Research · the live claims database

1,524 listings. 587 brands. Every phrase ruled.

Every listing here is ruled phrase by phrase against the full standards list in force at its scan. Rulings, not scraped copy. This is the benchmark your audit ranks you against.

The claims database

Updated 14 Sep 2026
Phrase checks
149,000+
every listing checked against all 98 standards
Fail outright
75%
of ruled listings fail our standards; 39% of brands carry a Critical
Violations flagged
6,419
findings. 615 Critical, across 587 brands
Listings ruled
1,524
top sellers across eight categories, from 3,800+ listing scans run

Ruled against standards drawn from FDA and FTC rules, Health Canada, and Amazon and Meta policy.

Phrases that fail most

The ten most-ruled findings in the database.

If your listing uses one of these, you are not being original, you are being counted. Each is struck through, exactly as it would appear in your report.

01 “missing fda disclaimer” High 324
02 “best / most effective / strongest” Medium 306
03 “non-gmo (undocumented)” Medium 295
04 “ingredient weight with no serving size stated” Medium 228
05 “gmp certified (unverified)” Medium 226
06 “balances hormones” High 153
07 “boosts immunity / immune booster” High 140
08 “normalizes / restores” High 120
09 “studies show (uncited)” Medium 120
10 “all-natural” Medium 106

Brand counts are distinct brands, among the 587 ruled, that carry the phrase or a direct variant.

The enforcement record

Four years of tightening, dated.

Every event below changed what a supplement listing may say, and each one is already reflected in the standards.

APR 2023
FTC puts 670 brands on penalty notice

Supplement, OTC and functional-food marketers received formal Notice of Penalty Offenses: unsubstantiated claims now carry $53,088 per violation, per statutory schedule.

Rulebook 02 · FTCReflected in the standards
FY 2024
Class actions reach the listing

Plaintiffs’ firms mine listings under state consumer-protection and false-advertising statutes. The same phrase that fails an FDA standard is the exhibit.

Litigation · civilReflected in the standards
MAR 2026
The Supplement Facts panel match, as reported

Trade press reported that from 31 March 2026 Amazon began requiring detail-page content to align with the Supplement Facts panel, enforced automatically across titles, descriptions and images. Amazon has not published the policy, so this is reported rather than official.

Trade press · AmazonReflected in the standards
JUN 2026
FTC v. Amare Global

The commission sued over supplements marketed as treating depression, anxiety and ADHD, the exact named-condition territory most listings drift into.

Rulebook 02 · FTCReflected in the standards
2026
Prescription-equivalence claims prohibited by policy

Amazon’s published supplement policy names GLP-1 and prescription-equivalence framings as prohibited outright, alongside disease claims and disease terms in backend keywords.

Rulebook 03 · AmazonReflected in the standards

Updated as letters post · free to cite

The full enforcement tracker

Every dated action above, plus each FDA warning letter as it posts, with its source and the standard family it maps to.

Open the tracker

Sources: FTC penalty schedule, FDA warning-letter database, Amazon seller policy, public litigation record.

Category coverage

Where the failures concentrate.

The share of ruled listings that fail our standards outright. Stress is the least compliant shelf in the store, with gut and digestion close behind. Every category has a screenshot-ready scorecard with its scan window and standards version printed on it.

Every category sits in the category claims-risk index, published with its method, its date and a line to cite it by.

Stress
82% FAIL
All eight categories
75% FAIL

Fail means at least one Critical or High finding on the live listing. Based on 1,524 listings across sleep, gut, stress, immune, cognitive, joint, sports and energy.

Inside the 6,419 findings

Where the exposure actually lives.

Severity across every finding flagged to date. The Critical share concentrates in a handful of brands rather than spreading evenly.

By severity · 6,419 findings

615
5,804
Critical · 10%High, Medium and Low · 90%

615 Critical findings sit across 587 brands, which is why 39% of brands carry at least one: Critical exposure clusters, it does not spread evenly.

The database in four numbers

6,419
findings flagged
615
Critical
587
brands ruled
39%
of brands carry a Critical

Counts from the live database · updated 14 Sep 2026

The peer benchmark

Your audit scores you against all of this.

Every audit includes your exposure score and your position against the ruled listings in your category, before and after the rewrites go live.

Acquirers and retail buyers ask where a brand sits. This is the only database that can answer with rulings instead of impressions.

Peer benchmark · SleepExtract
Exposure score · 0 is cleanHighest we have ruled: 310
You after · 6 You before · 88
88
Before
6
After rewrite
75%
of ruled listings fail
Representative audit: figures illustrate a typical first audit

Briefings

From the research team.

Working analyses drawn from the database. Available to clients on request. Press and researchers may cite headline figures with attribution to Claims Verified.

Published quarterly · free to cite

The State of Supplement Claims

The full quarterly report: the verdict split, fail rates by category, the ten most-flagged phrases quoted verbatim from live listings, and the enforcement record behind them. Brands are never named.

Read the report

Briefing 01

The “GMP certified” problem

One of the most-ruled phrases in the database is a certification claim almost no brand can substantiate as written. Why it survives, and the compliant form that converts as well.

On request · clients and press

Briefing 02

Backend keywords are in scope

Amazon’s policy reaches fields shoppers never see. What we found in listings that stuffed condition terms into backend search.

On request · clients and press

Briefing 03

The GLP-1 gold rush

“Nature’s Ozempic” and prescription-equivalence copy: Amazon’s published supplement policy names GLP-1 drugs, and what that means for weight-loss claims.

On request · clients and press

Briefing 04

What AI tells shoppers about you

ChatGPT, Claude, Gemini, Perplexity, Grok and Amazon’s Rufus answer for your brand whether you like the answer or not. How ruled claims spread into AI shopping answers, tested monthly.

On request · clients and press

Cite usJournalists and researchers: every headline figure on this page is free to use. Credit “Claims Verified” and link here. Want to know how a number was built, or need a full briefing? Email the research team.

audit@claimsverified.org →

Method

Rulings, not scrapes.

  • Listings enter through the same engine that runs paid audits, every phrase ruled against the full standards list
  • Paid verdicts are signed by a licensed pharmacist
  • Figures update as listings are added. Individual brands are never named in public data
  • Method and standards are documented on the Standards page

Pharmacist-in-the-loop

See where I sit: free check