What Does “Third-Party Tested” Actually Mean?

Published August 27, 2026 · Artemis Labs

Research Buyer FAQ › Certificates of Analysis & Verification › What “Third-Party Tested” Actually Means

“Third-party tested” actually means one thing: a laboratory that is not the seller ran a test. The phrase does not say which test, on which batch, how often, or who chose the sample. It is a starting claim, not a conclusion — and the follow-up questions are where a buyer learns anything.

How this page was built. The definitions below apply to any vendor’s testing claims, including ours. Published findings under References are transcribed from the papers themselves.

What does the phrase actually claim?

The claim is that an outside laboratory, rather than the seller’s own staff, performed a test on the seller’s material. That is the entire content of the phrase, and it is narrower than it sounds. A vendor can print it truthfully after sending one sample of one product to one lab, once, years ago. The phrase carries no batch, no date, no method, and no frequency. Those details live in a different document — the Certificate of Analysis for a specific lot — which is why the phrase on a homepage means little until a batch-level document backs it up.

What is the difference between a third-party COA and an in-house one?

The author. An outside lab with no stake in the sale writes one kind; the seller’s own quality team writes the other. Both can be honest, and in-house testing is a normal, useful part of day-to-day quality control. The difference is what happens when a number is inconvenient. An outside lab has no commercial reason to round a purity figure up, no inventory riding on a batch passing, and a reputation that depends on its reports being right. In-house data answers “did our process work today?” Outside data answers “is the claim on the label true?” Those are different questions, and a purity figure being used to sell something should be answered by the second kind.

A document usually tells you which kind it is: a report from an outside lab carries that lab’s own name and letterhead, not the vendor’s. If a vendor will not say which of the two you are holding, that silence is itself the answer.

Who counts as a third party?

A lab the seller does not own, staff, or control — and the strength of the claim rises with how checkable that independence is. A named lab with a findable address and a working phone number is a real third party, because you can ask it to confirm the report. An unnamed “independent laboratory” is a claim you cannot check at all. Between those poles sits accreditation: some labs hold ISO/IEC 17025 accreditation, the international standard for testing-laboratory competence, which means an outside auditor periodically examines how the lab runs its methods. Accreditation is not a requirement for honest work, but it is one more thing a forger cannot easily fake and a buyer can verify from home.

One quiet detail matters more than most buyers expect: who chose the sample. If the vendor picked which vial to send, the lab tested the vendor’s best foot forward. That does not make the result false — it makes it a result about the sample, not the shelf. Lot-by-lot testing, where each production batch generates its own report, is the version of the claim that covers what you actually receive.

What does “third-party tested” not tell you?

Four things, each of which has to be checked separately.

The phrase is silent on Why it matters Where the answer lives
Which tests were run Identity and purity are different tests; a batch checked for one has not been checked for the other The methods section of the COA
Which batch was tested A test of last year’s batch says nothing about this one The lot number, matched to the vial
How often testing happens One sample, once, makes the claim legally true Whether every lot has its own report
What was not measured A standard peptide COA covers identity and purity — not sterility, not endotoxin The list of methods actually named on the document

The last row is the one vendors least like to discuss. Purity by HPLC and identity by mass spectrometry are chemistry results. They do not certify anything the method never measured, and research-use-only material is not sterile-filled. A 2024 test-purchase study of semaglutide vials from unlicensed online sellers makes the gap concrete: the samples’ labels claimed 99%, measured purity ran from 7.7% to 14.37%, and endotoxin was detected in every sample — a result no label claim would have surfaced (PMID 39509151).

Can you trust a vendor that will not show a COA?

No, and you do not need a further reason. A batch report is the cheapest document a real vendor can produce; refusing to produce one is not a policy, it is an answer. The same logic extends one step: a vendor who shows a COA but will not name the lab behind it has shown you a claim, not a verification. Independent re-testing is not a formality — when researchers re-examined a thymosin preparation by high-resolution mass spectrometry in 2018, they found a composition different from the one assumed for decades (PMID 30063862). If long-studied material can be re-measured and corrected, a commercial vendor’s unverified claim deserves at least the same scrutiny.

What should you ask a vendor instead of “is it third-party tested”?

Ask four narrow questions, because the broad one is too easy to answer with a yes. Which lab tests your material, by name? Is every lot tested, or a sample now and then? Which tests does the report include — identity, purity, or both? And does the lot number on the report match the lot number on the vial you will send me? A vendor with real testing behind them answers all four without friction, because the answers already exist on paper. How to audit the paper itself — the batch numbers, the traces, the tells of a forgery — is covered in how to spot a forged Certificate of Analysis.

References

All compounds sold by Artemis Labs are for laboratory research use only. Nothing on this page is medical advice, and no statement has been evaluated by the FDA.