Every farm keeps a labor record, because you cannot pay somebody without writing down that they worked. So the record exists. The question is not whether you keep one; it is what you keep it for — and most farms have quietly decided that without noticing they decided it.
A labor record answers two entirely different questions, for two entirely different readers. The first reader is the one who pays and the one who inspects: how many hours, at what wage, was this person eligible to do this work, were they trained, were they certified. The second reader is the one who runs the place: which enterprise ate those hours, and what did that block of work cost the crop or the herd it went into. The same clock-in serves both. A farm that records hours only for payroll has thrown away half of what it already paid to collect.
This page is about the second half — the management value hiding inside the compliance record, and the handful of fields that let one record do both jobs.
This is not legal or employment advice. Employment law, wage rules, hour thresholds, worker-protection requirements, and mandatory trainings vary enormously by jurisdiction and change over time. Nothing here tells you what your farm is required to do, and you should be wary of anything that claims to. What follows is about the process of capturing the work honestly. For what the rules actually are where you farm, ask your regulator, your labour authority, and a lawyer. A record proves what you did; it does not decide what you were required to do.
The same clock-in answers two questions, and most farms ask one
Picture the hour a worker spends thinning a block of apples. To the payroll reader, that hour is a row: a name, a start, a stop, a rate, and eventually a number on a paycheck. Once the paycheck clears, a farm keeping records only for payroll will never look at it again.
To the management reader, that same hour is a cost that belongs somewhere. It was not spent on “the farm” but on that block, on that crop, and if you want to know what the block cost to bring in, the hour has to land against the block, not into a general pool. The payroll reader needs the name and the wage; the management reader needs the wage and the enterprise. The only field they disagree about is the one most farms leave blank.
The whole argument, compressed: add “on what” to “who and how long,” and one record answers both readers. Leave it off, and you pay to collect the hard part — the hours — and throw away the cheap part, where they went.
“We told them” is not a record
Worker safety training is where the gap between doing a thing and recording a thing does the most damage, because the thing itself is invisible. A treatment leaves a bottle. A harvest leaves a bin. Training leaves nothing but a changed person, and a changed person is not evidence.
“We told them how to handle it” is a true statement and a worthless record. It names no person, fixes no date, describes no content. Six months later, when someone asks whether the worker who did a task had been trained for it, “we always cover that at the start of the season” is an argument, not an answer. The record has to resolve to an individual: this person, this training, this date, evidenced by something that outlives the conversation — a signature, an acknowledgement, a logged entry with a name on it.
Notice the structure is the same one that makes a spray record defensible: not that a thing was done, but that a specific person did or received a specific thing on a specific day. A training log that reads clearly to a stranger a year later reads clearly to an auditor, an inspector, or an investigator — and, like every other record on the farm, it is worth the least on the day you write it and the most on the day someone asks.
A certification is a fact about the future
Most records describe the past. A labor log says who worked yesterday; a training record says what someone learned last spring. They are testimony: inert, backward-looking, safe to file and forget.
A certification is different in kind, in exactly the way a withdrawal period is. When a person is certified, a date in the future acquires a property: until that date they are eligible to do a thing; on that date they stop being eligible, whether or not anyone is watching the calendar. The certificate in the folder describes the past — a course was passed. The expiry date constrains the future — a day is coming when the past stops being enough.
This is why filing certificates is not the same as tracking them. A drawer full of certificates proves everyone was once qualified. It does nothing to stop a worker from doing certified work the week after their certification lapsed, because the fact that matters — the expiry date — is buried inside a document nobody re-reads. The record that protects you is not the certificate; it is a list of expiry dates, sorted, read before the dates arrive, exactly the way a withdrawal end date has to be read before the sale and not after.
The applicator licence is a record with a hard edge
Certifications shade from soft to hard. A general safety orientation is real and worth recording, but its edge is blurry — being a week past a refresher is a problem of degree. A pesticide-applicator licence is the opposite. It has a hard edge: either the person who applied the product held a current licence at the moment of application, or they did not, and that is a fact that a spray record and a labor record together are supposed to be able to prove.
This is where the two record systems have to touch. An input application record names the person who applied the product; a certification record holds that person’s applicator licence and its expiry date. The question an investigation asks — was this application made by a currently-licensed applicator — can only be answered if the two records point at the same named person and the licence’s expiry date is a thing you tracked rather than hoped about. A licence filed as a scanned certificate is one you will discover was expired only after it mattered. Write down the edge, not just the achievement: the course passed is the justification, the expiry date is the constraint, and a farm that records only the first is keeping a scrapbook, not a control.
Who did the work is a field, not a courtesy
Once the paycheck is settled, there is a strong temptation to record labor as a crew total: eight people, six hours, that block. It pays everyone correctly and feels efficient. It also erases the most useful thing a labor record can carry — the name of the individual attached to the individual hour.
Who specifically did the work is not a matter of credit or blame. It is a thread. When a food safety audit asks who handled a lot that later tested badly, a crew total cannot answer and a name can. When traceability requires you to reach back from a complaint to the people who touched the product, the labor record is one of the threads you pull. And a training or certification can only be tied to a person who was recorded doing the work in the first place. A crew number settles wages and closes every other door.
The cost of keeping the door open is one field: not “eight people, six hours” but the eight people, by name, against the hours and the block. It is more to write. It is also the difference between a labor record that only ever explains a paycheck and one that can answer a question nobody knew they would ask.
The seasonal crew is where the system breaks
A farm’s recordkeeping is only ever as good as its worst week, and for most farms the worst week is the one when the workforce triples. A permanent crew of four, each known by name and whose certifications you tracked over the winter, is an easy record to keep. Forty seasonal or contract workers arriving inside a two-week harvest window is where every clean system quietly falls apart.
It falls apart in predictable places. The people are new, so their hours get logged against a name that resolves to no one. Their trainings happen in a rush on the first morning and get recorded, if at all, as “the crew was briefed.” Their certifications — the ones with hard edges — are taken on trust because there is no time to check, and “we assumed they were licensed” is the seasonal-labor version of “we told them.” The work that most needs a clean record is done by the people the record is least ready for.
The fix is not heroic effort in the worst week; it is doing the setup before the week arrives. An employee record created before the first shift, a training logged against a name, a certification checked and its expiry date entered on arrival — each is a small act that is possible in June and impossible mid-harvest. The season does not forgive a system you meant to set up once things calmed down.
Labor is the largest cost and the least attributed
On most farms, labor is the single largest line of cost, and also the one attributed the least carefully. Land, seed, and inputs come with invoices that already say what they were for; an hour of work comes with nothing but a paycheck that says only that it happened. The most expensive thing the farm buys is the thing it is worst at knowing the purpose of.
This is why the “on what” field is not pedantry. If the largest cost on the farm lands in a general pool rather than against the enterprises that consumed it, every cost-of-production number you compute afterward is built on a guess about where the labor went. You can know exactly what the seed for a block cost and be entirely wrong about what the block cost, because the labor — the biggest piece — was allocated by memory in December instead of recorded by name in the field while the worker stood in it.
Attribution is not a report you run at the end; it is a field you fill at the start, on the same line where you wrote the name and the hours. The record you keep for the reader who pays is one field away from the record you keep for the reader who decides.
Where the labor record meets the software question
None of this requires software. It requires an employee record that resolves to a real person, an hour written against the enterprise that consumed it, a training logged against a name, and a certification carried as an expiry date you read before it lapses. A spreadsheet with the right columns does all of that, and for many farms it is the honest answer — a case worth its own reckoning under farm software.
Where software earns its place is in refusing to let the two readers drift apart. This is the design principle behind how Farm40 handles labor, and its limit belongs in the same breath as its use.Farm40 records employees, keeps labor logs that tie each block of hours to the enterprise that consumed it so the cost lands in cost of production, and stores certifications with the dates they expire; it will also export an employee labor log as CSV. The limit is the boundary this page has been drawing: it is a record, not a payroll system and not a time clock. It will not pay anyone, it will not calculate overtime eligibility, and it does not know your jurisdiction’s employment rules. For paying people and staying inside the law you still need a payroll provider and professional advice — Farm40 sits next to them and answers the question they do not: what did this work cost, and who did it.
Keep the labor record for both readers, and you stop paying twice — once to collect the hours and again, later, to reconstruct where they went. The name and the hour you already write. The enterprise and the expiry date are the two small additions that let the most expensive record on the farm earn what it costs to keep.