The Factory’s Report Is Not Your Report: Witnessed Sampling Explained

Article

October 26, 2026

An owner in the UK once asked to see the CCTV survey of the drains under a house he was buying. The contractor said yes, opened the file on his own laptop in the kitchen, and let him watch the whole thing. When the owner asked for a copy, the answer was no. He had seen the evidence. He did not have it, and without it he could not use it for anything.

That gap — between seeing something and being able to rely on it — is the whole subject of this article, moved from a kitchen table to a factory in China. Your supplier has a test report. It is probably genuine. It may even be for your order. And in a dispute three months later, it will very likely be worth less than you assumed, for reasons that have nothing to do with anyone lying.

What follows is not an argument that suppliers cheat. Most do not. It is an argument about evidence position: who commissioned the test, who owns the document, and who has to prove what, to whom, if the material turns out to be wrong.

Three Ways to Commission a Test — and What Each One Gives You

There are only three arrangements, and they differ on three things that matter more than the price of the test.

The middle row is the one most buyers end up in once they start caring, and it is a genuine improvement — but note what it does and does not change. Witnessing fixes the provenance of the sample. It does not fix the ownership of the report unless the commissioning party is you.

The bottom row costs more and takes longer, and it is the only one where a single party is accountable for both the sampling and the testing. When something goes wrong at the seam between those two steps, that matters.

Why a Passing Report From the Factory Is Weak — and It Is Not About Honesty

A test report is a chain with three links: the sample, the transport, and the test. The laboratory controls the third. The factory controls the first two. A report is only as strong as the link nobody independent touched.

Consider what a factory-commissioned report actually establishes. Some material was selected by the party with an interest in the result. It was packed and shipped by the same party. It arrived at a laboratory that was told what it was. The laboratory did its job correctly and reported that the material in front of it met the specification. Every sentence in that chain is true, and the conclusion the buyer wants — this lot is good — is not among them.

This is a structural point, not a moral one. It applies to a scrupulous supplier exactly as it applies to a careless one, which is why it is worth fixing in the paperwork rather than in the relationship. If you would not accept a self-audit as an audit, do not accept a self-sample as a sample.

There is a second, quieter problem: substitution is easy and leaving no trace is easy too. A sealed sample with a number, photographed at the moment of sealing, with both parties’ signatures on the record, is not tamper-proof. It is tamper-evident, which is the standard that actually exists in practice.

The Witnessed Sampling Checklist

Seven items. Each one is cheap at the factory and expensive to reconstruct later.

1. A unique sample number, written before anything is sealed. Not the order number, not the date. A number that appears on the seal, on the photograph, on the dispatch note and on the report, and nowhere else.
2. Sealing in front of both parties, and the seal photographed closed. The photograph is the evidence that the seal was intact at the start, not just at the end.
3. Both signatures on one sampling record — the factory’s representative and yours, or your agent’s. One sheet, one date, one description of where in the lot the sample came from.
4. The sampling position described, not just named. “From the middle of the stack” and “from the top layer of the second pallet” are different claims about a lot, and only one of them can be checked later.
5. A retained parallel sample, held by the factory or by you, unopened. This is what makes a retest possible without arguing about whether the first sample was representative. Agree the retention period in writing — and who pays for the retest if the two results disagree.
6. Transport conditions recorded. For anything moisture-sensitive or temperature-sensitive, the sample can be perfect at sealing and ruined on arrival. Note the packing, the desiccant, the transit time.
7. The report addressed to you. This is the item buyers forget, and it is the one that decides whether the document is yours to use.

None of this requires a laboratory in the room. A competent inspection agent can witness sampling, and for most orders that is the proportionate answer. What it does require is that the person standing there is paid by you and reports to you.

What to Write in the Purchase Order

Everything above collapses into a paragraph of contract language, and the paragraph is worth more than the witnessing itself. Put it in the PO before the first shipment, not in an email after the first problem.

Two details in that text do most of the work. Issued to the Buyer moves the document out of the supplier’s file and into yours. Retain the balance of the lot, unaltered and identifiable is what makes a retest possible at all — without it, the retest is a test of whatever is left, which is not the same thing.

What Witnessing Does Not Fix

Four limits, stated plainly, because a buyer who over-trusts a witnessed report is in a worse position than one who never asked for it.

It proves the sample, not the lot. A witnessed sample is still one sample. Whether it represents a hundred tonnes depends on the sampling plan — how many increments, from where, combined how. Witnessing guarantees the plan was followed; it does not guarantee the plan was adequate.

It does not make the laboratory accredited. A perfectly witnessed sample sent to an unaccredited laboratory still produces a report that may not travel. The two questions are independent, and both need answering.

It does not work at the destination end. We can witness sampling in China. If your dispute is about what happened to the material after it arrived — storage, handling, installation — witnessing at origin is irrelevant to that question, and no amount of paperwork at the factory will answer it.

It costs time. Sampling to a schedule, sealing, photographing and dispatching properly adds days. That is the actual price of the evidence, and it is worth paying before production rather than after a failure, when the same days become a delay on a site.

The Point Is Not Distrust

Most suppliers will hand over a sample, sign a record and address a report to your company without any objection at all, because none of it costs them anything and it removes a future argument they would rather not have. The ones who refuse are worth knowing about early.

So the practical version of this article is short: ask for the sample to be witnessed, put the clause in the PO, keep a parallel sample, and have the report addressed to you. Four sentences of paperwork, executed before the first container, that decide what you can prove eighteen months later.