The dose makes the poison: why an ingredient is not a verdict

Hazard asks: under some set of conditions, can this substance cause harm? Risk asks a narrower and more useful question: given the actual concentration, the actual exposure route, and the actual frequency of use, how likely is harm here? Chlorine gas is hazardous. The trace chlorine in tap water is not something you assess the same way, because the dose and exposure are entirely different.

A hazard-based scoring method skips the second question. It flags an ingredient for the worst thing it has ever been associated with at any concentration, in any study, on any route of exposure, and applies that flag uniformly whether the ingredient sits at 0.01% or 10%. That approach is easy to build and easy to market, because a red flag is simpler than a paragraph. It is also why it misleads: it cannot tell a reader whether a real product poses a real risk, only that a word appears on an ingredient list.

Cosmetic labeling rules require INCI ingredients to be listed in descending order of concentration down to 1%, after which order can vary. That single rule is the closest thing a shopper gets to a concentration reading without lab equipment, and it is the backbone of formulation-based scoring.

An ingredient near the top of the list, just after water and a few emollients, is likely doing real work at a real percentage. The same ingredient appearing eighth on a fifteen-item list is probably present at a fraction of a percent, often as a preservative, chelator, or pH adjuster included at the minimum level needed to do its job. Phenoxyethanol is a clean example: the EU's Scientific Committee on Consumer Safety concluded it is acceptable as a preservative up to 1.0% (see our phenoxyethanol dossier), and in practice it is formulated at or below that ceiling, not as a headline active. Position on the list is consistent with that use.

We treat the point where an ingredient list stops requiring strict descending order (commonly around 1%) as a rough dividing line between "likely a functional ingredient at a meaningful percentage" and "likely a minor additive." Below that line, an ingredient's theoretical hazard matters less, because the achievable exposure is small.

We are direct about the catch: brands almost never disclose exact percentages, and INCI order is a legal requirement, not a certified concentration figure. Two ingredients can be genuinely tied in percentage and listed in either order. So position is an estimate we use to weight a score, not a number we report as fact. When we cannot tell whether an ingredient sits above or below that line, we say so rather than guess with false precision.

Concentration is only half of exposure. The other half is contact time and route. A shampoo surfactant like sodium lauryl sulfate sits on skin for well under a minute before it goes down the drain; the same surfactant in a leave-on lotion would sit on skin for hours. Our sodium-lauryl-sulfate dossier grades it "watch if sensitive" specifically because it is a stronger cleanser that some people find drying in leave-on or high-concentration use, while a rinse-off formulation at typical use levels is a different exposure altogether.

This is why PlainScore discounts concern for rinse-off products: the same ingredient, same percentage, poses less real-world exposure when it is on skin for seconds rather than all day. A hazard-based list that only reads the ingredient name cannot make that distinction, because it never gets to exposure at all.

None of this is an argument to wave off every concern. Fragrance is graded "watch if sensitive" in our system regardless of where it sits on the list, because undisclosed fragrance blends are the single most common cause of cosmetic contact allergy, and "undisclosed" means position on the label tells you nothing about what is actually in it (see our fragrance dossier). Sensitization risk does not always scale down neatly with concentration the way irritation does, so we do not apply the same dose logic to every category of concern. Retinol is graded for irritation and a sun-sensitivity note that applies at ordinary use levels, not just at unusually high ones.

The point is not "nothing to see here." The point is that a defensible score has to ask three questions, hazard, concentration, and exposure route, before it lands anywhere, and it has to be honest when it cannot fully answer the second one.

Every PlainScore blends three weighted pieces: published concern (55%), transparency of the formula (20%), and formulation context (25%), which is where INCI position and leave-on versus rinse-off status come in. The concern weighting itself draws only from CIR, SCCS, FDA, and peer-reviewed sources, never from a hazard list that ignores dose. The full method, including how we handle ties and missing data, is on our /about page.