Methodology
How InsideTheFormula gets to the numbers it shows.
How products get added
Products are added when they meet two tests: the brand publishes a full ingredient list with each active at an identifiable dose, and the product is something a real shopper would actually find when searching the category. "Proprietary blends" that hide per-ingredient doses are excluded — there's nothing to compare. New additions get prioritized by category gaps, reader requests, and obvious omissions in the current set.
What data sources are used
Per-product dose information comes from the manufacturer's public supplement-facts panel or active-ingredient list, with a sanity check against retailer listings to catch reformulations. Ingredient dose ranges and side-effect notes come from primary literature read informally — randomized trials, meta-analyses, and position stands from professional bodies — supplemented by manufacturer technical documentation where it exists. The ingredient pages summarize what the human research consistently shows; they don't exhaustively cite every paper. When the evidence is thin or conflicting, the ingredient page says so directly.
How we assess dosing
For each ingredient we record an "effective minimum" — the dose at or above which human studies consistently show the benefit the ingredient is taken for. The figure is drawn from randomized trials, meta-analyses, and position stands, with preference given to doses that match real-world use rather than outlier protocols.
On a product page, every ingredient gets one of three labels. "Clinically dosed" means the product hits or exceeds the effective minimum. "Underdosed" means the product carries the ingredient but at a dose below what the human research establishes as effective — a common gap in proprietary-blend-adjacent formulations and label-window products. "No data" means we don't yet have an effective-dose figure for that ingredient, or the labeled units don't match what the studies measured.
A product's overall verdict aggregates these per-ingredient calls. "Fully dosed" means every assessed ingredient meets its effective minimum. A mixed verdict ("3 of 5 clinically dosed") calls out the count so a reader can decide whether the underdosed ingredient is the one they cared about. The judgment is deliberately about labeled dose, not about absorption, bioavailability quirks, or stacking effects — those belong on the ingredient page.
How the comparison ranks and sorts
Default sorts are alphabetical or by category. There is no hidden ranking signal that boosts a product because it has an affiliate payout — pricing and dose are shown directly so a reader can sort on what they care about. When comparison features mature, the ranking inputs will be documented here in the same place.
Corrections
Mistakes are inevitable on a site that publishes specific numbers. If a dose is wrong, a price is stale, an ingredient is mis-classified, or a study has been superseded, send a note and it gets fixed. Pages carry a visible "Last updated" date that reflects the most recent content change, so readers can judge how fresh the information is.