The EU Fragrance Allergen Deadline Passed on July 31. None of the New Names Are on Our Labels.

July 31 passed without anyone mentioning it. No supplier email, no note from a distributor, nothing from the two trade newsletters we actually read.

That was the day a four-year countdown in EU cosmetics law ran out. Commission Regulation (EU) 2023/1545 added a block of fragrance substances to the list that has to be named individually in an ingredient list, and 31 July 2026 was the last day non-compliant product could be placed on the Union market. We are not an EU seller. We noticed anyway, because we were re-counting our own fragrance data for something else and the two things ran into each other.

What 2023/1545 actually did

Article 19 of Regulation (EC) No 1223/2009 is the part that makes an EU ingredient list an ingredient list. Among other things it requires certain fragrance substances to be spelled out by name instead of disappearing inside the word "parfum". Until 2023 that list ran to twenty-six names. Linalool. Limonene. Coumarin. The ones you have already seen in four-point type at the bottom of a shampoo bottle without knowing why they were separated out.

2023/1545 inserted entries 327 to 371 into Annex III. That is forty-five entries. The Commission's public description of the same change says fifty-six additional fragrance allergens.

Both figures are correct, and we could not reconcile them by hand. We spent an hour on it and stopped. Some entries carry more than one substance, whether that is an isomer pair, a salt, or a plant extract listed alongside its isolated marker, so an entry count and a substance count were never going to land on the same number. If you have seen "24 to 81" attached to this rule anywhere, that pair of numbers appears in neither the regulation nor the Commission's material, and we would not repeat it.

There is also a correction sitting on top of all this: a corrigendum published in November 2025 that fixes the wording of entry 157, the rose ketones, and touches nothing else.

Two dates, and only the first has passed

The transitional arrangements are not in Article 2, which does the ordinary entry-into-force work and nothing more. They sit in footnotes attached to the annex itself, with the reasoning laid out in the recitals. We mention this because it is the sort of thing that gets miscited, and we nearly miscited it ourselves.

Those two phrases are not synonyms, and the gap between them is the whole practical story. "Placed on the market" means the first supply into the EU. "Made available" means every supply after that, the shelf included. A jar that entered the EU in June carrying an old label is still legal to sell today, and stays legal until the summer of 2028.

So if you are in Europe waiting for labels to change on 1 August, they didn't. They change as stock turns over.

So we read our own fifty-six labels

Fifty-six products. Every ingredient row in every one of them, pulled again this morning instead of trusting the counts in our own notes, because the catalogue went from fifty-four products to fifty-six in the last week and our numbers have gone stale on us before.

Sixteen products carry a generic fragrance declaration. They do it three different ways.

Three strings for one idea is sloppy record-keeping and it is ours. The line underneath it is worse. The identical string Fragrance (Parfum) is graded high on the againme scalp shampoo and the deepeer hair pack, and mid on the jhp hair essence and the eiio scalp scaler. Same eleven characters, two different risk grades, in a database we built. We have no defence for that and no explanation beyond the obvious one, which is that the grades were entered by hand at different times.

Eight of our products go further and print individual allergen names. Here they are, with what each one names.

Two of those eight name individual allergens while declaring no generic fragrance at all, which is the opposite of what you would expect and another sign the field was filled in inconsistently.

Which leaves ten products that say fragrance and then say nothing else: lacelltix, reslow, cure-alpha, myabis, noon-sup, neer-g, in-usth, foellie, againme, eiio.

Across the whole catalogue we name ten individual fragrance allergens, twenty-eight rows in total.

All ten are from the pre-2023 twenty-six. Entries 327 to 371 contribute nothing to our labels. Zero.

Hydroxycitronellal is the one to look at twice. We grade it high on jhp and deepeer and mid on both birthtii products. One molecule, two grades, and the split falls along which product it sits in rather than anything about the molecule. That is the same failure as the Fragrance (Parfum) problem above, in a different column, and we have now found it often enough that we treat it as structural rather than as a series of typos. We have written before about the ingredients we flag most; the honest footnote to that piece is that our flags are not always consistent with themselves.

The oils that declare nothing at all

Here is the part that changed how we read our own data.

Twelve of our fifty-six products contain an essential oil that we ourselves graded mid. Eighteen rows, fourteen distinct INCI strings, and not a single one graded anything other than mid. Ten of those twelve products name no individual allergen anywhere on the label.

Essential oils are where these allergens come from. Clove leaf oil is mostly eugenol. Geranium oil carries geraniol and citronellol. Rosa damascena oil carries both as well. Those are Annex III names. They are simply not printed as Annex III names, because the oil goes onto the label under its botanical INCI instead, which is legal and ordinary and is also precisely the gap the individual-naming rule exists to narrow.

Our dr-orga conditioner is the clearest case in the catalogue: four mid-graded essential oils in one formula, no generic fragrance declaration, no named allergen. Nothing about that is a violation of anything. It is just a label that tells you less than it looks like it does.

And we cannot take this further, because we do not have concentrations. We have said so before and it remains true. Without a percentage you cannot say whether any of this crosses a declaration threshold in any market, and we have never asked a supplier for one. That is a request we should have made two years ago.

What our screening found

What this is worth at the shelf

If you react to fragrance, the useful move is not to hunt for the word "parfum" and stop. It is to read past it. A list that names linalool and citral is telling you more than a list that says fragrance and moves on, and more again than a list that says nothing but carries four botanical oils.

Longer ingredient lists on European stock will start showing up gradually rather than all at once, for the shelf-life reason above. A Korean product sold in Korea is under Korean rules and this regulation does not touch it, so do not read a short allergen line on a Korean label as evidence of anything.

And the thing we would want someone to do with our own numbers: ask a brand for a concentration. We are going to start asking ours. If you want to see how the scores in this piece are put together, our screening method is written up in full, and the two companion pieces are our audit of fragrance-free claims in our own catalogue and our list of the ten ingredients we flag most. There is also a piece on which label claims carry a legal definition, which is the closest thing we have to a map of this territory.