Cosmetic Efficacy Testing vs Safety Testing: What’s the Difference?

“Is it safe?” and “Does it work?” are two different questions, and in cosmetics they are answered by two different kinds of tests. Brands often mix them up, ordering an efficacy study when the assessor needs safety data, or printing a claim that no study supports. Safety testing shows that a product does not harm people under normal use and is required for every cosmetic; cosmetic efficacy testing shows that the product does what its claims say and is needed whenever you make a claim that requires proof. This article compares the two side by side and shows how to plan both without wasting time.

The short answer

Safety testing feeds the safety assessment under Article 10 of Regulation (EC) No 1223/2009 and ends in a signed CPSR. Efficacy testing supports the claims you make under Article 20 and the common criteria of Regulation (EU) No 655/2013, and its results go into the Product Information File as proof of the claimed effect. Every product needs the first. The second depends on what you say on the label, the website and in advertising.

What safety testing answers

Safety testing asks whether the finished product can be used without harm, including reasonably foreseeable misuse. Typical evidence includes:

• a stability and compatibility test, so the product stays safe and unchanged in its packaging during its shelf life;
• a microbiological test, so the batch is clean;
• a challenge test, so the preservative system protects the product during use;
• ingredient toxicology and exposure calculations by the safety assessor;
• optionally, a patch test to confirm skin tolerance.

Read also Cosmetic Safety Testing: What It Covers and Who Needs It

What cosmetic efficacy testing answers

Efficacy testing asks whether a specific claim is true. “Moisturises for 24 hours”, “visibly reduces fine lines in four weeks” or “SPF 30” are statements of effect, and the brand must be able to back them up. Efficacy tests measure a defined effect, under defined conditions, on a defined group of volunteers or in a defined laboratory model, and compare it with a baseline, a placebo or an untreated area.

The legal basis compared

For safety, the law sets out what information is needed: Annex I of the Cosmetics Regulation, with guidelines in Implementing Decision 2013/674/EU. For efficacy, the law sets principles rather than a list of tests. Regulation (EU) No 655/2013 lists six common criteria for claims: legal compliance, truthfulness, evidential support, honesty, fairness and informed decision-making. The Technical Document on Cosmetic Claims, agreed by the Commission’s working group in 2017, adds best practice on the type of evidence; it is not legally binding but is widely used by authorities. The Product Information File must contain proof of the claimed effect where the nature or effect of the product justifies it.

Common efficacy test methods

• Instrumental measurements on volunteers, for example skin hydration, transepidermal water loss, elasticity or image analysis of wrinkles.
• Expert grading, where a trained assessor or dermatologist scores visible changes on a defined scale.
• Consumer perception studies, where users answer structured questionnaires; suitable for claims about how a product feels or is perceived.
• Laboratory and in vitro methods for some effects, such as in vitro sun protection measurements.
• Sun protection tests following ISO methods, such as ISO 24444 for in vivo SPF or the in vitro method ISO 23675 published in 2024, in line with Commission Recommendation 2006/647/EC on sunscreen efficacy.

The right method depends on the exact wording of the claim. A claim about perception (“skin feels softer”) needs different proof from a measurable claim (“increases hydration by 30%”).

Read also CPSR vs PIF: What’s the Difference?

Which claims need which kind of proof?

• Ingredient claims such as “with hyaluronic acid”: the formula and the ingredient list, provided the ingredient is present at a meaningful level.
• Sensory and perception claims such as “leaves skin feeling soft”: a consumer perception study is often suitable.
• Measurable claims such as “24-hour hydration” or “reduces wrinkles”: an instrumental study on the finished product, with a defined panel and statistical analysis.
• Sun protection claims: SPF and UVA tests under recognised ISO methods.
• Tolerance-related wording: safety and tolerance data, worded truthfully and without suggesting the product is special simply because it complies with the law.

Why safety comes first

Efficacy studies on human volunteers are only ethical once the product is expected to be safe. That is why efficacy labs usually ask for a safety assessment, or at least key safety data, before they start. In practice, the formula, packaging and safety data should be final before you invest in efficacy studies. A change to the formula after an efficacy study can make both the safety data and the claim evidence outdated.

Where the patch test fits

A patch test is a safety test, not an efficacy test. The product is applied under an occlusive patch for 48 hours, and the skin is graded for irritation. It confirms tolerance and gives your safety assessor additional human data. It does not show that a product hydrates, firms or brightens. If you want to mention testing on the label, the wording must describe truthfully what was tested; agree it with your safety assessor or a claims specialist.

Read also Advantages of testing cosmetics with us

Planning both: a practical order

1. Freeze the formula and packaging.
2. Run stability, microbiological and challenge tests in parallel.
3. Add a patch test if the product is leave-on or aimed at sensitive skin.
4. Obtain the CPSR.
5. Draft your claims and choose efficacy studies that match their exact wording.
6. Finalise label and marketing texts only after the evidence is in your file.

What we do and do not test

Cosmetic Testing Lab focuses on the safety and quality side: stability, microbiological, challenge and patch tests, the CPSR and SDS. We do not run efficacy or claim-substantiation studies, including SPF testing. For those, work with a specialist efficacy laboratory, and bring them your finished safety file.

A worked example

A brand launches a face serum and wants to say “hydrates for 24 hours” and “suitable for sensitive skin”. It first completes the stability, microbiological and challenge tests, adds a patch test and obtains the CPSR. It then commissions an instrumental hydration study from a specialist efficacy lab, using the same formula and packaging. The sensitive-skin wording is discussed with the safety assessor, who checks it against the safety data. Only when both pieces of evidence are in the Product Information File does the brand print the claims on the label and use them in advertising.

Common mistakes

• Using a patch test to support an efficacy claim.
• Writing the claim first and searching for a study later.
• Running an efficacy study on a formula that later changes.
• Relying only on an ingredient supplier’s data for a claim about the finished product.
• Assuming that a safe product needs no proof for its claims.

Frequently asked questions

Is cosmetic efficacy testing mandatory?

There is no fixed list of required efficacy tests, but every claim needs adequate and verifiable evidence. For many measurable claims, that means a study on the finished product.

Can one study cover both safety and efficacy?

Some efficacy studies record tolerance as a side result, but they do not replace the safety data required for the CPSR.

Can I use supplier data for my claims?

Supplier data can support ingredient-level claims. For claims about the finished product, evidence on your own formula is generally stronger and more defensible.

Key takeaways

• Safety testing answers “is it safe?”; efficacy testing answers “does it work as claimed?”.
• Safety data is required for every cosmetic; efficacy evidence is required for claims.
• Complete the safety side first, then match efficacy studies to your exact claims.
• A patch test confirms tolerance; it is not proof of efficacy.