Evidently

Methodology

The supplement industry runs on confident claims and weak data. Evidently exists to do the opposite: rate each supplement against the actual scientific literature, surface what trustworthy experts say, and point to brands that have earned it. Our guiding principle is simple — improve health and well-being, and reduce the noise.

The Evidently score

Every supplement gets a single score out of 10, built from three questions asked in a fixed order:

  1. Step 1

    What is it acting on?

    We name the mechanistic target — creatine on the phosphocreatine/ATP system, caffeine on adenosine receptors, psyllium on bile-acid binding. If a supplement has no plausible human mechanism, that is a warning sign before we even look at trials. The mechanism informs the review but does not add points on its own: plenty of things work in a petri dish and fail in people.

  2. Step 2

    How well do human RCTs support it?

    The RCT grade (0–10) is based only on randomised controlled trials in humans and meta-analyses of them. Animal, cell, and observational data can raise a question but never raise the grade. Every supplement page cites the trials we relied on.

    9–10Many large human RCTs plus meta-analyses; consistent, clinically meaningful effect.
    7–8.9Multiple human RCTs with a reliable effect, some inconsistency or population limits.
    5–6.9A few RCTs, small or mixed effects, or benefit confined to a subgroup (e.g. deficient people).
    3–4.9Mostly small, short, or poor-quality trials; promising but unproven.
    0–2.9No supportive human RCTs, or well-run RCTs that found nothing.
  3. Step 3

    What could it cost you?

    Risk is applied as a multiplier, not a footnote. A supplement with documented harms loses a large share of its score even if it "works" — because the goal is your health, not a product's efficacy in isolation. This is why caffeine grades 9.5 on RCTs but scores 6.7 overall.

    Minimal risk×1.0

    Long-term human safety data; side effects are rare, mild, and reversible.

    Low risk×0.9

    Mild, dose-dependent side effects or narrow interaction concerns; safe at evidence-based doses.

    Moderate risk×0.7

    Documented meaningful harms (organ toxicity reports, dependence, sleep disruption, or serious interactions).

    High risk×0.4

    Serious or frequent adverse events; harms may outweigh benefits for most people.

Formula
Evidently score = RCT grade × risk multiplier

Example: creatine 9.5 × 1.0 = 9.5 · caffeine 9.5 × 0.7 = 6.7 · ashwagandha 6.5 × 0.7 = 4.6

Evidence levels

Alongside the numeric grade, each supplement carries one of four qualitative ratings summarising the state of the human trial literature.

Strong evidence

Multiple high-quality RCTs and meta-analyses converge on a clear effect.

Moderate evidence

Reasonable RCT or meta-analytic support, with some inconsistency or limitations.

Limited evidence

Mostly small trials, observational data, or mechanistic reasoning.

Insufficient evidence

Claims outpace the data; primarily preclinical or anecdotal support.

How we choose experts

We deliberately mix mainstream evidence-based communicators (Attia, Patrick, Huberman, Galpin, Examine.com) with respected skeptics (Layne Norton, Nick Tiller, Stuart Phillips, Kevin Bass). Both perspectives matter — popular experts often surface useful tools, while critics keep the field honest about effect sizes and study quality.

How we choose brands

We weight three things: independent third-party testing (NSF, USP, IFOS, public CoAs), formulation quality (clinically-used forms and doses, no megadosing or junk fillers), and fair pricing. We recommend across budget, moderate, and premium tiers — high price doesn't equal high quality.

Important

Evidently is educational. Nothing here is medical advice. Speak with a qualified clinician before starting, stopping, or combining any supplement — especially if you have a medical condition or take prescription medication.