auditor looks at your records, and then at your farm to see whether the two agree. The question being asked is not whether your practices are correct in the abstract — it is whether you can demonstrate that you did what your own food safety plan says you do, consistently, on days when nobody was checking. That is why the review spends more time at a desk than in a field. A practice you cannot show a record of is, for the purposes of the audit, a practice you did not perform.

Which records get requested in a food safety audit?

The common list includes input and chemical application records, water test results, worker training and hygiene records, harvest and lot records, traceability from a lot back to its inputs and forward to its buyers, supplier or input approvals, corrective action records, and the food safety plan itself. The exact set depends on your scheme, your commodity, and your buyer, and no article can substitute for the standard you are being audited against. What every version shares is that each record has to connect to the others — a harvest lot has to reach the water test and the applications behind it. Isolated records that cannot be joined are the most common gap.

Why would having no corrective actions on file count against me?

A recorded corrective action is evidence that your monitoring is real — that someone was watching closely enough to catch a problem, and that the farm did something about it. A record with no corrective actions at all suggests the opposite: either nothing is ever monitored, or problems are caught and quietly fixed without being written down. An experienced auditor reads a spotless file with suspicion, because real operations generate exceptions. A documented problem you found and corrected is worth more than a sheet with nothing on it.

Can an auditor tell that records were written just before the audit?

Usually, yes, and more easily than people expect. Records created contemporaneously carry the texture of the days they describe — varied handwriting or timestamps, corrections, gaps that match weather or a holiday, entries in a different hand when someone was away. A batch reconstructed the week before tends to be too even, too complete, and internally consistent in ways real weeks are not. Auditors cross-check dates against invoices, weather, delivery records, and payroll, and a fabricated set fails those checks. The reconstruction also usually damages the thing it was meant to prove, because it turns an honest gap into a misrepresentation.

What is a mock audit and why run one?

A mock audit is a full dry run of the real thing, done by someone other than the person who keeps the records — a colleague, a consultant, or a member of your own team who did not write them. They work from the standard, request the documents, and try to trace a lot end to end, noting every place the trail goes cold. The value is that it surfaces the gaps while you still have time to fix them honestly, rather than discovering them in front of the auditor. Run it far enough ahead that a missing record can still be rebuilt from a real source, not invented.

Food Safety Audit Preparation for Farms

A food safety audit feels like it is about your farm. It is not. The auditor cannot watch you spray, cannot stand at the wash line for a season, cannot follow a lot from the field to the buyer. They arrive for a day, sometimes less. What they can do in that day is read what you wrote down while they were not there, and check whether it holds together and matches what they see.

So the audit does not test the farm. It tests your records about the farm. Those are two different problems, and a farm excellent at the first can fail badly at the second. You can grow clean produce, treat every animal correctly, and wash every bin, and still not pass — because passing is not a claim about the work. It is a claim about the evidence of the work.

Once you see the audit that way, the exercise reorganises itself. The goal stops being “do everything right” — you were already doing that — and becomes “be able to show, from a record made at the time, that you did.” This page is about that second job: what a reviewer is really checking, which documents get pulled, and the habits that make the week before an audit uneventful.

The scheme is the authority. Audit programs differ by buyer, by commodity, and by jurisdiction — a buyer-driven audit, a third-party certification scheme, and a government program each ask for their own list, against their own standard. This page describes the shared shape of preparation, not any one scheme’s checklist and not your obligations. Your certifier and the standard you are audited against are the authority for what you must keep and for how long. A record proves what you did; it does not decide what you were required to do.

An audit tests your records, not your farm

Everything the auditor most wants to know is in the past. The spray you made in April, the water test you ran, the worker you trained, the lot you shipped — all of it is gone by the time they arrive, and the only thing standing in for it is a record. This is why an audit is a documentary exercise dressed as a farm visit. The walk-through checks that the paper and the place agree — a sign-off that claims the cooler is monitored, against a cooler with no thermometer, is a finding — but the substance is in the file. A farm that has done everything correctly and written none of it down has nothing to show, and “trust me” is not a record. The audit cannot reach the work directly; it can only reach the trail the work left.

The auditor is checking whether you did what you said you would

It is tempting to think the auditor is grading your practices — whether your wash step is good enough, your buffer wide enough. Mostly they are not. The standard sets the bar; your food safety plan is where you committed to clearing it. What the auditor checks is narrower and harder: whether you actually did the thing you wrote down that you do, every time, including the times no one was watching.

That last clause is the whole game. Anyone can do it right on the day of the audit. The plan says you sanitise the harvest bins between uses; the auditor wants to see that this happened on a wet Thursday in September when you were behind and short-handed, and the way they see it is a record made that Thursday. Consistency is the thing under test, so a “good” farm is not the one with the most impressive practices but the one whose records show the same practice, unremarkably, week after week, with the exceptions honestly marked. Discipline of the sort in farm recordkeeping is worth more here than any single well-executed step.

The documents that get requested are a short, predictable list

The specific list is your scheme’s to set, but across programs the same categories recur, because they map to the ways food gets people sick. Know them cold, so none of them is a scramble.

Input and chemical application records — what was applied, the rate, the lot, and the intervals that had to pass before re-entry or harvest: the same record covered in spray and input records.
Water testing — results for the water that touched the crop, for irrigation or washing, sampled and tested the way your standard specifies.
Worker training and hygiene — who was trained, on what, when, and the sign-offs that show it happened, alongside the labor and safety records behind them.
Harvest and lot records — what was harvested, from where, when, and under what lot code, so a shipment resolves to a specific set of days and fields.
Traceability — the ability to move from a lot back to its inputs and forward to its buyers, the subject of traceability, and often the record the audit stresses hardest.
Supplier and input approvals — evidence that the seed, transplants, amendments, and packaging came from sources you vetted rather than whatever was cheapest that week.
Corrective actions — the log of problems found and what was done about them, which deserves its own section below.
The food safety plan itself — the document every other record is measured against, matching what you actually do.

What matters more than the list is that the items connect. A harvest lot has to reach the water test that covered it and the applications made to its field; a buyer complaint has to reach the corrective action it triggered. Records that each exist but cannot be joined are the most common finding, because the auditor’s real test is a trace, and a trace fails at the first broken link.

A clean sheet is a red flag

Here is the part that reverses most people’s intuition. You would expect a corrective action log with nothing on it to be the ideal — proof that nothing ever went wrong. To an experienced auditor it reads as the opposite: proof that nothing is ever caught. Real operations generate exceptions — a cooler drifts out of range, a delivery arrives without its paperwork, a worker skips a step, a water result comes back off. A farm actually monitoring itself finds these things, because finding them is what monitoring is for.

So an empty corrective action file does not say “this farm is perfect.” It says “this farm is not looking,” or worse, “this farm finds problems and does not write them down” — both findings. A documented problem, spotted and fixed and verified, is stronger evidence than a blank page, because it demonstrates the one thing the auditor most wants to be true: that the system catches its own failures with no auditor present to force the issue. So the instruction is uncomfortable but simple. Do not hide the exceptions; record them, and record what you did next. A corrective action is not an admission of a bad farm — it is the clearest proof of a working one.

A record written the week before is visible from across the room

There is a difference between a record made at the moment the thing happened and one made the week before the audit to describe a thing that happened months ago. It is not subtle, and it is one of the first things a trained reviewer reads. A contemporaneous record carries the texture of the days it describes. The handwriting changes because different people made entries; the ink or the device changes; there are corrections and cross-outs; there are gaps that line up with a holiday, a breakdown, or a week of rain. A batch reconstructed in one sitting has none of that. It is too even, too complete, filled in by one hand in one pen — a week that never existed, made too tidy. And auditors cross-check: the entry says you sprayed on the ninth, but the chemical invoice is dated the eleventh; the harvest record shows a crew of six, but payroll shows three that day.

The reconstruction does not just fail to help; it harms you, because it converts an honest gap into a misrepresentation, and that is a far more serious finding than a missing record. A blank is a lapse; a fabricated record that contradicts an invoice is a question about whether anything in the file can be trusted. The only record worth having is the one made when it was made — an argument for a system you can complete in the field, in the moment, not at a desk under deadline.

Run the audit before the auditor does

The strongest preparation is a mock audit: a full dry run conducted by someone who did not create the records. That last condition is the important one. The person who keeps the records cannot audit them, because they read what they meant rather than what is there, and fill every gap silently from memory. Hand the standard to a colleague, a consultant, or a teammate who was not the recordkeeper, and have them do what the real auditor will do.

Give them a lot code and ask them to trace it — back to the field, the inputs, the water, the harvest crew, and forward to every buyer who received it — and to request each document on the list and time how long it takes to produce. Note every place the trail goes cold, every record that lives in someone’s head or a different truck, every join that cannot be made. Those cold spots are what the real audit will find, with the difference that now you have time to fix them from a real source instead of inventing them.

Run it early. A mock audit two days out only tells you what you are about to fail; one with weeks of runway lets you rebuild a genuinely missing record from an invoice, a delivery slip, or a test result that still exists — an honest reconstruction from a real source, which is a different thing from fabrication. It is also where the connections between records get tested, the same way withdrawal and residue and organic certification records get stress-tested by tracing a single lot through all of them.

The week before an audit should be uneventful

There is a single test of whether your recordkeeping is real: how does the week before the audit feel? If it is a scramble — reconstructing logs, chasing down water results, backfilling training sign-offs, rehearsing what to say — then the records were never a genuine account of the farm. They were a story assembled for the occasion, and an assembled story is exactly what the auditor is trained to detect.

When the recordkeeping is a habit rather than an event, the week before holds nothing to do. The records already describe the farm, because they were made as the farm ran. You confirm the food safety plan still matches practice, make sure the files are within reach, and go back to work. The audit becomes a reading of a document you have been writing all year, not the frantic authoring of one. Boring is what a true record looks like from the outside.

The exports are a view of your recordkeeping, not a substitute for it

When the auditor asks for the input applications, the treatment log, and the traceability packet, what you hand over is a formatting problem on top of a recordkeeping problem — and only the first is something software should touch. Farm40 ships twelve one-click CSV exports, among them chemical and input use, the livestock treatment log, harvest traceability, and a traceability packet: the records an auditor commonly asks for, already laid out in a form you can hand across the desk. The limit — the important half of the sentence — is that an export is only a view of what you entered. It formats your recordkeeping; it does not perform it. It reproduces every application you logged and says nothing about the one you forgot to log, because it has no way to know that spray happened. The discipline the audit tests still lives with you and the moment you did the work; the export only makes what you already captured legible.

Which returns to where this started. The audit reaches the farm only through its records, and the records are only as true as the habit that made them. The thing being audited — that you did what you said you would, consistently, when no one was watching — is decided in the field, on the ordinary days, long before anyone asks to see the file.

Frequently asked questions

What does a food safety auditor actually look at?: The auditor looks at your records, and then at your farm to see whether the two agree. The question being asked is not whether your practices are correct in the abstract — it is whether you can demonstrate that you did what your own food safety plan says you do, consistently, on days when nobody was checking. That is why the review spends more time at a desk than in a field. A practice you cannot show a record of is, for the purposes of the audit, a practice you did not perform.
Which records get requested in a food safety audit?: The common list includes input and chemical application records, water test results, worker training and hygiene records, harvest and lot records, traceability from a lot back to its inputs and forward to its buyers, supplier or input approvals, corrective action records, and the food safety plan itself. The exact set depends on your scheme, your commodity, and your buyer, and no article can substitute for the standard you are being audited against. What every version shares is that each record has to connect to the others — a harvest lot has to reach the water test and the applications behind it. Isolated records that cannot be joined are the most common gap.
Why would having no corrective actions on file count against me?: A recorded corrective action is evidence that your monitoring is real — that someone was watching closely enough to catch a problem, and that the farm did something about it. A record with no corrective actions at all suggests the opposite: either nothing is ever monitored, or problems are caught and quietly fixed without being written down. An experienced auditor reads a spotless file with suspicion, because real operations generate exceptions. A documented problem you found and corrected is worth more than a sheet with nothing on it.
Can an auditor tell that records were written just before the audit?: Usually, yes, and more easily than people expect. Records created contemporaneously carry the texture of the days they describe — varied handwriting or timestamps, corrections, gaps that match weather or a holiday, entries in a different hand when someone was away. A batch reconstructed the week before tends to be too even, too complete, and internally consistent in ways real weeks are not. Auditors cross-check dates against invoices, weather, delivery records, and payroll, and a fabricated set fails those checks. The reconstruction also usually damages the thing it was meant to prove, because it turns an honest gap into a misrepresentation.
What is a mock audit and why run one?: A mock audit is a full dry run of the real thing, done by someone other than the person who keeps the records — a colleague, a consultant, or a member of your own team who did not write them. They work from the standard, request the documents, and try to trace a lot end to end, noting every place the trail goes cold. The value is that it surfaces the gaps while you still have time to fix them honestly, rather than discovering them in front of the auditor. Run it far enough ahead that a missing record can still be rebuilt from a real source, not invented.