This site's vendor table has a column for lab reports, and for most vendors it's a dash. That emptiness is the finding: in a market where every store claims 99% purity, only a minority will show batch-level paperwork, and only a fraction of that paperwork survives scrutiny. Having now reviewed a lot of COAs — including the 16 independently tested batches published on our COA-verified vendor profile — the quality pattern is consistent enough to write down.
The five traits of a trustworthy report
1. Purity and identity, together, from one sample. The single biggest separator. HPLC purity says "this material is X% one substance"; LC-MS identity says "and that substance has the right mass to be what the label claims." Either alone is half a report — a vial of the wrong peptide can still be 99% pure. The best reports run both on the same physical sample and say so. (Full explainer: HPLC vs mass spec.)
2. A lot number that matches the product. A COA describes one batch, not a brand. The best reports carry a lot/batch number you can match to the vial in front of you; the worst are undated, lot-less "representative" certificates recycled across every listing in the store.
3. Lab-side verification. A PDF is trivially edited. The reports worth trusting resolve in the testing lab's own database — Janoshik's public results portal, Freedom Diagnostics' searchable COA database, Chromate's verify page. Every COA we publish carries its search code for exactly this reason. If a report exists only as a vendor-hosted PDF, its trust ceiling is the vendor's honesty.
4. A named, real lab. The report names the laboratory, and the laboratory actually exists, takes public submissions, and appears on other vendors' paperwork too. Our testing-lab guide covers who's who. A COA from a lab with no findable existence outside one vendor's website is a red flag with letterhead.
5. A plausible number. Real testing across many batches produces variation — 98.9 here, 99.6 there, the occasional 97 that an honest vendor publishes anyway. A store whose every report reads 99.9% isn't showing you test results; it's showing you marketing with a decimal point.
The three failure modes
Nearly every bad COA we've seen fails one of three ways: recycled (one old report standing in for every batch since), partial (purity with no identity test — or a mass-spec trace with no purity), or unverifiable (no lot, no lab portal, sometimes no lab). None of these requires forgery, which is why they're so common — a vendor can technically "have COAs" while none of them tells you anything about your vial.
Grade a report in one minute
Purity test present? Identity test present? Same sample? Lot number matches? Lab named and real? Resolves in the lab's own database? Six yeses is a report you can lean on. Four or fewer, treat the product as untested — because functionally, for the batch you'd receive, it is. The full COA walkthrough shows each element on a real report.
For research use only. This article is educational and makes no health, dosing, or usage claims of any kind.