How it works — Dirty Scent

Two tracks, kept separate

Disclosure is a set of independent boolean tests against the published ingredient list or a business's stated policy. Each one cites the exact ingredient or evidence that triggered it.

Reports are what members observed, shown as counts and dates. Four members, most recent in March — not a rating out of five.

We don't average them together. A product whose label looks clean but that three people reacted to isn't a contradiction to smooth over. It's the most important thing on the page.

Positive and negative findings are not symmetric

This is the idea everything else follows from. Finding an allergen on a label is proof it's there. Not finding one is proof of absence only if the list is complete and every term is specific.

So there are three answers, not two.

State	What it means
Contains	Named on the label. Certain — survives incomplete lists, generic terms, everything.
Does not contain	Only when the list is on file, every term resolved, and no generic term present.
Cannot determine	A generic term is present, or the list is missing or unreadable. Not a softer version of "does not contain."

What `Parfum` does to a label

A single word on an ingredient list can stand in for a mixture of hundreds of compounds. When we see one, we cannot tell you the product is free of linalool, or of anything else undisclosed.

So a generic term doesn't attach a warning to a clean result. It removes the clean result. We are not entitled to that claim, so we don't make it.

How list completeness and generic terms affect what can be reported
List complete	Generic term	Allergen named	What we can report
yes	no	yes	Contains — certain
yes	no	no	Does not contain — no substance we track is disclosed
yes	yes	yes	Contains — certain
yes	yes	no	Cannot determine — voided by generic term
no	—	yes	Contains — certain
no	—	no	Cannot determine — list incomplete

Only two of those six rows let us tell you something is absent.

Why a clean label is weak evidence

We don't say "fragrance-free."

A label lists ingredients. It does not tell you what a product releases into the air, and those are different questions. Disclosure rules were written for other purposes and other exposure routes, and they carry concentration cut-offs — a great deal of what matters for something you breathe isn't required to be named at all.

So the strongest thing even a complete, fully specific list supports is: no substance we track is disclosed on this label. That is not the same as unscented, and we don't present it that way.

What we track, and why

The substances we look for aren't taken from any one regulator's list. A legal list records what somebody must label — which reflects jurisdiction, timing and lobbying as much as it reflects harm, and most of it was written about skin rather than lungs.

So every substance we track carries its own cited evidence, and you can read it. Where a substance also appears on a regulatory list, we record that as a fact about the substance — not as the reason we track it.

The list is community-sourced and evidence-gated: anyone can propose a substance, but we don't add one without sources anyone can check.

Natural is not a synonym for safe

Lavender oil, bergamot, citrus oils and "botanical blends" are fragrance ingredients. They're common triggers. We flag fragrance function, not fragrance origin, so a plant-derived fragrance ingredient is recorded exactly like a synthetic one.

Any system that treats "no synthetic fragrance" as clean is useless to the people this is for.

Facts stack. Nothing is scored.