Most brands meet the Certificate of Analysis late — usually the week a retailer asks for one. By then the testing has already happened, and the certificate on file was written for the manufacturer’s own records rather than for a buyer’s compliance team. That mismatch is a common reason a listing stalls.
What the document is actually for
A Certificate of Analysis records what was tested, by which method, what result came back, and what specification that result was measured against. Its job is not to say the product is good. Its job is to let someone who has never met you verify a claim independently.
That distinction matters, because it explains what reviewers look for. They are not reading for reassurance. They are reading for traceability: can this number be tied to a defined method, a defined specification, and an accountable laboratory?
The four things a reviewer looks for first
Identity before potency. A potency figure is meaningless until the material has been confirmed to be what the label says it is. Identity testing — FT-NIR, FTIR, or a chromatographic method depending on the matrix — comes first in the sequence and should be visible on the certificate.
A method beside every result. “Vitamin D3: 102.4%” is not a result. “Vitamin D3, HPLC-UV, 102.4% of label claim, specification 90–120%” is. A number without a method and a specification cannot be checked by anyone.
Contaminants, not just actives. Heavy metals by ICP-MS and a microbiological panel are usually mandatory for onboarding regardless of category. Brands that test only for actives are frequently sent back to the laboratory at exactly the wrong moment.
Accreditation and sign-off. Whether the issuing laboratory is ISO/IEC 17025 accredited, and whether the certificate carries evidence of review, changes how much weight it is given. At curí every certificate passes second-scientist review and Technical Director sign-off before issue.
Why reformatting later is expensive
A certificate written for internal use often cannot simply be re-typed into a retailer’s template. Fields that were never captured — a specification range, a method reference, a lot linkage — are not editorial gaps. They are testing gaps, and closing them means running work again against the correct specification.
The cost is rarely the retest itself. It is the calendar. Vendor onboarding windows do not usually move because a laboratory needs another week, and a launch date that slips past a category reset can cost a season.
What to do before you test, not after
Decide which channels the product is going into before the first sample is submitted, and have the specification written against those requirements. Amazon, a national grocer and an international e-commerce platform do not ask identical questions, but the union of their requirements is knowable in advance and cheap to satisfy up front.
The practical version of this is short: bring the retailer’s requirements to the laboratory at the start. A certificate designed for its destination rarely needs to be rebuilt.
This note describes curí’s own laboratory practice and is written for brand and regulatory teams. It is not regulatory advice on a specific product. All programs are designed to support structure/function claims permissible under DSHEA; curí does not design studies to support disease claims.