Only 7.5% of Supplements Are Third-Party Certified (2026)

For scale: the FDA estimates there are more than 100,000 dietary supplement products on the US market, up from about 4,000 when DSHEA passed in 1994. Every certified listing in the country, from all three programs, would account for roughly 5.3% of that.

Read that ratio carefully. Registry listings and FDA product counts are not the same unit. The 5,317 listings include products sold only outside the United States, and multiple flavour and size variants of one formula. It is an upper bound on how much of the market carries a certification, not a precise share — the true figure is lower. We use it only to establish the order of magnitude, and every product-level finding below is measured directly rather than inferred from it.

2. The USP Verified mark is essentially four retailers

USP Verified is the mark consumers are most often told to look for, and the one most likely to appear in a “how to choose a supplement” guide. It is also the smallest of the three: 159 products from 16 brands. Here is the entire program.

Several of the remaining names — Kirkland Signature, Member’s Mark, trunature — are warehouse-club private labels. The most trusted quality mark in American supplements is, in practice, a compliance product bought by a handful of mass retailers to stand behind their own store brands.

3. What happens when you check real products

We took 730 supplement products across 116 categories and 171 brands, and asked one question of each: can this exact product be found in a certifier’s registry?

135 of the 171 brands we track (79%) appear in no certification registry at all — not one product, not one program.

4. The halo problem: a certified brand is not a certified product

The middle bar is the one shoppers are most likely to misread. 137 products (19%) come from brands that do hold certification — on something else. A brand certifies its creatine and its whey, because a sports retailer requires it, and the reputation quietly extends across a catalogue of botanicals and vitamins that were never submitted to anyone.

Nothing on those bottles is false. But a shopper who has learned that a brand is “a certified brand” will draw a conclusion about the bottle in their hand that the registry does not support.

5. Claims of testing that lead nowhere

127 of the 730 products advertise third-party testing, lab verification, or a certificate of analysis without naming a certifier. Of those, 117 (92%) could not be located in any registry we compiled.

Language like “third-party tested” is unregulated. It may describe a genuine contract-lab assay, which is a real thing a responsible brand does and a real cost it absorbs. It may also describe nothing in particular. The distinguishing feature of a certification program is that it leaves a public record a shopper can check — and for 92% of these products, no such record exists.

The difficulty extends to products naming a specific mark. 57 products display a named certification mark on the listing. We confirmed 31 in the corresponding registry; 26 (46%) we could not.

We are deliberately careful with that figure. A failure to confirm is not evidence of a false claim. Certifications lapse, registries lag behind label printing, and a certified formula is often listed under a different size or flavour than the one on sale. The finding is not that these brands are lying. It is narrower and, for a shopper, worse: doing exactly what the advice columns recommend — looking the product up in the certifier’s own database — fails about half the time.

6. Whole categories with no certified option at any price

77 of 116 categories (66%) contain no certified product at all among the products we track. In these aisles, paying more cannot buy verification, because verification is not on sale.

• L Glutamine (14)
• Saw Palmetto (14)
• Boswellia (13)
• Elderberry (13)
• Msm (13)
• Garlic (12)
• Cordyceps (11)
• Ginkgo (11)
• L Tyrosine (11)
• L Citrulline (10)
• Spirulina (10)
• Ginseng (9)
• Manganese (9)
• Saffron (9)
• Taurine (9)
• Glycine (8)
• Colostrum (7)
• Probiotics (7)
• Betaine Tmg (6)
• Bromelain (6)
• Carnosine (6)
• Citicoline (6)
• Gaba (6)
• Grape Seed Extract (6)
• Hmb (6)
• Iodine (6)
• Krill Oil (6)
• L Arginine (6)
• Lutein Zeaxanthin (6)
• Mct Oil (6)
• Nmn (6)
• Phosphatidylserine (6)
• Pine Bark Extract (6)
• Tocotrienols (6)
• Tryptophan (6)
• Vitamin A (6)
• Algal Omega 3 (5)
• Alpha Gpc (5)
• Boron (5)
• Branched Chain Amino Acids (5)
• Cla (5)
• Cod Liver Oil (5)
• Coq10 (5)
• Digestive Enzymes (5)
• Fisetin (5)
• Glucomannan (5)
• Hyaluronic Acid (5)
• Inulin (5)
• Lactoferrin (5)
• Lysine (5)
• Vitamin K2 (5)
• Ashwagandha (4)
• Astaxanthin (4)
• Beta Carotene (4)
• Calcium Citrate (4)
• Choline (4)
• D Mannose (4)
• Lions Mane (4)
• Pqq (4)
• Shilajit (4)
• Spermidine (4)
• Vitamin K (4)
• Berberine (3)
• L Theanine (3)
• Maca (3)
• Methylfolate (3)
• Nac (3)
• Potassium (3)
• Same (3)
• Sea Moss (3)
• Chondroitin (2)
• Glucosamine (2)
• Nr (2)
• Selenium (2)
• Urolithin A (2)
• Zinc Lozenge (2)
• Magnesium Oxide (1)

7. Certification tells you the label is true. It does not tell you the dose works.

Everything above measures one failure: a label nobody has checked. There is a second failure, and it is the one shoppers actually worry about. A label can be perfectly accurate about a dose too small to do anything.

These are independent problems. A certifier confirms that the 100 mg on the bottle is really 100 mg. It does not ask whether the clinical trials used 500 mg. So we tested the two against each other: for the 205 products where we have both a registry status and a clinically studied dose range, does certification predict a useful dose?

Read this one carefully, because it is easy to over-read. Certified products met the studied dose less often here — but with only 28 certified products to evaluate, that gap does not reach statistical significance (Fisher exact, p = 0.185). We are not claiming certified supplements are worse dosed. The finding is the null result, and it is the more useful one: certification carries no information about whether the dose is effective. The seal and the dose are unrelated properties.

Which produces the number that matters most in this report:

Only 16 of 205 products (8%) are both independently verified and dosed at the level the research actually used. The two things a shopper wants from a supplement almost never arrive in the same bottle.

And paying more does not fix it. The more expensive half of these products met the studied dose 67% of the time; the cheaper half, 72%. Price predicts neither verification nor dose. Both of the heuristics a shopper is handed — look for a seal and buy the premium one — are solving at most half the problem, and frequently neither half.

8. So is the bottle over-dosing you, under-dosing you, or neither?

This is the question underneath everything else. It has three answers, because three different things can go wrong. Two are smaller than the headlines suggest. The third is the one almost nobody names.

Over-dosing: essentially not happening

Nine ingredients we track have a defined supplemental upper limit — the level above which routine daily intake stops being sensible. Among the 39 products in those categories, 4 exceed it, the worst by 1.14×.

Three magnesium products at 400 mg against a 350 mg limit — and that limit exists because supplemental magnesium loosens the bowels, not because it is toxic. On this evidence the mass-market supplement aisle is not overdosing anyone. Worth saying plainly, because it is the fear the category usually attracts and our data does not support it.

Under-dosing: real, and roughly a third

Of the 205 products comparable against a clinically studied dose, 63 (31%) sit below the low end of the range the research used. Not dangerous. Just unlikely to reproduce the effect that made anyone want the ingredient.

The real answer: for a lot of products, nobody can tell you

In many categories the thing on the label and the thing in the studies are different quantities — a chelated mineral, an oil, a standardised extract. “1,000 mg magnesium citrate” is the weight of the whole compound; the magnesium is a fraction of it. A krill softgel measured in milligrams of oil delivers a much smaller number of milligrams of the omega-3 actually studied.

So we looked for the 18 categories where the active amount is provably a real, disclosable number — proven because at least one competitor in that category prints it on the label. In those categories a brand's silence is a choice, not a quirk of the ingredient. Across the 100 products in them:

In nearly half of these products you cannot work out the dose of the studied substance at all. Not because it is hidden in fine print — because the number is not printed anywhere. The shopper cannot know it, we cannot compute it, and no certifier reports it.

If you want to know what a specific bottle actually declares, look it up in the NIH label database — 207,000 labels, exactly as filed. If the active amount is not there, it is not anywhere.

Why the honest brands look worst on paper

There is a trap in measuring this, and it is worth showing because it inverts the obvious conclusion. You can only calculate how far a label number sits from the active dose when a brand discloses both. So the products with the most alarming-looking gap are, necessarily, the transparent ones.

The widest gap in our catalogue is a spermidine supplement whose label leads with 800 mg of wheat germ extract and states that it delivers 1 mg of spermidine. That is an 800-fold difference between the two numbers, and it is on the bottle precisely because the brand printed both. It is the most informative label in its category.

Its competitors are the problem. One states 1,500 mg of wheat germ extract and an “equivalent of 15% spermidine” that is wildly out of line with the well-under-1% spermidine content of wheat germ. Another lists roughly 998 mg of extract blend and no spermidine figure at all. Neither appears in any gap calculation, because neither gives you a second number to compare. Punishing the disclosed ratio would mean rewarding the brands that told you nothing.

Which is where the whole report closes on itself. Not one of the products with a 2×-or-worse disclosed gap is certified — and certification would not have helped if it were. A certifier confirms that a label is accurate. These labels are accurate; the silent ones are accurate too, because omitting the active amount breaks no rule. The most consequential way a supplement leaves a shopper in the dark is fully legal, fully compliant, and structurally invisible to the entire verification system this report has just finished measuring.

9. Verification is allocated almost exactly opposite to need

If certification existed to protect confused shoppers, you would expect to find it where confusion is highest: the new, expensive, heavily marketed ingredients nobody can evaluate. It is the other way around.

Every legacy vitamin category we track has at least one certified option. 14 of 17 trend ingredients have none at all — not one certified product at any price:

• Astaxanthin
• Berberine
• Colostrum
• Cordyceps
• Fisetin
• Lions Mane
• Nmn
• Nr
• Pqq
• Saffron
• Sea Moss
• Shilajit
• Spermidine
• Urolithin A

Vitamin C is a century old, costs pennies, and has a certified option. Urolithin A is new, costs a great deal, makes ambitious claims about cellular ageing, and has none. The verification infrastructure covers the aisle where a shopper is least likely to be misled and is entirely absent from the aisle where they are most likely to be.

That is not a conspiracy. It follows mechanically from section 9: certification is bought by mass retailers for commodity private labels and by sports brands for banned-substance liability. Neither of those buyers operates in the longevity aisle — so nobody there is paying for verification, and nobody is required to.

10. Why the industry looks like this

The pattern above is not an accident, and it is not an enforcement failure. It follows from how the market is built.

• No premarket approval. Under the Dietary Supplement Health and Education Act of 1994, a supplement needs no FDA sign-off before sale, and no mandatory test for identity, potency or label accuracy.
• Certification is voluntary and costs money — per product, with facility audits and per-lot retesting.
• Shoppers cannot price the signal. A certified product sits beside an uncertified one that looks identical and costs less. The certified brand carries the cost and captures little of the benefit.
• So certification is bought where a gatekeeper compels it. Professional and college sports require banned-substance certification, which is why the sports programs dominate the listing counts. Mass retailers underwrite verification for their own private labels, which is why warehouse-club brands fill out the USP program.

That is the shape of the whole finding: third-party certification was never built as a consumer-quality signal. It is a compliance product, sold to athletes and to retailers. The vitamin aisle is not under-certified because anyone is cutting corners. It is under-certified because nobody standing in it is required to be.

11. What to do with this, if you take supplements

None of this is an argument for not taking a supplement you have a reason to take. A vitamin D deficiency does not become less real because the certification system is small. The argument is narrower: the two shortcuts people are handed — look for a seal, buy the expensive one — carry much less information than they appear to. What follows actually works.

• Prefer a certified product when one exists — it just usually will not. The seal is a real check on label accuracy and worth having. In 77 of our 116 categories there is nothing to prefer.
• Do not read a missing seal as a warning. 92.5% of products lack one. At that prevalence, absence carries almost no signal about an individual product.
• Check the dose against the studied range yourself. This is the part certification never covers, and it takes a minute. Easiest route: put the supplement you already take through our free checker, which compares its form, dose and cost against the research. To do it by hand, our dosage guides list the studied range for each ingredient.
• Watch compound weight versus elemental dose. “1,000 mg magnesium citrate” is not 1,000 mg of magnesium — the elemental figure is a fraction of it. The same trap runs through iron, calcium, and zinc, and it is the single most common way an honest label still misleads.
• Be most skeptical exactly where certification is absent. The newest, most-marketed ingredients are the ones with no verification infrastructure at all, and they are also the ones making the largest claims.

12. What this analysis does not show

The limits matter as much as the findings, so they are stated plainly rather than buried.

• This is not a safety finding. We tested nothing. An uncertified supplement is not thereby contaminated, underdosed or dangerous. The claim is narrow and entirely about the public record.
• Absence from a registry is not absence of testing. Brands run private contract-lab testing that no public registry records. This measures verifiability, not whether testing happened.
• The certified count is an over-estimate, on purpose. Our matching rule is deliberately generous: a product counts as certified when any plausible brand-and-product match exists, including different flavours and doses in the same line. A stricter rule returns fewer certified products and a starker headline. We chose the rule least favourable to our own conclusion.
• Three programs, not all of them. We ingested NSF, USP and Informed Sport/Choice. IFOS (fish oil) and BSCG were not ingested; their product lists are not machine-readable. 4 products in our catalogue advertise IFOS. Counting every one of them as certified would move the headline from 7.5% to 7.9% — a difference of 0.4 percentage points, which does not change the finding.
• The dose comparison covers part of the catalogue, and is under-powered. A clinically studied dose range exists for 205 of the 730 products; the rest are in categories with no single defensible “effective dose” to compare against. Only 28 of those are certified, which is why the certified-versus-uncertified gap is reported as a null result rather than a difference. The 8% overlap figure is descriptive of that subset, not an estimate for the whole market.
• The disclosure finding covers 100 products in 18 categories, not the whole catalogue. A category only qualifies when a competitor proves the active amount is disclosable there, which deliberately excludes pure-powder categories where the label already is the active dose. That keeps the 48% honest, but it is a statement about those 18 categories rather than about supplements in general.
• Non-disclosure is not evidence of a bad product. A brand that does not print an active amount may still be selling a well-made supplement. What we can say is narrower: nobody outside the company can check.
• “Meets the studied dose” is a floor, not a verdict. We test whether a serving reaches the low end of the range used in published trials. It does not account for bioavailability differences between forms, or for whether the evidence behind that range is strong in the first place.
• The catalogue is not a random sample. It over-represents large, established brands, which are the brands most likely to certify. A representative sample of the market would almost certainly show less certification, not more.

Methodology

1. Registry capture. Complete public listings downloaded for NSF/ANSI 173 Dietary Supplements (finished products; NSF’s listing page states it is current as of Tuesday, August 25, 2026), USP Verified dietary supplements, and Informed Sport/Informed Choice. Combined: 5,317 listings. Raw captures are kept unmodified for provenance.
2. Catalogue. 730 supplement products across 116 categories and 171 brands, each with a live retail listing at the time of capture. Selection is editorial: the leading options per category. It is not a sales-ranked sample, and we make no best-seller claim.
3. Brand matching. Brand strings are normalised, corporate suffixes removed, and “dba” trading names resolved. Descriptive words are deliberately not stripped — collapsing “Nature’s Lab” to “Nature’s” would falsely match every other Nature’s-prefixed brand. Brand strings too generic to resolve safely are excluded rather than guessed.
4. Product matching. Brand tokens are removed from both names before comparison; without that step a product matches any other product from the same brand. A match requires at least two shared distinctive tokens, or one plus an identical dose, at 60% overlap or better.
5. Classification. Each product receives exactly one status: certified, brand-certifies-but-not-this-product, or brand-absent.
6. Reproducibility. The registry captures and the audit script ship in the repository; re-running the script regenerates every number on this page. All source registries are public and free to query.

Download the full audit (CSV) · JSON · all 730 products with brand, category, listing claim, registry status, and the matched registry listing where one exists.

Frequently asked

Related

• What supplement listings disclose about certification
• The same supplement can cost far more per day depending on brand
• Supplement statistics (2026)
• The open dataset