Claims substantiation starts with what your claim tells a shopper. Write down what the wording and packaging communicate, which product and users they concern and what result would back them up. With that in hand, a qualified testing provider can tell you which evidence could support the claim and how far that evidence would reach.

This guide helps brand teams prepare that conversation. It includes a claim-to-evidence table and a study brief you can send to providers. The cosmetic product development roadmap shows how claims evidence fits with safety assessment, documentation and launch.

What the EU rules require

A claim study shows whether a benefit can be supported. Safety, and whether the product is a cosmetic at all, are assessed elsewhere. The common claim criteria apply only once a product has been assessed as a cosmetic, and in the EU it is national authorities and courts that decide case by case which framework applies. A claim that implies treating a condition can raise that question, and claim testing will not settle it.

In the EU, Article 20(1) of Regulation (EC) No 1223/2009 bans text, names, trade marks, pictures and other signs that imply characteristics or functions the product does not have. Regulation (EU) No 655/2013 sets six common criteria for claims: legal compliance, truthfulness, evidential support, honesty, fairness and informed decision-making. Under its Article 2, the responsible person must make sure claim wording meets those criteria and is consistent with the documentation proving the effect in the Product Information File. A study mainly addresses evidential support, so the wording still needs checking against the other five.

Other markets have their own rules. The EU launch checklist covers EU responsibilities and the MoCRA guide covers US safety records.

Step 1: Turn the wording into an evidence question

Write the claim exactly as it will appear, then ask what a reasonable shopper would take it to mean. Read it in context, with the headline, small print, images and placement. Article 1 of Regulation 655/2013 covers claims conveyed explicitly or implicitly through texts, names, trademarks, pictures and other signs, whatever the medium.

For each benefit claim, record:

• Product: formula version and pack.
• Population: who the claim is about, and who would be studied.
• Benefit: a feeling, a measured property or an assessed visible result.
• Use conditions: how and how often the product is applied.
• Timing: immediate, after repeated use or long-lasting.
• Comparison: against a baseline, another product, or implied superiority.
• Presentation: qualifiers and images that change the meaning.

Ask the provider to turn the benefit into an endpoint (the thing measured or assessed) and explain why it fits the claim. Work from the claim to the measurement, so the wording never ends up larger than what was measured.

Keep each wording version with its evidence in the Product Information File. If “feels softer” later becomes “visibly smoother skin”, treat it as a new claim and check its evidence again.

Step 2: Know the three kinds of evidence

Each type of evidence answers a different question:

• Consumer perception: participants report how the product felt or performed for them. Ask for the exact questions, answer options, population and use conditions. Their answers show how users experienced the product.
• Instrumental measurement: a device measures one defined property under controlled conditions. Ask which endpoint is measured and how far the result can be interpreted. A reading shows that the property changed, though not whether consumers noticed.
• Expert assessment: a trained assessor grades a defined outcome against stated criteria. Ask who assesses, with what qualifications, on which scale and with which controls.

One study can combine these, as long as each conclusion rests on the part that supports it. When a quote says “clinical testing”, ask the provider to describe the actual work.

The Annex to Regulation 655/2013 asks for adequate and verifiable evidence, including expert assessments where appropriate, that takes state-of-the-art practice into account. Studies used as evidence must be relevant to the product and the claimed benefit, well designed and conducted, valid, reliable, reproducible and ethical. The Annex does not assign a study type to each claim. It also makes allowance for obvious exaggeration and abstract statements that no one would take literally; this guide covers claims a shopper could take at face value.

Step 3: Map each claim to the evidence it needs

Use this table to find gaps before you ask for quotes. The wordings are examples, not approved claims.

Add the formula reference, claim version, evidence owner and open issue to each row. Where no evidence route fits yet, note the gap and keep the claim on hold.

Step 4: Send providers a study brief

Commission work against a written question and ask the provider to justify the design. Copy and adapt:

PROJECT / REQUEST VERSION / DATE:
BRAND CONTACT / TECHNICAL AND CLAIMS REVIEWERS:
TARGET MARKETS AND ANY CLASSIFICATION QUESTIONS:
DECISION THIS WORK SHOULD INFORM:

CLAIM AND PRODUCT
- Exact proposed wording, images and qualifiers:
- What the claim is meant to communicate:
- Formula version, sample source and pack:
- Intended users and directions for use:
- Existing evidence and its gaps:

QUESTIONS FOR THE PROVIDER
- Proposed evidence approach and why it fits this claim:
- Study population, selection criteria and recruitment:
- Endpoints and how they relate to the claimed benefit:
- Use conditions, time points and rationale:
- Baseline, comparator and controls:
- Sample size and statistical analysis rationale:
- Questionnaires, measurements or grading criteria:
- Ethics approval, consent and participant protection:
- Handling of deviations, dropouts and unexpected findings:

REPORT
- Product identity and work actually completed:
- Methods, results, analysis and limitations:
- Exact questionnaire wording or grading details:
- Which claims the provider considers supported, and which not:
- Report access and evidence-register reference:

REVIEW
- Who reviews the final wording against the report:
- Decision owner, open questions and commitments on hold:

Sample size, duration and statistical thresholds are for the provider and your reviewers to set. Ask them to explain each choice in terms of the claim. Agree in advance how incomplete work, deviations and limitations will be reported.

Step 5: Check the final claim against the report

Check the final claim against what the study found. The honesty criterion in the Annex to Regulation 655/2013 says presentations of a product’s performance must not go beyond the available supporting evidence.

Ask your reviewers to check:

1. Identity: the report covers the formula version you will sell.
2. Scope: which population, conditions, comparisons and time points were actually covered.
3. Results: what the data support, including limitations and deviations.
4. Wording: the claim adds no population, duration, effect size, comparison or certainty beyond the study.
5. Whole pack: images, headlines and small print tell the same story.

Ingredient data needs its own link to your product. The Annex says that extrapolating ingredient properties to the finished product needs adequate and verifiable evidence, for example showing the ingredient is present at an effective concentration. A supplier’s study on an active ingredient is a starting point for that argument.

If you quote a percentage of participants, keep the question they answered and the base it was calculated on. If a share of participants agreed their skin felt smoother, the wording should say what they reported, for example “X% said their skin felt smoother”.

For the EU, Article 11(2)(d) of Regulation 1223/2009 requires the Product Information File to contain proof of the claimed effect where the nature or effect of the product justifies it. In the US, the MoCRA safety substantiation requirement concerns product safety records.

Step 6: Record the decision, including what is on hold

Each claim decision should name the open question and what may go ahead. A typical mismatch: the proposed claim is “increases skin hydration”, but the provider’s quote covers only a questionnaire on perceived softness. The record would hold the claim, and any artwork that depends on it, until a provider proposes a way to assess hydration and your reviewers accept it. Changing the wording is also an option, but the new wording then needs its own evidence check.

Where to start

Pick the claim that matters most to your launch, fill in the Step 1 list for it and send the study brief with the current formula and pack references. The provider’s answer will tell you whether that wording can be supported or needs to change before artwork is committed. The EU launch checklist, MoCRA guide and PIF guide cover the other launch decisions.

National rules and other markets are not covered.