What a CPSR is, what separates Part A from Part B, and who may sign it

The one-sentence version

Before a cosmetic product is placed on the Union market, the responsible person has to ensure it has undergone a safety assessment and that a cosmetic product safety report is set out in accordance with Annex I (Art. 10(1), Reg. (EC) 1223/2009).

That report is the CPSR. It sits inside the product information file (Art. 11(2), Reg. (EC) 1223/2009), which is the folder that holds it and everything it was built from.

Part A and Part B are different kinds of thing

Annex I divides the report into two parts, and the division is not cosmetic. It is the difference between evidence and judgement. What follows describes what each part contains; the item numbering is Annex I's own, and we reproduce it once it has been transcribed from the Official Journal text and dated.

Part A is the safety information. It is what is true about the product: the quantitative and qualitative composition, the physical and chemical characteristics and stability, the microbiological quality, impurities and trace materials and information about the packaging material, normal and reasonably foreseeable use, exposure to the product and exposure to its substances, the toxicological profile of each substance, and any undesirable effects on record.

Part A is data. It can be assembled, it can be checked, and every line of it should be traceable to a document in the file — a supplier specification, a certificate of analysis, a stability study, a challenge test, a published toxicological endpoint.

Part B is the safety assessment. It is what a qualified person concludes from Part A: the reasoning, the conclusion on the safety of the product, any warnings and instructions for use that the conclusion depends on, and the assessor's credentials and approval.

Part B cannot be assembled. It is somebody's professional judgement, and it is the thing you are actually buying.

The practical test of whether you have a real report is whether Part B refers to Part A. If Part B is three paragraphs of boilerplate that would read identically against a different formula, you have a template with your product's name typed into it.

Who may sign it

Art. 10(2), Reg. (EC) 1223/2009 requires the assessment to be carried out by a person holding:

a diploma or other evidence of formal qualifications awarded on completion of a university course of theoretical and practical study in pharmacy, toxicology, medicine or a similar discipline

Four things follow.

Can a pharmacist sign a CPSR? Yes. Pharmacy is named in the article, ahead of toxicology and medicine.

Does the assessor have to be in the EU? Art. 10(2) is a qualification requirement, not an establishment requirement. The establishment requirement lives in Art. 4(1), and it applies to the responsible person, not to the assessor. These two are constantly conflated, in both directions.

Is the assessor the same person as the responsible person? Not necessarily. They can be. Art. 10(1) makes the responsible person ensure a report exists; Art. 10(2) governs who may produce it.

Is the name visible? It should be. Every report we issue is signed by David Vandevoort, pharmacist (Master, KU Leuven; PhD, KU Leuven 2010). If you are comparing providers, ask each of them who signs, and ask for that person's qualification. It is a fair question about a document whose legal weight rests entirely on the qualification of the person who wrote it.

What actually stops a signature

We could not find a provider that publishes this, so here is ours.

A report does not get signed here when:

• A substance cannot be resolved. A raw material arrives under a trade name, or a traditional name with no INCI equivalent, and the composition behind it is unknown. You cannot assess what you cannot identify.
• A restricted substance exceeds the condition in force for that product type. The Annexes — II (prohibited), III (restricted), IV (colourants), V (preservatives), VI (UV filters) — carry conditions by product category. If the formula is outside the condition, the answer is reformulate, not argue.
• A toxicological endpoint is a placeholder no one has accepted. Every scientific value we use starts out marked unconfirmed, and stays unconfirmed until a person has checked it against a source. A report is not signed over an unconfirmed default unless a named person accepts that default — and where one does, the acceptance, the inputs it covers, who accepted them and when are printed on the signed document itself.
• A margin of safety falls below the threshold and Part B does not address it. A figure that is uncomfortable can be discussed. A figure that is uncomfortable and unmentioned cannot.
• The registry the screen ran against has not been reviewed by a person. A regulatory change is proposed automatically and accepted by a named human being. Nothing is assessed against an unreviewed registry.
• Composition is a range so wide that the assessment would be meaningless. "Between 1 % and 20 %" is not a composition.

In each of those cases you get the finding — the substance, the annex entry, the number — and the signature follows once the file can support it.

With us, the portal turns this into your own checklist →

One product, or one variation?

This is the question indie brands actually want answered, and it is usually answered badly. One maker put it plainly on a soapmaking forum:

With regard to colour/scent do I need to use up each of my 6 variations with each change of colour or fragrance/eo?

Here is the honest structure of the answer.

The Regulation does not set a variation count. We could not find one in Regulation (EC) 1223/2009. Where a provider sets a variation limit, or prices a report per variant, that is that provider's commercial rule. It may be a perfectly sensible rule. It is not the law, and you should be told which is which.

What decides it is the flattened composition. Take the product down to substance level, with a concentration range for each substance. Two products whose substance-level composition falls inside the same assessed ranges, with the same product type, the same use and the same exposure, are the same assessment problem. Two products where a change pushes a substance outside its assessed range, or introduces a substance that was not assessed at all, are not.

So, in practice:

• A colour change is often within scope, if the colourant is Annex IV-listed, used within its conditions, and the assessment covered a range that contains it.
• A fragrance change is usually not. Fragrance is a mixture. A different fragrance is a different set of substances at different levels, with different allergen content and a different toxicological profile. This is the change people most expect to be free and most often is not.
• A supplier change depends on whether the material's specification is the same. Same INCI, different supplier, different impurity profile is a real change.
• A pack-size change is usually not a formula question, though it can be an exposure question.

What we will do is tell you, before you commit, which of your variants sit inside one assessment and which need their own — and we will tell you the rule we applied, so you can predict the next one yourself. How we price a base versus a variation →

Is a very cheap report safe to use?

One UK provider lists its CPSR as shop items, from £55.00 for a single product with no variants up to £1,728.00 for sixty-four (read on phoenixcosmeticsafety.com, 20 September 2026; how published prices sit against our bands →). The honest question is not whether £55 is too little. It is what £55 can have looked at.

An assessment has a fixed floor of work: resolving every raw material to a substance, flattening the formula, checking each substance against the annex entry in force for that product type on that date, finding a defensible toxicological endpoint for each, calculating exposure and a margin of safety, and then writing a conclusion that refers to all of it. That floor does not move because the price did.

So a very cheap report is one of two things. It can be a template written against a fixed base recipe you must not deviate from — that model is sold openly, and the makers who use it describe the trade-off in the quote below. Or it can be a real assessment of a genuinely simple product — an anhydrous balm or a straightforward soap — where the floor of work really is low. The second is legitimate. Which one you have is visible from the report itself: run the two checks below.

Two checks you can run yourself, free: does Part B refer to Part A in terms specific to your product, and is the person who signed it named with their qualification.

Do I have to use your recipe?

No. A pre-made report sold with a fixed base recipe and a permitted fragrance list is a different product from an assessment of the thing you actually make. As one maker described the trade-off on a forum: you can buy a premade recipe, "but you are tied to their recipes/fragrance."

We assess your formula. If something in it needs a document or a change, we name which part and why, and you decide what to do about it.

FAQ

The registry behind it

Every Annex II–VI entry as consolidated on 18 May 2026 — 2,385 entries, 36,023 substances — re-checked daily against EUR-Lex, CosIng, SCCS and IFRA.