How we check: the Fine Print Ledger
The Fine Print Ledger is how we turn scattered lab tests into one clear answer. Here's every step, so you can check our work.
1. Find every public test
We search for every published lab test of a product type: testing groups, supplement makers, journalists and creators who share lab reports. We record the date of each search on the page. We can miss tests that aren't public, such as members-only results, and we say when we know of one.
2. Check each number against its source
We read the original report wherever it's public, not a summary of it. We record who ran the test, which lab, the date, the product, the flavor and lot if given, the amount on the label, and the amount found. When sources disagree, we keep every result.
3. Rate how strong each test is
Not all tests are equal. For each one we note who paid for it, whether the lab is named and accredited, and what's missing, such as lot or flavor. Every article shows this in a "How strong is each test?" table.
- Independent: the brand being tested didn't pay for it. A competitor's test counts, but we say it's a competitor.
- Brand-paid: shown and marked, but never counted in our totals and never enough for a pick.
4. Count results one way
- Met the label: at least 90% of the labeled amount in every tested batch. We use 90% because small gaps come from normal batch and lab variation.
- Fell short: under 90% in at least one batch. "Mixed" means another batch did reach 90%.
- Almost none: a batch under 5% of the label.
- Not detected: the lab found less than its detection limit. None found: the lab reported zero without a published limit.
Each result describes the batch that was tested, not the brand forever. Products change, so every article shows test dates. No public test isn't the same as failing; we only list what was measured.
5. Work out the real cost
Cost per dose is today's price divided by the number of doses, using the amount the lab found. We count any result above 100% as 100%, so extra never earns a lower cost. We record the date we read each price.
6. Pick by a written rule
Each article states its rule before the pick. Usually: the product met its label in an independent test, it's sold in the US, and among those it costs the least per real dose. We say what we didn't verify and what would change the pick.
A worked example: our creatine pick
Here is how one product, BulkSupplements creatine monohydrate powder, goes through the steps.
| Step | BulkSupplements, 1 kg |
|---|---|
| 2. Check the source | SuppCo bought it and had a lab test it. Its June 2025 report lists 98.6% of the 5 g on the label. Our other pages round it to 99%. |
| 4. Count the result | 98.6% is at least 90%, so it met its label. |
| 5. Work out the real cost | $27.97 for 200 doses is $0.140 a dose. The lab found 98.6%, so a real 5 g costs $0.140 ÷ 0.986 = $0.142, about $4.26 for 30 days. Price read Oct 22, 2026. |
| 6. Pick by the rule | It met its label, it's sold in the US, and it is also NSF Certified for Sport. Among powders with both, it costs the least per real dose. See the pick. |
7. Re-check on a schedule
Every article has a review log with the date we last checked and the date we'll check again, at least every three months for picks. Lab tests older than two years are flagged as old. If something changes, we update the page and note it on the corrections page.
What this method can't do
We can only be as good as the tests that exist. A product with no public test may be fine or may not. Three groups' tests aren't a random sample of the whole market. We say this on every page, and we'd rather tell you what we don't know than guess.
Last updated .