How we test and audit
Last updated October 21, 2026
Every product we review gets a research audit of its label, its doses and the studies behind them, and our ratings rest on that audit. Where we've also taken a product ourselves, the review says so. A lot of supplement sites blur that line. We think the distinction is the whole game, so we use two different words. An audit is what we did to the label and the evidence. Testing is what we did with our own mornings, and we won't claim it where it hasn't happened.
What every product audit involves
Every audit starts with the Supplement Facts panel. We record each printed amount, compare it with the study dose we use as that ingredient's benchmark, and grade that research. We check the sales page's claims against the studies and the brand's own documents, and we check whatever lab reports the brand publishes.
The audit also records the things brands would rather leave vague, such as whether "choline 250 mg" means 250 mg of choline or 250 mg of choline bitartrate that supplies about 100 mg of choline, whether a money-back guarantee quietly deducts a handling fee, and whether "third-party tested" comes with a published certificate of analysis or just a sentence on a sales page. When we can't find anything to check a claim against, the review calls it Unverified, not false.
A dose audit is more than a number on a label
These are two entries from our Vyvamind review. The pale band shows the amounts used in the studies. The dot is the amount in one serving.
citicoline sodium · Preliminary evidence
Doses used in studies that showed a benefit
Vyvamind has about 191 mg (200 mg citicoline sodium on the label). The study dose we use as this ingredient's benchmark is 250 mg, so it's 59 mg short.
Study dose benchmark from McGlade 2012, 250 or 500 mg a day, 28 days
Moderate evidence
Doses used in studies that showed a benefit
Vyvamind has 150 mg, inside the 100 mg to 400 mg used in studies that showed a benefit.
Study dose benchmark from Dassanayake 2023, One dose of 100, 200 or 400 mg
The range matters, but the study question matters too. A dose can look adequate on a graph and still be irrelevant to the result a reader wants. The full context and source links are in the Vyvamind dose audit.
What an audit can and can't establish
A label audit can tell you whether a formula claims doses that match the research. It can't confirm what's inside the capsules or tell you how they'll affect you.
It can also tell you whether a claim has a source behind it and what a product cost on the date we checked. Only a lab report can show what's actually in the capsules, which is why every review grades the testing separately, from Fully verified down to Unverified.
How we'll conduct first-hand testing
We'll buy every product we test at retail, the way you would. We won't accept free samples, and if that ever changes, the review will say so. Before the first dose, we'll record the batch number, the expiry date, the price we paid and the label's directions.
Each day we'll log the dose and when we took it, how we slept, how much caffeine we had, and notes on concentration, mood and any side effects. Those notes get written as we go, not reconstructed at review time.
A test will run as long as the trials behind the product's main ingredients did, not a flat 30 days. The main bacopa and citicoline memory trials ran for 12 weeks, so a product built on either gets 12 weeks.
Before any test starts, a doctor or pharmacist will check the tester's medications. We'll stop a test early if side effects appear, and we won't start one if any ingredient could interact with a medication the tester takes. Each test will involve one or two people.
Negative and inconclusive results will be published with the same prominence as positive ones. If we felt nothing, the review will say we felt nothing.
What first-hand testing can and can't prove
An audit can't tell you whether the capsules smell like a pond, whether the "no crash" claim survives a 3 p.m. meeting, or whether week three feels different from week one. That's what testing is for.
One or two people without a control group are open to placebo effects and to ordinary good and bad weeks. So what we report from testing is experience, not evidence. It can describe what taking a product was like, but it never establishes that the product works.
How this affects our ratings
Hands-on results stay out of the audit rating. They'll appear beside it in the review, so a score means the same thing whether we've taken the product or not. We're researchers, not doctors, and nothing here is medical advice.