Is the Yuka app accurate for skincare?

Partly. Yuka reads an ingredient list correctly, but it scores ingredients in isolation. Because cosmetic formulas rarely disclose concentrations, it cannot judge the dilution, pH or delivery system that actually determine whether a product irritates you. Cosmetic chemists interviewed by Glossy, NewBeauty and Hypebae have consistently raised this as the core limitation.

Why does concentration matter so much?

Because nearly every ingredient in skincare changes character with dose. A preservative at 0.1% keeps a product from growing bacteria; the same molecule at 10% would be a different proposition entirely. An essential oil at 0.05% and the same oil at 3% carry very different sensitisation risk. An ingredient list tells you what is present and roughly in what order, but not the number, and cosmetic chemist Jane Tsui has pointed out that ratings which ignore percentage are therefore not fully accurate.

What does Yuka get right?

It parses labels reliably, it is free, and it has pushed a large number of people to look at ingredient lists rather than front-of-pack marketing. That is a real gain. It is also useful as a search tool: if you are avoiding a specific ingredient for a known reason, such as a diagnosed fragrance allergy, an app that flags that ingredient quickly is doing something genuinely helpful.

Is Yuka the same as EWG Skin Deep?

They are different organisations with a shared structural problem: both rate hazard, the capacity of a substance to cause harm at some dose, rather than risk, the likelihood of harm at the dose you are actually exposed to. Yuka's algorithm is also proprietary, so its reasoning cannot be independently checked the way a published safety assessment can.

What should I use instead?

Use INCIDecoder to understand a label rather than to rank it. It tells you what each ingredient is and what it does, and leaves the judgement to you. For actual safety assessments, the Cosmetic Ingredient Review panel in the US and the EU Scientific Committee on Consumer Safety publish exposure-based opinions with their reasoning shown. For a product that keeps irritating you, a dermatologist and patch testing beat any app.

Can an app tell me if a product will break me out?

No, and neither can we. Individual reactions depend on your barrier, your history, what else you are using, and concentrations no app can see. The only reliable test is a careful patch on your own skin, one new product at a time, over a couple of weeks.

Is the Yuka App Right About Your Skincare?

Researched, not tested. We have not personally used this product or protocol. This page answers the question from published sources, every one of which is linked. Our first-hand testing lives in reviews.

Evidence: mixed Mixed evidence — credible sources disagree; both sides are shown below.

Scanning a barcode and getting a number out feels like getting an answer. It's worth understanding exactly what kind of answer it is, because the app is doing something narrower and more mechanical than most people assume — and the gap between what it measures and what you wanted to know is where the trouble starts.

The short version

Yuka reads the ingredient list on the back of the box and scores the ingredients it finds. That part it does competently and quickly.

The problem is what an ingredient list leaves out. It tells you what is in the product and, above the 1% line, roughly in what order. It does not tell you how much, at what pH, in what delivery system, or buffered by what else. Those are the variables that decide whether a formula is soothing or stinging.

So the score answers a question you probably weren’t asking: how concerning are these molecules in general? rather than is this product good for my face?

Steelmanning the app first

It would be easy and wrong to be dismissive here.

Yuka got millions of people to turn a box over. Before these apps, the average shopper’s information source was the front of the package — the part written by marketing. Moving attention to the INCI list is a genuine improvement, and industry people who dislike the ratings often concede this point.

It is also free, fast, and consistent. If you have a diagnosed allergy — say you’ve patch-tested positive to a specific fragrance material — an app that flags that name in two seconds is doing something useful and hard to replicate by squinting at small print.

And its critics are not neutral. A lot of the loudest pushback comes from the cosmetics industry, which has an obvious commercial interest in consumers not having a pocket ingredient auditor. That doesn’t make the criticism wrong. It does mean you should want the criticism to hold up on its own logic, which is what the next section is about.

Where the reasoning breaks

It scores ingredients, not formulas. This is the central issue and it isn’t a matter of opinion. Two products can share an ingredient list and behave completely differently depending on percentages and pH. A cosmetic chemist quoted by NewBeauty made the point directly: the rating system doesn’t account for percentage within a formula, and percentages generally aren’t disclosed.

Dose is not optional information. Chemists interviewed by Hypebae described the core failure as oversimplifying toxicology. Every substance has a dose at which it causes harm, including water. A rating that ignores dose is measuring something, but not safety.

Essential oils are the clearest example. These apps commonly flag them as allergens regardless of dilution. Essential oils genuinely do sensitise — we say so elsewhere on this site — but at 0.05% in a cream, the flag is not telling you what it appears to be telling you.

The algorithm is proprietary. You can’t audit it. Compare that to a Cosmetic Ingredient Review or SCCS opinion, where the reasoning, the exposure assumptions and the safety margins are published and can be argued with.

Alarming words do work on people. Reporting in Glossy, Beauty Independent and Marie Claire has covered both the criticism from industry safety specialists and a knock-on effect worth naming: ingredient anxiety. People throwing out products that were fine, avoiding effective actives, and reading a red score as a health warning rather than a precautionary flag.

The honest limit on our own position

Two things we won’t overclaim.

First, no one has published a systematic accuracy audit of Yuka’s skincare ratings against a recognised standard. The criticism is expert commentary, not a study, and expert commentary from a conflicted industry is worth exactly what that implies. We rate the evidence here mixed for that reason.

Second, “the app is unreliable” does not mean “everything it flags is fine.” Some flagged ingredients genuinely are sensitisers for some people. The app’s failure mode is imprecision in both directions, not a systematic bias you can simply invert.

What to do instead

The replacement isn’t a better score. It’s a different question.

Use INCIDecoder to understand a label, not to rank it. It tells you what each ingredient is and does, and stops there. You get comprehension instead of a verdict.
For actual safety assessments, read the grown-up sources. The Cosmetic Ingredient Review panel in the US and the EU’s Scientific Committee on Consumer Safety publish exposure-based opinions and show their working.
Know your own list. Most people don’t need to evaluate 3,000 ingredients. They need to know the three or four that have burned them before. That list comes from your own experience, or from patch testing.
Patch test one new product at a time. Unglamorous, slow, and the only method that measures the thing you actually care about, which is your skin.

What an EWG Score Actually Measures — the hazard-versus-risk error in detail, explained once so it works everywhere.
Parabens: What the 2004 Study Actually Found — the single study behind the most-flagged ingredient class in these apps.
The Castor Oil Eye Trend — the same logic in reverse: an ingredient with no app flag at all, used in a way that can genuinely hurt you.
Don’t Put Neosporin on Your Face — a top-tier contact allergen sitting in a product almost no scanner treats as alarming.

Sources

Researched, not tested. This is an assessment of a rating method, not a review of a product we use. Evidence rating is mixed: the concentration limitation is a matter of formulation fact, but no published audit has measured the app’s overall accuracy, and several of its critics have commercial interests.