How to tell a real test report from a photocopy

1. The model. Compare the model or type designation on the report against the goods on your purchase order, character for character. A trailing letter, a suffix, a "-B" is a different product and often a materially different one. This is the check that finds the most problems and takes the least time.
2. The applicant. The company that commissioned the test is named on the report. Compare it with the company invoicing you. A report commissioned by a component maker, a module supplier or a different factory is a real report about a real thing, and it may or may not be usable for your product — but you need to know which situation you are in.
3. The date, and the edition of the standard. An old report may predate component changes; a report against a superseded edition of a standard is testing against requirements that no longer apply. Both are staleness rather than fraud, and staleness is the commonest defect in a supplier compliance file.
4. The sample description. What was actually tested — the configuration, the accessories, the firmware version, the power supply used. A report on a sample without the radio, or with a different antenna, or with a different adapter, does not describe the product you are shipping.
5. The accreditation and its scope. Which body accredits the laboratory, and does the accreditation cover this specific test? This is the check almost nobody runs and it is where a subtle failure hides.
6. The results, not just the conclusion. Read the measured values against the limits. A report with a pass on the summary page and marginal values in the body is telling you something about production variation that the summary hides.
7. Verification with the laboratory. Many laboratories will confirm whether a report number is theirs and what it covers. That call is short and it settles the authenticity question completely — which is why it belongs last: by then you already know whether the document is about your product, which is the harder question.

A factory is asked for "the CE report" and sends what it has. What it has may be the report for the model it sells most, or for the platform your product is derived from, or for the module inside it. None of that is deception; it is a supplier answering a vague question with the nearest document.

Which means the fix is on your side: ask for the report FOR YOUR MODEL, by the model designation, and say what you will do with it. A specific request produces a specific answer or an honest "we do not have that", and the second answer is far more useful than a plausible substitute.

The component-report trap. A cell has a report, a module has a report, a power supply has a report — and none of them is a report for the finished product. Component reports are real inputs to a technical file; they are not conclusions about your device.

The platform-report trap. Where a product is built on a solution house's reference design, the report may belong to that design and to a different assembler. Whether you may rely on it is a commercial and legal question with a real answer, and it should be asked before production rather than after an enquiry.

And the configuration trap, which is the subtlest. A report is for what was tested. If your version has a different antenna, a different adapter in the box, a bigger battery or a firmware change that alters transmit behaviour, the report may not cover it — and nothing on the document will say so.

Accreditation is a formal recognition that a laboratory is competent to perform specific tests by specific methods. The important word is specific: the accreditation document lists the tests and methods covered, and anything outside that list is not accredited work even when the laboratory performing it is accredited for other things.

So the question is two-part: which body accredits this laboratory, and is this test within the accredited scope? A laboratory will answer both, and the accreditation is normally verifiable with the accrediting body.

Why it matters commercially. Some destination regimes and some marketplaces require test data from an accredited laboratory, and some require accreditation for that particular test. A report that fails that requirement is not evidence you can use, however competently the testing was done.

This page names no laboratories, no accreditation bodies and no verification portals. Naming a laboratory is an endorsement this page cannot support, and a portal address is a link that rots — whereas the check itself does not.

A related check worth doing once: ask whether the testing was performed at the accredited laboratory or at the manufacturer's premises under a witnessed arrangement. Both can be legitimate, but they are different things and the report should say which.

Put the request in the purchase order. Reports for the exact model, naming the applicant, against the current edition of the standard, from a laboratory whose accreditation covers the test — as a deliverable with a date, alongside the samples.

Keep a one-page register per product: which reports exist, for which model, in whose name, from which laboratory, against which standard and edition, and when you checked them. That register is what turns a request from a marketplace or an authority from a scramble into an email.

Add the change-notification clause, because it protects every report you hold at once: no change to named critical components, the radio module, the antenna, the cell or the enclosure material without written notice and updated documentation.

Re-check on a schedule rather than on an incident. Standards get new editions, candidate lists grow, and factories change components. A file that was correct at launch is a file, not a fact.

And treat "we do not have that" as good news. A supplier who says it is telling you something true about where you stand, which is worth more than a document that looks like an answer and is about somebody else's product.