Every product gets one of three verdicts: clinically dosed (label dose meets the amount used in trials), underdosed (disclosed, but below it), or undisclosed (the label doesn't let you verify). The rules below are applied identically to every product; no brand can pay to change them.
Doses come from the NIH Dietary Supplement Label Database (DSLD), the U.S. government's archive of supplement labels. Every product page links its exact DSLD label, so you can check our reading of it.
Each category is scored against one dose from published clinical research (e.g. creatine: 3 g/day, ISSN Position Stand 2017; ashwagandha: 300 mg/day at 5% withanolides, the most-studied KSM-66 protocol). The citation appears on every page in the category.
A blend's printed weight is the weight of the whole mixture, not of any ingredient in it. If a label hides the target ingredient inside a blend without its own amount, the product scores undisclosed — we never count a blend's total as an ingredient's dose. The one exception: a blend whose disclosed composition is entirely the target ingredient (e.g. a "creatine blend" of six creatine forms) counts at its stated total.
If a label discloses several forms of the ingredient with their own amounts (say creatine HCl 1 g + creatine citrate 1 g + creatine peptide 1 g), we sum them.
For botanicals (like ashwagandha), raw milligrams aren't comparable across products: a 10%-withanolide extract is ~2× as potent per mg as a 5% one, and plain root powder discloses no potency at all. We normalize each ingredient row by its own disclosed standardization before comparing to the threshold. Undisclosed-potency material can only ever downgrade a product to undisclosed — it never counts toward a pass.
Label databases lag reformulations, and extraction at this scale has edge cases. If we scored a label wrong, email [email protected] with the product and the DSLD link from its page; we fix errors and say so.
We score label transparency against studied doses. We do not lab-test products, verify label claims are true, or rank brands by quality. A "clinically dosed" verdict means the label discloses enough of the ingredient — not that the product is effective or right for you. Not medical advice.