The dose in the trial is the dose in the jar.
By the Paper sells doses, not products. Each formula is a claim: this ingredient, at this amount, produces this measured outcome in humans. The claim ships with the trials that support it, the certificate of analysis for the lot in your hand, and the paperwork from the supplier who made the raw material. If any of those three is missing, the product does not ship.
A constraint cannot be traded off.
| No. | Rule | What it means in practice |
|---|---|---|
| 1 | Dosed as studied. | If human trials did not use the ingredient at this dose for this outcome, it is not in the formula. Not a smaller amount, not a “hint”. |
| 2 | One dossier per ingredient. | Outcome, dose range in the literature, our dose, evidence grade, the trials, the supplier, the specification, the lot tests. No dossier, no ingredient. |
| 3 | Nothing that isn’t doing the job. | No emulsifiers, flavor systems, sweeteners, colors, flow agents or fillers. Where a format needs a carrier (a serum needs a solvent), it is listed by name and justified in the dossier. |
| 4 | Function decides the format. | Protein without lecithin clumps in a shaker. The label says “blend it” instead of the tub containing lecithin. The trade-off is the customer’s, and we say so. |
| 5 | Flavor is last, and usually never. | Unflavored is the default SKU. Where a flavor is offered it is one sourced ingredient, with its own dossier line, as a separate SKU. |
| 6 | Formulas are versioned. | Whey 1.0 becomes 1.1 when anything changes. The change, the reason and the evidence are public. Old versions stay readable. |
| 7 | Every lot is tested, and the numbers are published. | The lot number on the package resolves to the certificate of analysis for that lot. Not a representative certificate. That lot. |
| 8 | Anyone can dispute a formula. | A public page: cite the paper we missed. Accepted disputes credit the submitter by name and bump the version. Rejected disputes are published with the reasoning. |
| 9 | We say what we don’t sell. | When a category does not clear the evidence bar, we publish the review and we do not sell the product. The absence is a claim too. |
Numbers before adjectives. If a sentence has no unit in it, ask what it is for.
Grades are a promise about the kind of evidence, not about how well it works. A and B ship. C is published and not sold. F is published as a warning.
| Grade | Means | Ships |
|---|---|---|
| A | Multiple RCTs or meta-analyses in humans, at this dose, for this outcome. | Yes |
| B | At least one good RCT at this dose, or consistent RCTs with heterogeneity in effect. | Yes, grade printed on label |
| C | Mechanistic, animal, or small uncontrolled trials only. | No. Review published. |
| F | Evidence of no effect, or of harm, at the doses sold elsewhere. | No. Review published. |