How We Rate Products

What the score measures

The score answers one question: how well does this product stand up to a careful buyer's scrutiny? It is not a measure of how likely it is to work for you, because for most products in these categories nobody knows that, including the manufacturer. It combines four things.

What we weigh	What earns points
Evidence	Published human trials of the finished product rank highest, then trials of its ingredients at the doses actually used, then preliminary or animal work, then nothing.
Transparency	Ingredients named and amounts printed. A proprietary blend that hides its doses loses points, because it makes the evidence uncheckable.
Marketing honesty	Whether the claims match the evidence. Invented statistics, fabricated research and fake experts are the heaviest penalty on this page.
Commercial terms	A refund policy processed by a third-party retailer, clear pricing, and an upfront total. Anchor prices and undisclosed upsells lose points.

The scale runs from 3.5 to 5.0

3.5 is the floor. We do not publish a score below it. Every product we review has cleared a basic bar before it reaches this page: it is a real product, from a real seller, that ships and can be refunded. That is what 3.5 certifies, and it is the least we will say about anything we cover.

This means the number alone will not tell you everything. Where a product's conduct would rank lower on an open-ended scale, we do not express that as a smaller number. We state it in the verdict, in the product's own words, and we recommend reading it.

What each band means

4.5 to 5.0: strong evidence, fully transparent

Human trials support the finished product, doses are printed, and the marketing does not outrun the data. Nothing on this site currently scores here, and that is worth saying plainly: it reflects the state of these categories, not our generosity.

4.0 to 4.4: real product, disclosure good, marketing restrained

Ingredients are named with amounts, the claims stay close to what the evidence supports, and the commercial terms are clear. The evidence is still usually ingredient-level rather than product-level. This is the strongest band anything on this site currently occupies.

3.8 to 3.9: real product, evidence thin, mechanism overstated

The product exists, it ships, the refund route works, and the ingredients are plausible. Doses are often undisclosed, and the marketing usually overstates a real mechanism. That describes most of this market. A score here is not a recommendation. It means we found nothing dishonest, and nothing proven.

3.5 to 3.7: our floor, and a documented integrity problem

A product lands in the floor band for one of two reasons, and both are about conduct rather than efficacy:

• False claims. Invented statistics, research attributed to institutions that never produced it, fabricated experts, or safety absolutes such as "100% safe" and "no side effects".
• Risk to the reader. Marketing that discourages medical care, manufactures fear of a diagnosis, or targets a condition where delay causes harm.

A product can be perfectly real, ship promptly and refund reliably and still score here. The refund policy protects your money. It does not protect someone who skipped an appointment because a sales page told them their doctor was the problem.

Because the floor compresses this band, the distance between a product at 3.5 and one at 3.7 is smaller than the distance between what they did. Read the verdict, not just the number. Every review in this band names the specific claim, quoted from the seller's own page, that put it there.

Why there is a floor

We only review products we have vetted to a minimum standard: a real seller, a real product, and a refund route processed by a third-party retailer rather than at the seller's discretion. Anything that fails that standard does not get a review at all, so it never receives a score. The floor reflects that filter. It is not a statement that every product is fine, and the trade-off is stated openly above: a compressed band carries less information than an open scale, so the written verdict does the work the number cannot.

How the scale has changed

In July 2026 we re-scored every product against this rubric twice. The first pass corrected an inconsistency in our own scoring: reviews published earlier had scored around 3.6 to 3.7 for the same faults, undisclosed doses and an unproven mechanism, that were pulling newer reviews down to the mid-twos. The second pass introduced the 3.5 floor described above and shifted every band upward to keep them distinguishable. Several scores moved in each pass. Where a score changed, the review text explains why.

Our ratings are editorial opinion based on the published evidence and the seller's own materials. They are not medical advice, and no score is a substitute for talking to your doctor. Our full method is in our review principles and scientific review process.