How we grade evidence
We grade each compound one outcome at a time, because a compound can help one thing and do nothing for another. This is the rubric every grade on the site follows. It is versioned, and commercial relationships can never change a grade.
The grades
- A, strong
- At least one meta-analysis or three independent randomized trials in the relevant healthy population, pointing the same way, with low or some risk of bias, at doses legally available.
- B, moderate
- Two or more randomized trials with mostly consistent results, or one large, well-run trial, or a meta-analysis with high heterogeneity.
- C, limited
- One small randomized trial, open-label or pilot studies, or results only in clinical or older populations being applied to healthy adults.
- D, preliminary
- Animal, cell or mechanism evidence only. No controlled human outcome data.
- N, no benefit shown
- Adequate human randomized trials found no meaningful effect.
The ring fills with the grade: a solid disc for A, a hollow ring for N. The hollow ring stands for the placebo; the solid disc for the active compound.
Safety is a flag, not a grade
A separate amber flag marks a credible safety signal, an interaction, a regulatory warning, or a quality or adulteration problem. It never raises or lowers a grade, and a grade never hides a flag.
Modifiers
Shown next to a grade when they apply:
- Industry-funded: most supporting trials were funded by a company that sells the compound.
- Dose gap: typical products contain less than the doses that showed an effect.
- Short-term only: trials lasted days or weeks.
- Single research group: most of the evidence comes from one team.
- Subjective outcome only: the effect was self-reported, not measured.
How a grade is set
Every grade stores a short rationale, the studies used, the search string and date, the grader, the editor’s sign-off, the date and the previous grade. Risk of bias for each study uses a simple published checklist: randomization, blinding, attrition, pre-registration, selective reporting, and funding or conflicts of interest.
Grade history
Grades change only when new eligible evidence passes review. The grader proposes a change, the editor approves it against this rubric, and the change is logged with its reason and date. Every compound page links to its grade history.
Independence
Grade records cannot reference affiliate data, and the editor’s checklist confirms it. Our affiliate partner’s ingredients are graded by this same rubric. See how we make money.