How we grade evidence.
Every ingredient claim on this site carries a letter grade. A grade nobody can audit is worth less than no grade at all, so here are the exact bands, applied the same way every time.
Strong evidence
Five or more independent studies, 500+ combined participants, consistent positive results, including randomised controlled trials or meta-analyses.
Moderate evidence
Two to four studies with adequate sample sizes. Positive overall, but less consistent than grade A.
Limited evidence
One or two studies, or small sample sizes. Promising but inconclusive, and a larger trial could reasonably overturn it.
Weak evidence
Mechanistic or theoretical support only, animal studies, or human results that contradict each other.
Not supported
The majority of available evidence shows no effect, or shows the opposite of the marketed claim.
Graded per claim, not per ingredient
Grades are assigned to a specific claimed benefit, never to an ingredient as a whole. This is the single most important rule in the method. An ingredient is not simply “good” or “unproven” — it has been studied for particular outcomes, and the evidence for one can be strong while the evidence for another is absent.
Glucomannan is the clearest case on this site. As a bulk-forming fibre it has acceptable support for regularity. For weight loss, which is what almost every product containing it is sold for, the NIH Office of Dietary Supplements concludes it has little to no effect. One averaged grade would flatten that into something reassuring and useless.
What a grade does not tell you
A grade measures the state of the published evidence for a claim. It does not measure whether a particular product contains a useful amount, whether the form used matches the form studied, or whether you personally need it. A grade A ingredient in a product that contains a tenth of the studied dose is still a poor purchase, which is why the label overviews check the panel separately.
It is also not clinical advice. Evidence strength in a trial population says nothing about your own medication, blood levels or history. See the medical disclaimer.
Why some grades are low on purpose
A grading system where nothing scores below B is not rigorous, it is marketing with letters attached. Applying these bands honestly means some widely sold ingredients grade D or F for their headline use, and publishing those is the point of having a method rather than an opinion.
Scoring a product, not an ingredient
A letter grade describes research. It cannot tell you whether a particular tub is worth buying, because that also depends on how much active ingredient a serving contains, what the label discloses and what you are paying for it. Reviews therefore carry a separate score out of 10, built from five criteria applied the same way every time.
Each criterion is scored out of 10 and the overall figure is their mean, printed as a breakdown on every scored review so the arithmetic can be checked. Label overviews carry no score at all: they record what a manufacturer published and are not assessments.
Evidence for the marketed claim
How strong the research is for what the product is actually sold to do — not for the ingredient in general.
Scores 10: The headline claim carries an A-grade body of evidence in the studied population.Scores 0: The headline claim is unsupported, or the supporting research is for a different outcome.
Dose against the studied amount
Whether one serving delivers the amount used in the research behind the claim.
Scores 10: A serving sits inside the studied range without taking extra scoops.Scores 0: A serving is a fraction of the studied amount, or the amount is not disclosed at all.
Label transparency
Whether the Supplement Facts panel is published in readable text and every active amount is disclosed.
Scores 10: A readable panel on the manufacturer’s own site, every active quantified.Scores 0: No readable panel, or a proprietary blend hiding the individual amounts.
Value against the generic equivalent
Cost per gram or per dose of the active ingredient, compared with the plainest version of the same thing.
Scores 10: At or near the cost of the generic equivalent.Scores 0: A large multiple of the generic price with no functional difference to show for it.
Safety and tolerability
Known risks and interactions, how well characterised they are, and whether the label warns about them.
Scores 10: A well-characterised ingredient with clearly communicated precautions.Scores 0: Poorly characterised risks, known interactions, or precautions the label omits.
The five are weighted equally on purpose. A product whose active ingredient works, but which does not publish a readable panel and charges a large premium over the generic equivalent, should not be rescued by its chemistry.
When grades change
Grades are revisited when a relevant trial or systematic review is published, and every ingredient page shows when it was last updated. A change that moves a grade is described rather than made silently — see how pages are produced and corrected.
Common questions
How is an evidence grade assigned?
Against the published bands on this page: the number of independent studies, the combined participant count, how consistent the results are, and whether randomised controlled trials or meta-analyses exist. A grade is assigned per claimed benefit, not per ingredient.
Why does one ingredient have several different grades?
Because evidence strength varies by use. Glucomannan has reasonable support as a fibre for regularity and effectively none for weight loss. Publishing one averaged grade would hide the difference that matters to someone deciding what to buy.
Does a high grade mean I should take it?
No. A grade describes the strength of the evidence for a claim in the studied population. It says nothing about whether you personally need it, whether it interacts with your medication, or whether your levels are already adequate. That is a question for a clinician.
How is a product score different from an evidence grade?
An evidence grade describes the published research for one claim, so it belongs to the ingredient rather than to any brand. A product score out of 10 judges a specific product: whether this tub, at this price, with this label, is a sensible purchase. A grade A ingredient can sit inside a product that scores poorly, which is the whole reason the two are kept separate.
Why are the five scoring criteria weighted equally?
Because a product that hides its Supplement Facts panel and charges a large premium should not be rescued by its chemistry. Weighting evidence above everything else would produce high scores for expensive, opaque products built on well-studied ingredients, which is the outcome the method exists to avoid. The overall score is the mean of the five, so it can always be checked against the breakdown printed on the page.
Do grades change?
Yes. A new trial can strengthen or weaken a grade, and pages carry a last-updated date so you can see when the assessment was last examined. A grade that has not been revisited in years is a reason to check the primary sources yourself.