The COA They Sent You Is Real. That Doesn't Mean It's About Your Vial.
Five fields decide whether a certificate of analysis describes the vial in your hand or one somebody tested once, two years ago — and you can check all five in thirty seconds.
"It's third-party tested — here's the COA."
I get that PDF forwarded to me constantly. Group-buy organizers send it. Resellers send it. Guys three orders in who keep a folder of screenshots on their phone send it, usually with some version of does this look right to you? And most of the time the document is perfectly real. A real lab, on a real instrument, ran a real sample. That is not the problem.
The problem is that nothing on the page connects it to the vial that showed up at your door.
A certificate with no lot number is a receipt with no date, no amount and no item on it. It proves a transaction happened somewhere. It says nothing about yours. And I'll name the metaphor rather than leave you to find it: that's the whole article.
"Is this COA legit" is the wrong question
Everyone asks whether the document is fake. Almost nobody asks whether it's about their batch. Those are different questions, and only one of them you can actually answer from your kitchen table.
Here's the incentive structure, stated plainly. Testing costs money per batch. A document costs nothing to re-upload. If "third-party tested" is what buyers compete on, the cheapest way to be competitive is to test once and let the file do the work forever. That's not fraud. It's just marketing, wearing a lab coat.
What I can't tell you is how often it happens. Nobody has a census of this market and I'm not going to invent one. What I can tell you is that the failure mode is structural, so it's checkable — and that a buyer can be handed a document that is unverifiable by design and never notice, because nobody was ever taught which fields to look at first.
The Five-Field Test: Lot, Date, Method, Lab, Scope
In that order. Before you read a single purity number.
Lot. Find the lot or batch field. Compare it character-for-character to the vial label. Not "looks similar." Identical. If the field is blank, cropped out of the screenshot, or absent from the document entirely, stop reading the results — you're reading about a different vial.
And don't let anyone tell you batches are interchangeable. Batch-to-batch consistency is something you measure, not something you assume. In one reported peptide workflow, mass-spec analysis across independent batches showed 78–83% peptide overlap — good, consistent, directionally reliable, and still not 100%. That was a workflow built to be reproducible. In an analysis of a whey hydrolysate, RP-HPLC and LC–MS/MS found a complex but reproducible peptidome across production batches with a conserved set of sequences, and a separate assay on that material reported low intra-batch variability. Notice the word. Intra. Within one batch. That is the honest unit of a purity claim.
Date. Two dates matter: when the lab received the sample, and when it analyzed it. A document with only a report date tells you nothing about how long that vial sat before anyone looked at it. And the reason recency matters is that labs have bad weeks. In a multi-lab benchmarking study, layered quality control caught column degradation, a misconfigured liquid chromatography setup, and site-specific pipetting failures. Those were real problems inside real labs, found because someone looked. A certificate from 2023 is not evidence about a column in 2025.
Method. The document should name the analytical method. This is the field people skip because it looks like boilerplate, and it's the one I'd argue matters most. In that same benchmarking work — seven laboratories, two instrument platforms, six data-analysis software tools — the choice of software affected identification depth and quantitative accuracy more than the instrument vendor did. Read that again. The number depends on how it was processed, not just what it was processed on. An unnamed method makes the number uninterpretable.
Lab. A named laboratory and a report number you could theoretically call and quote. "Third-party" with no third party named is an adjective, not a fact.
Scope. What did they actually test for — identity, purity, or both? Purity without identity tells you the vial contains 98% of something. And you only find what you look for: a mass-spectrometry survey of 47 commercial human embryo culture media and supplements detected undeclared proteins in every media type, ranging from 22–57 in unsupplemented media up to 103 in one serum substitute. Different field, same lesson. A label is a claim about contents. An assay is a measurement of them, and only of the things it measured.
For contrast, here's what a batch-level number looks like when someone means it: a radiotracer preparation reported 97.0% ± 0.8% radiochemical purity, n = 3. A figure, an error, a sample size. That's a measurement. "≥99%" floating in a colored box is a design choice.
What to ask, in writing
Send this, exactly, and you won't sound naive: Please send the lot-matched report for the lot printed on my vial, including the analysis date, the method used, the laboratory name and the report number, and confirmation of whether identity as well as purity was assessed.
A vendor who has it will send it. A vendor who doesn't will change the subject. Either answer is worth having.
Us
Our certificates carry a lot identifier and received and reported dates. They confirm identity by LC-MS and print the mass-confirmation spectra, not just a purity percentage. Every listing prints a purity figure on the page and carries a COA.
Two of the five fields I've just held everyone else to, I'm auditing us on in public rather than claiming: whether the lot on the physical label is printed in a form you can compare character-for-character to the document, and whether a buyer can independently verify a report number with the laboratory that issued it. Where the answer is no, it goes on the testing page by name. Not quietly rewritten. Named.
I sell this stuff. That's exactly why I'd rather you check my paperwork than trust it.
Thirty seconds
- Lot on document = lot on vial, character for character.
- Sample receipt date and analysis date present.
- Method named.
- Laboratory named, report number present.
- Identity and purity both assessed.
Missing one? Ask. Missing three? You bought a claim, not a test — at the same price.
Our testing and documentation page lists which of these fields appear on the reports we provide, and which we currently cannot produce. Read that one before you read our purity numbers.
Testing is only evidence if it points at your vial; without traceability, a COA documents a vendor's intentions rather than your product. Check the five fields before you check the number, because the number cannot be more trustworthy than the document carrying it. And when a supplier can't produce the lot-matched report, that silence is data too.
Claims this piece draws on, by the strength of the research behind each one. Ordered strongest to weakest. Community discussion is shown separately because it indicates interest, not evidence.