How we decide
Every filter on this site is worked out from the ingredient list, not copied from the label. This page shows the reasoning behind each one, the places we deliberately stop, and the mistakes we have found in our own data.
Yes, we are stricter than what you are probably used to — but not without full justification. Where we draw a harder line than the usual advice, we say so, and we show the evidence for it. You can disagree with any call on this page and still see exactly how it was made.
On this page
How we judge whether an ingredient feeds Malassezia
Malassezia is a lipid-dependent yeast. It cannot make its own fatty acids, so it has to take them from its surroundings — from your skin's own oil, and from what you put on top of it. That single fact is what the whole site is built on.
The chain length of the fatty acid decides whether it is food. Fatty acids roughly in the C12 to C24 range support growth. This was established by Nazzaro-Porro and colleagues in 1975 and it is still the working range today.
Shorter chains behave in the opposite way. Caprylic acid, at eight carbons, is not usable as food and is actively inhibitory. That is why an ingredient built from caprylic or capric acid can sit in a product without being a problem, while the same structure built from palmitic or stearic acid is.
It is not only free fatty acids. Malassezia can split many esters apart to reach the fatty acid inside, and Dobler and colleagues showed in 2019 that the structure of the ester changes how easily that happens — the alcohol half matters, and unsaturated fatty acids stimulate growth more than saturated ones.
There are real exceptions, and we honour them rather than rounding them away. Fatty acids carrying a hydroxyl group behave differently: growth on hydrogenated castor oil derivatives was seen for only one species, and on some related materials for none at all. An ingredient that looks like food on paper is not always food in a dish, and where the published work says so, we follow the work.
Sources: Nazzaro-Porro et al., 1975 · Mayser et al., 1997 (doi:10.1046/j.1365-2133.1997.18071890.x) · Dobler et al., 2019 (doi:10.3390/cosmetics6030045)
What each filter means
Each filter answers one question about what is ina product. None of them answers whether the product is safe for fungal acne — that is a separate verdict, shown separately, and the two can disagree. An ingredient can be present and harmless.
We work these out from the ingredient list every time. Where a brand's own label and the ingredient list disagree, the ingredient list wins.
Ester-Free
- Counts as present
- Esters of fatty acids. That includes the ones most people recognise, such as isopropyl palmitate and glyceryl stearate, and also plant oils, butters and waxes, which are chemically triesters of fatty acids.
- Does not count
- Esters that are not built from a fatty acid: parabens, most sunscreen filters, carbonates, carbamates, sulfates and phosphates. Pure hydrocarbons such as mineral oil and squalane, which contain no ester at all. Essential oils, which are terpenes rather than fatty-acid esters.
- Worked example
- Sunflower Seed Oil is a triglyceride, which is three fatty acids joined to glycerol. That is an ester, so a product containing it is not ester-free. This is stricter than the common shortcut that esters are the ingredients ending in -ate, and it is the definition the underlying chemistry actually gives.
Oil-Free
- Counts as present
- Anything a shopper would call an oil: plant and seed oils, essential oils, mineral oil and petrolatum, plus butters and waxes.
- Does not count
- Ingredients that merely feel oily or act as emollients without being oils.
- Worked example
- Squalane is an oil, so a product containing it is not oil-free — even though squalane itself is safe for fungal acne. The filter describes what is in the bottle, not whether it is a problem.
Fatty Alcohol-Free
- Counts as present
- Long-chain alcohols derived from fatty acids, such as cetyl, stearyl and cetearyl alcohol.
- Does not count
- Simple solvent alcohols such as denatured alcohol and isopropyl alcohol. They share a word and nothing else.
- Worked example
- Cetearyl Alcohol is a fatty alcohol. Alcohol Denat. is not. A product can be alcohol-free and not fatty alcohol-free, or the other way round.
Fragrance-Free
- Counts as present
- Ingredients declared as Fragrance, Parfum or Aroma.
- Does not count
- Ingredients that happen to carry a perfuming role in a regulatory database while being present for another reason — botanical extracts, preservatives, antioxidants.
- Worked example
- A rosemary extract used as an antioxidant does not make a product fragranced. Counting it would fail as badly in one direction as ignoring a declared Parfum fails in the other.
EU Allergen-Free
- Counts as present
- The fragrance allergens the European Union requires to be declared by name on a label, identified per ingredient against the Commission's own cosmetic ingredient database.
- Does not count
- Ingredients that merely share a plant source with a listed allergen. The material form matters: an oil or extract may be in scope where the water is not.
- Worked example
- Lavender oil and lavender water are not the same entry, and we do not treat them as one.
Sulfate-Free
- Counts as present
- The sulfate surfactants people mean by the term — the cleansing agents.
- Does not count
- Mineral salts that happen to be sulfates, such as magnesium sulfate and zinc sulfate. Nobody buying a sulfate-free cleanser is avoiding epsom salt.
- Worked example
- Sodium Lauryl Sulfate counts. Magnesium Sulfate does not.
Paraben-Free
- Counts as present
- The paraben preservatives.
- Does not count
- Nothing of substance — this is the most clear-cut filter on the site.
- Worked example
- Methylparaben is a paraben. It is also, chemically, an ester — but not a fatty-acid ester, so it does not affect the ester-free filter. Different questions.
Alcohol-Free
- Counts as present
- Drying solvent alcohols: denatured alcohol, ethanol, isopropyl alcohol and their relatives.
- Does not count
- Fatty alcohols, which are emollients and behave nothing like solvents, and sugar alcohols such as glycerin.
- Worked example
- A product with Cetearyl Alcohol and no solvent alcohol is alcohol-free.
Silicone-Free
- Counts as present
- Silicones and siloxanes, identified both by name and by the chemical description held in the EU's database.
- Does not count
- Ingredients whose names merely resemble silicone names.
- Worked example
- At least one silicone in our catalogue gives no hint in its name and is only identifiable from its chemical description. Matching on names alone would have missed it.
Squalane: one ingredient, three answers
Squalane is an oil. It is ester-free. It is safe.
Three different questions. Three independent answers. One ingredient.
Squalane is a pure hydrocarbon — a chain of carbon and hydrogen with no fatty acid and no ester bond anywhere in it. So it is an oil by every ordinary meaning of the word, it contains no ester, and Malassezia cannot eat it.
This is why our filters do not collapse into one another. “Is it an oil”, “does it contain an ester” and “will it feed the yeast” are three separate questions, and a site that answers them with a single yes or no is throwing away information you need.
What we refuse to claim
We do not calculate whether a product is non-comedogenic.Where a brand makes that claim we will show you that the brand makes it, clearly labelled as the brand's own claim, and we will not dress it up as a finding of ours.
Three reasons, all from the published literature:
- The comedogenic ratings in circulation descend from a rabbit-ear test. A 1982 review in the Journal of the American Academy of Dermatology found an absence of correlation with what happens in human skin.
- The human model that replaced it reported that the rabbit test is more sensitive than human skin. The lists over-call.
- The assumption that a raw ingredient's score predicts the finished product's behaviour was examined and abandoned in a 2006 re-evaluation of the concept. A finished formula is not the sum of its ingredient ratings.
On top of that, a 2018 systematic review found the terms non-comedogenic, hypoallergenic, fragrance-free and oil-free on cosmetic labels are regulated by no governing body, and a 2022 paper still describes comedogenicity assessment as not yet established.
It also measures a different problem from ours. A comedo is a plug of keratin in a follicle. Fungal acne is a yeast growing in a follicle because something fed it. Different mechanisms, different causes — and in practice the two do not line up at all.
We can measure that across our own catalogue. Of the products whose brands claim non-comedogenic, roughly seven in ten are not fungal-acne safe by our reckoning, and about half are clearly unsafe (measured September 2026). A non-comedogenic label is not a fungal-acne verdict, and we will not let it behave like one.
Sources: PMID 6461674 · PMID 7138047 · PMID 16488305 · PMID 29616616 · PMID 29671667 · PMID 34632684
Where our data comes from
- CosIng, the European Commission's cosmetic ingredient database, for what an ingredient name officially denotes and for the EU fragrance allergen list.
- The Cosmetic Ingredient Review (CIR), the independent expert panel that reviews the published toxicology on a cosmetic ingredient and issues a finding on whether it is safe as used. We show that finding where we have it.
- EWG, for a second opinion on general skin safety. It is an advocacy organisation rather than an expert panel, and we treat it as one input, never as a verdict.
- PubChem and PubMed for chemical identity and for the primary research behind our reasoning.
- The product label itself, for the ingredient list.
A safety finding is not a fungal-acne finding.CIR judges whether an ingredient is safe to put on skin. That is a different question from whether Malassezia can eat it, and the two answers routinely disagree. Cetearyl alcohol is judged safe as used, and we class it as a trigger — both of those are correct, because they are answers to different questions. We never let one stand in for the other.
None of these is complete, and we would rather say so than imply otherwise. The EU database has no chemical description for more than half of the ingredients we needed to settle the ester question. That is why no single source decides any rule here: each one is checked against the others, and where they disagree we look at the primary literature and record the decision.
Coverage and limits
We publish filters across our whole catalogue. A small number of listings are accessories with no ingredient list at all, and we make no free-from claim about those whatsoever. An unanalysed product is not the same as a product free of something, and we will not blur that line to make a result look tidier.
A small number of products have very short ingredient lists. A short list is still a list and we treat it as one, but it is the group most likely to be incomplete on the packaging, so treat those results with a little more caution than the rest.
We also do not know concentrations. Ingredient lists are ordered, not quantified, and almost no brand publishes percentages. Every judgement here is about presence.
Formulas change
Brands reformulate without renaming the product and without announcing it. A tube bought this year may not match the one reviewed last year, and the packaging is often the only place the change appears.
So treat everything here as a reading of the formula we have on record, and check the label on the box in your hand — especially if a product matters to you. If yours does not match what we show, tell us and we will re-derive it.
Report a mistake
If something on this site is wrong, we want to know, and we would rather hear it bluntly than not at all. Tell us the product and what you are seeing on the packaging and we will re-derive it from the ingredient list.
Get in touch — or use the ingredient checker to run a label yourself and see how a verdict is reached.