Which Products Actually Score 100/100 on Yuka?

Twenty-four products in the table below read 100 out of 100 in the Yuka app. They come from twenty-three different companies. Nobody paid to be on the list, and several of them cost less than lunch.

A list like this is usually published by someone selling what is on it, which is why it tends to contain one brand. Ours contains a French supermarket own-label, a dermatology pharmacy staple, a bag of Epsom salt and a mascara. That mix is the point. It is also why the second half of this page is the ugly half: the fourteen readings we hold that fall below 76, nine of them from a single brand that half the internet recommends.

Every number here was published on this site against a specific formulation. We do not estimate a score, and we do not carry one across from a product with a similar name, because formulations differ between markets and change without warning.

The 100/100 Table, Grouped by What the Product Is For

Grouped this way because that is how the question usually arrives. Almost nobody wants a perfect score in the abstract. They want a face cream, or a body lotion, or a wipe that will not set off a rash.

Face care and moisturisers

Four of those nine are single-ingredient or near single-ingredient oils and waters, which is the first thing this table teaches. The shortest formula wins on Yuka every time, and a bottle of castor oil has no preservative to be flagged because there is nothing in it to spoil. The emulsified creams in that group are the interesting ones. We ranked the whole moisturiser shelf, high scores and low, in which face moisturisers score best on Yuka.

Cleansers, serums and eyes

A mascara at 100 is the entry that surprises people most, since eye makeup is the aisle shoppers assume is the worst offender.

The PanOxyl needs a footnote, and it comes from our own review rather than from the app. The reading is on the base version of the wash, which does not contain benzoyl peroxide. Scan a different product in the same range and you may well get a different number. That is the general rule here and the reason we record the formulation rather than the brand.

Body, hair, bath and wipes

Two product names in that block deserve a word of honesty. "Body lotion" and "Johnson's bp (w) 50g" are the names as the barcode database returned them, and we publish them that way rather than tidying them into something that reads better than the source. If a name looks like it came out of a scanner, it did.

A deodorant at 100 is the one worth stopping on. Deodorant is the category where scores usually collapse, because aluminium salts, synthetic musks and preservatives all live there. This one is built on a short list, and it reads clean.

What a 100 Certifies, and What It Does Not

A 100 on Yuka means one thing precisely: no ingredient in the declared list matched anything on the app's hazard database, at any risk level. It is a formulation screen run against a watchlist.

It is not a measure of whether the product works. Nothing in the score model looks at concentration, so a formula containing 0.1% of an active and a formula containing 10% of the same active are indistinguishable to it. Nor does it know whether the product suits your skin. Sweet almond oil scores 100 and will still break out plenty of people who use it on the face.

It also says nothing about price, packaging, testing on animals, or whether the brand behind it is one you want to fund. Those are real criteria. They are simply not in this number.

The strongest way to put it: a 100 tells you a specific worry is off the table, and leaves every other worry exactly where it was.

Why So Many of Our Readings Land on the Same Three Numbers

Here is something that only shows up once you have a few hundred readings in one place. Below 100, our scores pile up on three values, and they are each seven points apart: 93, 86 and 79. Twenty-two of our readings sit at exactly 93. Seventeen sit at 86, fifteen at 79. That is fifty-four of the hundred and thirty-six scored products on this site parked on three numbers.

Examples of the 93s: Garnier's Micellar Water All-in-1, Byoma's Hydrating Milky Toner and Beauty of Joseon's Ginseng Essence Water. Down a rung at 86 sits La Roche-Posay's Lipikar AP+ Cleansing Oil. Three unrelated products landing on the identical number is not a coincidence about the products. It is the shape of the arithmetic behind them.

The tempting conclusion is that each flagged ingredient costs seven points. We checked that against the readings where we recorded the flag count ourselves, and it does not hold. Medicube's PDRN Pink Collagen Toning Gel Toner Pad carries two flags and returned 91, not 86. Its Zero Pore Blackhead Mud Mask carries four and returned 80. So the deduction is weighted, not flat, and we are not going to publish a formula we cannot demonstrate.

What the clustering is genuinely good for is calibration. Treat a gap of seven points as noise. A product at 93 and a product at 86 are, in any practical sense, the same product with one more structural ingredient in it. A gap of sixty points is a different matter entirely.

The Near Misses: Everything We Hold Between 95 and 99

Seven products, and they are worth a look precisely because the thing standing between them and a perfect score is almost always trivial.

If you came here looking for a hand cream, the MKL at 98 is the best-scoring one on this site. Nothing in the two-point gap between it and a hundred should change your mind about it.

The Fourteen Readings We Publish That Fall Below 76

A list of winners from a site that only publishes winners is worth nothing. So here is the other end of our own shelf: of the hundred and thirty-six products we hold a score for, fourteen come in under 76.

Every single one of them is CeraVe or Medicube. That is not a verdict on those two brands relative to the rest of the market. It is an artefact of how we work: those are the two brands we have gone through product by product rather than sampling, and going deep is how you find the bottom of a range. Sample one product from a brand and you will usually catch it on a good day.

Now the part a shop page will never tell you. Low scores on this table do not all mean the same thing, and two of them are barely comparable. The Renewing SA Cream reads 0 because it contains two parabens, which Yuka classes as high risk and which cap a product below 25 on their own. The Facial Moisturising Lotion reads 39 because of phenoxyethanol, a preservative the EU permits up to 1% under Annex V, entry 29 of Regulation (EC) No 1223/2009, stacked with five structural ingredients that carry no restriction in European law at all.

One of those is a genuine regulatory argument. The other is a hazard app disagreeing with a chemist about emulsifiers. We laid out the whole brand-level picture in why CeraVe scores so low on Yuka, and every reading we hold sits on the CeraVe hub.

Medicube's range is a different story again, and one we had to correct in public. We originally argued that its toner pads scored low because of the format. A later reading broke that: one pad came back at 91 while another sat at 60, so the argument was wrong and we withdrew it. The current nine-product picture is in our Medicube score guide.

How We Would Actually Use This Page

If you are replacing something that scored badly, start from the group that matches what you are replacing and check the price before you check the score. Half the products on the 100 table are cheap. A switch that costs nothing is a switch worth making even if you think the app is oversensitive.

If you are chasing a perfect score across a whole bathroom, stop at the first product where the swap costs you something real. A dermatologist put that prescription moisturiser in your routine for a reason, and a number from a phone app is not a reason to overrule them. The comparison we ran across three gentle brands in CeraVe vs Cetaphil vs Vanicream works through exactly where that line sits.

And if you are here because something you have used happily for years scanned badly, the honest answer is that your own skin has more evidence than the app does. Keep the product.

This table grows as we read more products. It is a snapshot of one site's corpus, not of the market, and we would rather say that plainly than pretend twenty-four products are the whole answer.