Why Your Supplement Got Pulled From Amazon: The Five Compliance Failures We See Most Often

Amazon’s listing suspension doesn’t arrive with a detailed explanation. It arrives as a terse policy notification, usually on a Thursday or Friday, and by the time you’ve decoded which ASIN is affected and why, your weekend is gone and your inventory is sitting in an FBA warehouse earning nothing. Over the years we’ve processed thousands of COAs for brands selling through FBA, and the same five failure modes appear with enough regularity that we can usually spot them in a document review before a seller even submits to Amazon. The frustrating part: most of these failures are preventable, and none of them require a product reformulation to fix.

Amazon’s compliance program for dietary supplements has tightened substantially since 2022. The marketplace now requires third-party testing results from ISO 17025 accredited facilities for most supplement categories, and it uses automated document parsing tools to flag documentation that doesn’t meet its criteria before a human reviewer ever touches the file. Understanding what that automated review is checking — and where it reliably trips people up — is the most useful thing a supplement brand can know right now.

What Amazon’s Compliance Review Is Actually Evaluating

Most sellers picture Amazon’s compliance team as reviewers manually reading through PDFs. The process is more structured than that. Amazon’s Compliance Science team uses a combination of automated document parsing and human review to evaluate submitted COAs against a documented set of requirements. The primary checkpoints include: laboratory accreditation status, the specific analytes tested and the methods used, quantitative results compared against defined limits, lot and batch traceability, and whether the testing date falls within the required window — typically within 12 months for most category-specific compliance submissions, though this varies by product type and complaint history.

Lab accreditation is the gateway criterion. If a COA doesn’t originate from an ISO 17025 accredited laboratory — specifically one whose scope of accreditation covers the analytes and methods being reported — the document can fail before a single result is evaluated. This catches a surprising number of brands who used domestic labs without confirming accreditation scope, or who sourced testing from foreign facilities whose accreditation status Amazon doesn’t recognize within its review framework.

The analytes required depend on product category. For finished dietary supplements, Amazon’s baseline expectation typically includes: heavy metals (arsenic, cadmium, lead, and mercury), microbiology (total aerobic plate count, yeast and mold, Listeria monocytogenes, E. coli, and Salmonella spp.), and identity verification for the stated botanical or active ingredient. Weight loss, sports nutrition, and energy products face elevated scrutiny and often require additional screening for undeclared active pharmaceutical ingredients (APIs) — a category where FDA warning letters have been especially active.

The Five Failure Modes That Come Up Again and Again

COAs from labs whose ISO 17025 scope doesn’t cover the reported analytes. This is the most common failure, and it’s deceptively easy to miss. A laboratory can hold ISO 17025 accreditation and still not be accredited for every test it performs. Accreditation scope is analyte-specific and method-specific — a lab accredited for inductively coupled plasma mass spectrometry (ICP-MS) heavy metals in food matrices may not hold an accredited scope for the same analysis in finished supplement capsules. We see COAs submitted from accredited labs where specific analytes were tested using methods listed as “non-accredited” in the fine print of the report. Amazon’s reviewers cross-reference A2LA and PJLA laboratory databases directly, and they’ve grown more sophisticated at identifying scope mismatches.

Label claims that drift into FDA drug claim territory. Amazon’s automated listing review scans product titles and detail pages for language that could constitute a disease claim under 21 CFR 101.93 and 21 CFR 101.14. The FDA distinguishes between structure/function claims (“supports immune health”) and disease claims (“reduces symptoms of influenza”) — but the practical line blurs quickly. Phrases like “reduces inflammation,” “lowers blood pressure,” “reverses insulin resistance,” or “clinical-strength” are reliable triggers. When a listing is flagged for a potential drug claim, the compliance documentation is frequently pulled into review simultaneously, meaning a copywriting problem in the marketing team cascades into a laboratory documentation problem for the operations team. We’ve seen this exact sequence enough times that we now recommend brands conduct a label claim audit before initiating a testing program — there’s no point investing in testing documentation for a label that’s already flagged.

Heavy metal results that exceed California Prop 65 maximum allowable dose levels. Even when testing is technically valid and performed by an accredited lab, the numbers themselves can be disqualifying. California’s Proposition 65 sets maximum allowable dose levels (MADLs) for reproductive toxicity and carcinogenicity. For lead, the reproductive toxicity MADL is 0.5 µg per day. For a product delivering a 3-gram daily serving of concentrated turmeric root powder, a lead concentration of 0.8 ppm in the bulk ingredient — a number that would look unremarkable in many food contexts — translates to 2.4 µg of lead per daily dose, nearly five times the MADL. Amazon’s compliance review for products distributed in California incorporates Prop 65 thresholds, and a COA that passes FDA/USP-based limits can still trigger a removal on Prop 65 grounds. Brands selling nationally with California as part of their footprint need to calculate dose-adjusted exposure for their highest-serving product — not just report raw concentration.

Microbiology failures on the finished product. Total aerobic plate count exceedances are less common on finished encapsulated supplements than on raw botanical powders, but they occur. More often, we see failures on pathogen testing — Salmonella in botanical powders, or Listeria in protein-based blends. The root cause is usually upstream: a brand tests finished goods but sources raw materials from a supplier who hasn’t been independently microbiology-tested, and contamination survives the encapsulation process. Amazon’s requirements don’t distinguish between finished-product testing and raw material traceability — they want clean numbers on the specific finished product being sold. A positive pathogen result on a submitted COA, or the complete absence of pathogen testing, is an immediate compliance failure with no path to resolution except a new COA from a new lot.

COA formatting that fails automated document parsing. This one is frustrating because it has nothing to do with the quality of the science. Amazon’s document parsing system — like most automated compliance pipelines — needs to extract specific structured fields: sample identification, test date, analyte names, quantitative results, units, specified limits, and laboratory contact information including accreditation certificate number. COAs that use non-standard table layouts, PDFs generated from scanned (non-searchable) images, reports with missing lot number fields, or documents that express results as ratios without a clear pass/fail designation against a named limit all create parsing failures. A flagged document goes to manual review, which adds processing time — and manual reviewers apply more scrutiny, not less. We’ve seen well-documented products with clean results get caught in extended review cycles simply because the lab’s report template didn’t conform to Amazon’s parsing expectations.

What a Testing Program That Actually Survives Review Looks Like

The brands that navigate Amazon compliance cleanly share a few operational habits that are worth naming directly.

Use a single ISO 17025 accredited laboratory for your core testing panel. Splitting chemistry and microbiology testing across two labs for cost reasons creates traceability problems. When Amazon requests a COA covering your full required analyte panel and your heavy metals results come from one lab and your pathogen panel from another, you’re presenting two documents with two separate chains of accreditation and two independent lot references. Consolidating with a lab that holds accredited scope for both chemistry and microbiology — and issues a unified COA — removes an entire category of document review complexity. It’s also easier to defend in an escalated review.

Calculate Prop 65 dose-adjusted exposure before you submit results. Don’t assume that USP or FDA compliance means Prop 65 compliance. Pull your serving size from the supplement facts panel, multiply the concentration result by the serving mass, and confirm the daily exposure clears the relevant MADL. For lead in botanical products, the 0.5 µg/day reproductive toxicity MADL is the binding threshold for California distribution. For cadmium, the cancer MADL is 4.1 µg/day. If you’re close to a threshold, pursue a product-specific Prop 65 exposure assessment from a qualified toxicologist before the listing goes live — not after a removal notice arrives.

Put a COA expiration calendar on your compliance dashboard. Amazon requires documentation to be current — generally within 12 months of submission. We’ve watched brands lose listings because a perfectly valid COA aged out and no one had a scheduled re-test in place. This is a purely administrative failure, and it’s one of the easiest to prevent.

Run a pre-upload checklist before every COA submission. Confirm: the issuing lab holds current ISO 17025 accreditation with an in-scope method for each reported analyte (verify directly in the A2LA or PJLA database, not just on the lab’s website); the PDF is text-searchable; the lot number on the COA matches the ASIN-specific lot being represented; every result includes both a unit and a specified limit; and the report includes the lab’s accreditation certificate number and scope reference.

The upfront investment in a comprehensive supplement compliance testing panel at an ISO 17025 accredited US laboratory — typically in the range of $800–$2,500 depending on the analyte panel and product type — is a fraction of the revenue exposure from a suppressed ASIN. We’ve spoken with sellers who absorbed five-figure losses in a single week waiting on a compliance review resolution that could have been avoided with a more thorough initial document review. Getting the testing program right before the listing goes live is still the most cost-effective insurance a supplement brand can buy.

---

Written by Nour Abochama, Vice President of Operations, Qalitex Laboratories. Learn more about our team

Talk to our team about your testing needs. Contact us

Related from our network

• FDA cGMP Audit Preparation for Dietary Supplement Manufacturers — Aurora TIC’s regulatory consulting team covers 21 CFR Part 111 readiness, FDA inspection prep, and corrective action planning.
• Supplier COA Verification and Raw Material Testing for Supplement Ingredients — Ayah Labs specializes in evaluating incoming ingredient documentation and independent raw material testing before production.