How to Verify Packaging Supplier Claims When You Can't Visit

Almost every claim a supplier makes has exactly one piece of evidence that would settle it — and the useful skill is knowing which one. “We are a factory” is settled by a business licence and an address, not by a photograph of a building. “We inspect before shipment” is settled by a report with sampling data in it, not by the word “passed”.

Most buyers who cannot fly out do one of two things: they accept the claim, or they ask for “proof” and receive a document that looks reassuring without settling anything. This is the third option — matching each claim to the specific evidence that would actually resolve it, and reading that evidence for the things that are easy to miss.

The rule: one claim, one piece of evidence

Write the claims down as a list, then put the settling evidence next to each. The exercise is uncomfortable precisely because it exposes how many claims are currently being taken on trust.

Claim	What actually settles it
“We are a factory”	Business licence showing manufacturing in the registered scope, plus the address goods actually ship from matching it
“We are certified”	The certificate itself, with the legal entity name on it matching your contract counterparty — see below
“We do our own molding”	Mold inventory, machine list with tonnage, or footage of the specific machine running your part
“We inspect before shipment”	A report containing sampling level, defect counts by class and a signature — not a certificate of conformity
“We have done this before”	Named references you can contact, or a sample of the physical output for a comparable part

Two things follow from writing it out. First, several claims collapse into the same evidence — one factory walkthrough settles four of them. Second, some claims have no available evidence at all, and those need a decision rather than more requests.

Certificates: the trap is not “do they have one”

The common failure is not a missing certificate. It is a certificate that exists, is genuine, and does not belong to the entity you are buying from. Certificates are frequently held by a parent, a sister factory, a trading arm, or a partner the supplier works with — and a scan in a PDF gives you no way to tell.

Four checks resolve it, and all four are done in an afternoon:

1. The legal entity name. Compare the name on the certificate with the name on your quotation, your contract and the bank details. If those are three different companies, that is the thing to resolve first — before the certificate.
2. The scope wording. A certificate covers a defined scope: specific sites, specific activities, specific product categories. A certificate for “trade of packaging products” is not a certificate for manufacturing them, and the scope line is where that shows up.
3. The issuing body's own record. Look the certificate up through the body that issued it rather than accepting the PDF. A certificate number that does not resolve in the issuer's own registry settles the question immediately.
4. Validity and site. An expired certificate, or one covering a site that is not where your goods are made, is the same as not having one for your purposes.

Our own position is simple and you can hold any supplier to it: ask for the certificate and the entity name on it, and ask which site it covers. A supplier who will not show you the document, or who answers a direct question about the entity name with a general assurance, has told you something useful.

Why we take this seriously: we got it wrong ourselves

This section is in the first person because the trap above is not hypothetical to us. It is the one we fell into.

In September 2026 we audited our own published pages and found that a batch of articles on this site had gone live carrying claims we could not support. They described production as certified to ISO 9001, a blockchain traceability system, and a customer relationship — none of which we held. One page displayed a certificate belonging to another company as though it were ours.

The root cause was not a bad document. It was a missing gate: the pages had been produced faster than they were verified, and nothing between writing and publishing was asking whether the claims were ours to make.

What we changed:

• the unsupported claims and the certificate image were removed;
• an automatic check was built into the publishing pipeline. It scans each article — title, description and body — for the specific claims we cannot support: named certification schemes, certification bodies, regulatory endorsements, traceability technology, and named customers or carriers. Publishing is refused until each one is either removed or explicitly exempted in that article's own frontmatter — which makes clearing a claim a deliberate, visible act with a name on it, rather than an oversight nobody notices;
• the exemption is per article, so an article that legitimately discusses one of those subjects has to declare that it is doing so.

We are telling you this for two reasons. The first is that the certificate trap described above is common enough to happen by accident, and we are a working example rather than a warning about other people. The second is that it tells you what to expect from us: if we cannot show you the document, we should not be claiming it. Hold us to the same test this article recommends you apply to everyone else — and if you ask us which entity commissioned a report or which site a certificate covers, you should get a straight answer rather than a general assurance.

What an inspection report has to contain to be worth anything

“Passed inspection” is not evidence. It is a conclusion with the reasoning removed. A report is useful when it contains enough that you could disagree with it:

• A report identifier and a date — so it can be referenced if the same goods are inspected again.
• The sampling plan — the standard and the level. Our published reports state the standard and the level for exactly this reason.
• Lot size and sample size — without them, the defect numbers mean nothing.
• Defects counted by class, against the limits — not a total. Our 2007 record for an 80g aluminium jar reads 6 major and 5 minor findings against limits of 10 and 21: two numbers, and a reader can see how close it came.
• Where the check was done and by whom — on site, by a named role, at a stated stage.
• Photographs that identify the goods — cartons, marks, the actual units, not a stock image.

A report missing the sampling plan is not a weaker report. It is a different kind of document, and it cannot be argued with, which is usually the point of sending it.

Two document families that look alike and are not

This is the distinction that costs buyers the most, because both documents arrive as a PDF and both say the goods were tested. They are not the same kind of evidence, and they answer different questions.

	Material / safety test report	Pre-shipment inspection report
Commissioned by	Usually the supplier	You, or someone acting for you
What is tested	A submitted sample	Your production lot
Covers	That sample, at that time	The lot sampled, at that stage
Answers	“Is this material composition acceptable?”	“Are these units, now, per specification?”
Repeats per order	No — a one-off screening	Yes — per lot

A material or safety test report — heavy metals, restricted substances, food-contact, phthalates — is valuable, and it is now standard practice for a supplier to hold one and send it with a quotation. Read it for three things before you treat it as evidence about your order:

• Who commissioned it. If the client named on the report is the supplier, then the supplier chose what to submit. That is not a reason to distrust the report — it is a reason to read the next two lines.
• What it says it covers. Reports of this kind generally state that the result applies to the sample submitted and no more. That sentence is doing real work: it means the report is not a statement about the units in your container.
• When it was done. A screening from a previous year, on a material grade you are not using, is a statement about a different order.

None of this makes such reports worthless — they are how you find out whether a material grade contains something you cannot ship into your market. It makes them the wrong document for the question “are the goods I am about to pay for correct?”, which is a pre-shipment inspection, and a different purchase.

The refusal is itself a finding

One of the most useful things an inspection report can contain is a record of what was not done. In our archive, a 2015 final inspection of 18,100 plastic sprayers found 80% of sampled units with untrimmed burrs, marked the result pending and required the factory to fix before release — and the report notes that the factory declined a drop test.

The refusal is in the document. That is what a report is for: an inspection you would happily show a buyer is one that records the awkward facts alongside the flattering ones. When you receive a report, ask what it records that did not pass. A document with no negative entries across many inspections is describing a process that was not really looking.

The same report carries a limitation worth internalising: the inspector does not make the commercial decision. Our archive states plainly that only the client holds the final reject or approve call. An inspection produces evidence; it does not produce a verdict on your behalf.

What video can and cannot prove

A live walkthrough — a call where someone carries a phone through the plant while you ask questions — settles a great deal, and costs nothing. It is also the single most requested and least structured form of remote verification.

What makes it worth something:

• Live, not a file. A recorded clip is edited by definition. A live call where your question changes where the camera goes is not.
• Ask for specific things, not a tour. The machine running a part of your size and material; the mold rack with your mold in it; the QC bench and how measurements are recorded; where finished cartons are staged before shipping.
• Ask for continuity. Warehouse, production floor and the address on the gate in one unbroken movement. A tour assembled from clips across different sites cannot be checked, and the seams are where the useful information is.
• Ask a question you do not know the answer to — how a specific tolerance is held, what happens to a rejected lot. Prepared material survives an open question less well.

What video cannot do is prove that the units in your container were produced on the machine you were shown, or that the process is repeatable. It is evidence about capability, not about your order. That distinction is what the inspection stage is for.

The sample you approved, and the run you receive

Most first-order disputes are not about quality in the abstract. They are about the distance between the approved sample and production — and that distance is controlled before production, not after.

• Approve a golden sample and keep it. The reference has to exist somewhere other than in an email thread, and both sides should hold it.
• Say what “same as sample” means. Colour, surface, wall thickness, function, print registration. “Same” unqualified is not a specification.
• Decide tolerance before the run. What deviation is acceptable, and what is a reject. If this is only discussed after goods arrive, it will be negotiated rather than measured.

Where a third-party inspection fits

Third-party inspection firms exist to send someone to look at goods on your behalf when you cannot. They are genuinely useful for one specific job: a witnessed, independent sampling of a defined lot against a defined specification, producing a report you did not commission from an interested party.

They are worth understanding for what they do not do:

• They sample the goods in front of them. They do not audit a management system, and they do not certify your supplier.
• They inspect against the specification you gave them. A thin specification produces a thin report.
• They do not follow the goods after the inspection. What happens between inspection and loading is outside their scope.
• They are an additional cost on the order, and usually least economic on small runs — which is precisely when a first order tends to happen.

A third-party inspection is a substitute for being there at that moment. It is not a substitute for a specification.

Questions that discriminate

Some questions get a useful answer from any supplier. These tend to get a useful answer only from the ones you want:

• Which legal entity will be on my contract, and which entity holds the certificate you sent me?
• Which site is my order made at, and is that the site your certificate covers?
• May I have the mold inventory for parts of this size and material?
• What did your last inspection report record that did not pass?
• Who has the final decision on releasing a lot that is borderline — you, or me?
• Will you hold a golden sample, and can I hold one too?

The pattern is that these ask for something specific and verifiable. General assurances are easy to give and impossible to check, which is why they are so common in replies.

When a claim simply cannot be verified

Some things are not verifiable remotely at any reasonable cost — a supplier's internal quality culture is the honest example. The response is not to keep asking. It is to decide explicitly how much of the order you are willing to put behind the unverified part, and to structure the commercial terms around that:

1. Verify what you can — entity, site, certificate, report, sample.
2. Name what you cannot, out loud.
3. Size the exposure — a first order small enough that being wrong is survivable is a legitimate verification strategy, and often a better one than a long questionnaire.
4. Put the critical items in the inspection plan, so that the part you could not verify beforehand is at least measured before it ships.

That is the honest version of due diligence at a distance: not certainty, but a named list of what is verified, what is not, and what you have done about the difference.