Does One Preservative Decide a Yuka Score? In Our Reads, Phenoxyethanol Did

In every product we have read in the Yuka app on a recorded date, phenoxyethanol decided which side of 50 the score fell on. Eight of the twenty-five products contain it. All eight scored between 24 and 49. Seven products scored 50 or more, and not one of those seven contains it.

That is a narrow finding, and it is worth stating exactly. Phenoxyethanol does not set the final number. It sets a ceiling at 49, and the rest of the label decides how far below the ceiling a product lands. It is also not the only way under 50: one product in the set reads 24 because of an antioxidant, and phenoxyethanol has nothing to do with it.

Inside those limits the pattern held without exception. The clearest case is two products with a single flagged ingredient each. A La Roche-Posay moisturiser with one low-risk flag reads 93. A hyaluronic acid serum whose only flag is phenoxyethanol reads 49.

One Flag Each, 44 Points Apart

On 12 October 2026 we read La Roche-Posay's Toleriane Double Repair Face Moisturizer. Yuka flagged one ingredient, dimethicone, at low risk, and passed everything else. Score: 93.

The same afternoon we read a hyaluronic acid serum, the one sold as Hyaluronic Acid 2% + B5 and now listed in the app with "and Ceramides" on the end. Yuka flagged one ingredient, phenoxyethanol, at moderate risk, and passed everything else. Score: 49.

Same number of flags. A 44-point gap. The only thing that differs is the level of the one flag, and that is the whole mechanism in a single comparison.

It follows Yuka's own published rule. Its help article "How are penalties calculated in the cosmetic product scores?" (help.yuka.io, article ih5pet4ffc, read on 20 October 2026) says the ingredient with the highest risk level determines the product's band: a product with a moderate-risk ingredient scores "between 0 and 49/100", one with a high-risk ingredient "between 0 and 24/100", and a product with only low-risk or risk-free ingredients "will systematically be greater than or equal to 50/100". Phenoxyethanol is moderate in every reading we hold. So wherever it appears, the product cannot reach 50, however clean the rest of the label is.

What the rule does not tell you is how often that happens in practice. That is what the table is for.

All 25 Dated Readings

Every row is an operator reading taken in the Yuka app on the date shown and logged on our Trello board the same day. The preservation column comes from a published ingredient list for the same product, named in the last column, and lists the ingredients on that label that do the preserving. The La Roche-Posay sunscreen readings from August recorded the score only, so their other columns are blank rather than guessed.

Ingredient lists on INCIDecoder were read on 20 October 2026, the Open Beauty Facts records the same day. CeraVe's own list for the day lotion was checked on 1 September 2026, when we published our review of it. Where a count matches the number of ingredients Yuka showed for the product, the table says so, because a match is the best evidence we have that the list describes the formula Yuka scored. Where it does not match, we say that too.

The La Roche-Posay sunscreens are written up in our Anthelios explainer. They stay in the table because they are real dated readings, but they cannot test the preservative question either way.

What the Readings Support, and Where It Stops

Take the sixteen readings where we know both the score and what Yuka flagged.

Eight contain phenoxyethanol. They score 49, 47, 43, 39, 39, 39, 39 and 24. In seven of them phenoxyethanol is the only ingredient Yuka rated above low risk. In the eighth, CeraVe's mineral sunscreen, it shares the moderate band with aluminium stearate, part of the coating on the mineral pigments.

Eight do not contain it. Seven of those score between 58 and 93, and none of the seven carries a flag above low risk. The eighth is the retinol serum at 24, held down by BHT, an antioxidant Yuka rates high risk. There is no preservative on that label at all: it is an oil with no water in it, which is the usual reason a product can do without one.

So the clean statement is this. Among products without a high-risk ingredient, phenoxyethanol was present in every reading below 50 and absent from every reading at 50 or above. Sixteen readings, no exception.

Below the ceiling, the rest of the label sets the exact number, and our readings fit a simple pattern. The serum with nothing else flagged reads 49. The niacinamide serum, with one extra low-risk flag, reads 47. The NMF cream, with three, reads 43. Four CeraVe products with five each read 39. Two points per extra low-risk flag, seven products out of seven. Yuka's help article lists a 2-point penalty for a low-risk allergen or irritant in a product with an orange ingredient, which is consistent with that, but it does not describe the starting point, so treat the arithmetic as our observation and not as Yuka's documented formula. The mineral sunscreen, with a second moderate ingredient and nine low ones, falls to 24, and the pattern alone does not account for it.

Two other preservatives in the table make the same point from the opposite direction.

Sodium benzoate sits in CeraVe's Foaming and SA cleansers, both at 39, and in the peeling solution at 93. Yuka passed it every time. It is phenoxyethanol, sitting beside it in the two cleansers, that separates them.

Chlorphenesin appears on four labels: CeraVe's mineral sunscreen, and The Ordinary's two serums and cream at 43 to 49. Yuka passed it every time as well, and two of those four lists match Yuka's ingredient count exactly, so it was in the formula Yuka read. This one deserves attention, because European law is stricter about chlorphenesin than about phenoxyethanol. It sits in Annex V of Regulation (EC) No 1223/2009 at entry 50 with a maximum of 0.3%, against phenoxyethanol's 1.0% (CosIng, the Commission's ingredient database, checked 20 October 2026). A tighter legal limit and no Yuka flag. Whatever Yuka's hazard list is measuring, it is not how tightly regulators cap a preservative.

What European Regulators Say About Phenoxyethanol

Phenoxyethanol is a permitted preservative in the EU. CosIng lists it in Annex V, entry 29, with a maximum concentration of 1.0% in the finished product (checked 20 October 2026).

That limit was reviewed, and kept. In its opinion SCCS/1575/16, final version of 6 October 2016, the EU's Scientific Committee on Consumer Safety concluded: "The SCCS considers 2-phenoxyethanol safe for use as a preservative with a maximum concentration of 1.0%, taking into account the information provided." The committee was asked specifically about young children and found that the margin of safety it calculated for children, about 50, covers that age group too. We re-read the opinion's PDF on the Commission's site on 20 October 2026.

The other side of the file is real. The same opinion records why it was written. France's medicines agency, ANSM, had concluded that the limit should be lowered to 0.4% in products for children under three and that phenoxyethanol should not be used in products for the nappy area. Other member states raised similar concerns. The committee reviewed those concerns and kept 1.0%. So Yuka's caution has a documented origin. The regulator's conclusion, reached after looking at the same worry, is that the permitted dose is safe.

The two positions can both be accurate because they answer different questions. Yuka's penalty article describes points per ingredient by risk level. It says nothing about concentration, and on the page we read it gives no way for the amount in a product to change the penalty. A product with a trace of phenoxyethanol and one at the 1.0% limit are treated the same. The regulator's opinion is entirely about the amount.

What a Score Below 50 Should Mean to You

Read it literally. For products like the ones in this table, a score in the 40s mostly tells you phenoxyethanol is in the bottle. It does not tell you how much, and on its own it is not evidence that the product is unsafe at the dose EU law allows.

That makes the score useful in one situation and much less useful in another.

If you have reacted to phenoxyethanol before, or a dermatologist has told you to avoid it, the score is a fast and reliable filter. In our readings, every score below 50 without a red flag meant phenoxyethanol. Scan, check the orange line, move on.

If you have not, a product at 39 and one at 93 may be much closer in practice than the 54 points suggest. CeraVe's whole range sits below 50 for this reason, and plenty of people with sensitive skin use it daily without trouble. Choosing between them is a question about your own skin and your own threshold for precaution, and the number cannot answer it for you.

One thing the score does not do, in either case, is reward a stricter preservative. Chlorphenesin, capped at less than a third of phenoxyethanol's limit, costs nothing.

If You Want a High Yuka Score, Look For…

A label without phenoxyethanol on it. That is the single most useful check, and you can do it in the shop. Every product at 50 or above in our readings was preserved some other way, most often with a combination of these:

• 1,2-hexanediol, caprylyl glycol and ethylhexylglycerin, the multifunctional glycols behind all five Medicube readings.
• Hydroxyacetophenone, alongside those glycols on the toner pad that read 91.
• Sodium benzoate with potassium sorbate, as on the peeling solution at 93.
• Caprylyl glycol with capryloyl glycine, as on Toleriane at 93.

None of these guarantees a high number. The Medicube capsule cream and peptide serum use the glycol system and still read 58, because six other ingredients each cost them points. What these systems do is leave the ceiling at 100 rather than 49.

From the dataset itself, the products that clear 90 are La Roche-Posay's Toleriane Double Repair Face Moisturizer (93), The Ordinary's AHA 30% + BHA 2% Peeling Solution (93) and Medicube's PDRN Pink Collagen Toning Gel Toner Pad (91). A caution on the second one: it is a strong acid peel meant for occasional use, and its score says nothing about how hard it is on your skin. As a daily moisturiser, Toleriane is the one to look at.

Two older readings sit outside this dataset because nobody recorded the day they were taken, and both point the same way. Cetaphil Gentle Skin Cleanser and Vanicream Daily Facial Moisturizer are both published here at 100. Cetaphil's published formula is preserved with sodium benzoate and citric acid, Vanicream's with caprylyl glycol, 1,2-hexanediol and pentylene glycol, and neither lists phenoxyethanol. Our three-way comparison with CeraVe goes through them in detail. We would not add them to the table until they are read again with a date.

How We Built This

The dataset is every product read in the Yuka app with a recorded date, as logged on our editorial board: 25 products from four brands (CeraVe, La Roche-Posay, Medicube and The Ordinary), read on 11 and 12 August, 31 August, 7 September and 12 October 2026. Each reading was taken in the app by the site's operator, screenshotted, and transcribed into the log the same day. For 16 of the 25 we hold the full list of flagged ingredients and their levels. For the nine La Roche-Posay sunscreen readings from August we hold only the score.

Older scores published on this site are not included. Most were recorded by an earlier workflow that kept no read date, and we have found that some of them no longer match the app.

This is the sample we chose to read, not a sample of the market. Four brands is a small base, most of the readings were taken because a brand was being written about, and the La Roche-Posay readings were filtered to that brand's lowest scorers. The finding is that within these readings the rule held without exception. It would take a much broader read to say how far it holds across a whole shelf, and we have not done that.

We used the Yuka app the way a reader does. We have no access to Yuka's database and do not reproduce its method; everything we say about how scores are built comes from Yuka's own help pages or from comparing our readings.

Corrections

Older reviews on this site do not all fit the pattern above, and we think the fault is ours. Three of them list phenoxyethanol among the ingredients while publishing a score at or above 50: the Neutrogena Hand Cream (Norwegian Formula) at 92, The Body Shop Moringa Body Butter at 100, and the Nuxe Men Multi-Use Shower Gel at 76. All three were published before we recorded read dates. Either the scores are out of date or the formula Yuka holds differs from the one we listed. Each page now carries a notice saying so, and all three are being re-checked in the app in November 2026. We will correct them with whatever we find.

The Answer

Yes, within one clear limit. In every product we have read on a recorded date, phenoxyethanol was the difference between a score under 50 and a score over it, unless something rated high risk got there first. It does not decide the exact number. It decides the ceiling.

Whether that ceiling should matter to you is a separate question, and Yuka does not try to answer it. The EU's safety committee looked at phenoxyethanol, including for small children, and kept it at 1%. If you have no reason to avoid it, a score in the 40s is telling you less than it looks. If you do, it is one of the most reliable signals the app gives.