A farmers market stand does not look like the setting traceability rules were written for. There is no invoice per customer, no receipt system most vendors bother with, and no way to know, after the fact, whether the person who bought six ears of corn on Saturday still has the bag. It is tempting to conclude from this that traceability simply does not apply at market scale — that the whole discipline assumes a wholesale account with a paper trail a cash table will never produce.
That conclusion mistakes what traceability actually asks for. It does not ask you to name every buyer. It asks you to know what left your farm, and a market stand can answer that question completely, even though it can almost never answer “who bought unit seventeen.” The honest version of market traceability is narrower than wholesale traceability, and it is genuinely achievable within that narrower scope.
The gap shows up clearly the first time someone actually asks a market vendor for a trace. Ask which tomatoes a specific customer bought last Saturday, and the honest answer is usually a shrug — nobody wrote it down, and nobody could have without turning a farm-stand transaction into a checkout system nobody wants. Ask instead which lots of tomatoes were on the table last Saturday, and, if the record exists, the answer is immediate. The second question is the one a recall actually needs answered, and it is the one a market table can answer completely.
Small volume does not mean no obligation
It is easy to reason that a market table selling a few hundred dollars a week is too small to matter to a recall. That reasoning does not hold up: the obligation to know what you sold is not scaled by revenue, and a small quantity of contaminated product reaching the public is still a public health concern regardless of how little money changed hands. What genuinely does scale down at market size is the effort required to meet the obligation — a market vendor’s complete traceability record is a fraction of the paperwork a wholesale account generates, not a fraction of the responsibility.
This is a relief once it is stated plainly. You are not being asked to build an invoicing system for a farm stand. You are being asked to write one line before you load the truck.
Record what you brought, not who bought it
The achievable unit of record for a market is the market day itself: which lots you loaded, in roughly what quantity, for which market on which date. That single note, made once before you leave the farm, is the entire traceability record a market stand needs to produce, and it answers the only question a recall can realistically ask of this channel: which lots reached the public, through which market, on which date.
This is deliberately less than what a CSA or a named-account direct sale can produce, and that is fine. A CSA has a member list that a market table does not; use it where it exists, and do not invent a fictional customer list where it does not. Recording what is real — lots and market days — beats pretending to a precision the channel cannot deliver.
Vendors who sell at more than one market in a week benefit particularly from keeping this note per market rather than per week. If a lot goes to a Tuesday market and a Saturday market both, a single weekly note that says “sold at markets” cannot tell you which market a specific implicated lot actually reached, while two dated notes can. The extra line is small; the gap it closes is not.
What genuinely cannot be recovered, and why that is not a failure
Be honest with yourself, and with anyone who asks, about the real limit: you cannot say which specific customer bought a specific unit at a cash table with no receipt system. If a concern is reported days or weeks later, the best a market vendor can typically do is narrow the candidate lots down to whichever ones were present at that market during the relevant window — which is often one or two lots, not the season’s entire output, provided the lots-and-dates record was actually kept.
This is not a gap in your diligence if the underlying lot record is solid. It is a structural property of cash retail sale, and it is the reason public health notices for retail-level concerns are issued publicly — reaching the population who might have bought something — rather than by individual notification. Your job is narrowing that population as much as your records allow, not achieving an impossible one-to-one match you were never going to get.
It also helps to say this out loud to the customer who reports a concern, rather than leaving them to assume you have more or less information than you do. Telling someone plainly that you can identify which lot was likely theirs, but not confirm it was definitely theirs, is more useful to them than either false confidence or a shrug, and it is the accurate description of what a market-day lot record actually gives you.
Check your specific market’s rules before assuming this is enough
The baseline described here — lots brought, by market, by date — is the practice this page can responsibly describe in general. It is not a substitute for reading your own market’s vendor agreement or your state’s direct-marketing and cottage-food rules, some of which impose specific labeling or recordkeeping requirements that go further than the general practice above. Those requirements vary enough by market and by state that no general guide can state them reliably; your market manager or your state agriculture department can.
Where a simple record removes the guesswork later
None of this requires a point-of-sale system. A notebook line per market day, kept alongside your harvest records, is a complete answer for this channel. Farm40 keeps that market-day note joined to the same lot and harvest records your other sales channels use, so a mock recall does not require a separate process for the market table versus the wholesale account. The limit is unchanged from everywhere else in the system: the join depends on the lot code being entered against the market-day note in the first place, and a note with no lot code attached is a diary entry, not a trace. For the fuller argument behind all of this, see the traceability guide, and for the discipline of testing whether any of it actually works under pressure, see running a mock recall.