How do I make sense of skincare research claims?

Laboratory studies test cells or tissue in a dish. They can suggest how an ingredient might work, but skin in a person is far more complex, so a laboratory result is early evidence only.

Ingredient studies in people test a particular concentration in a particular base. They support the ingredient in that context, not every product that contains it. Randomised controlled trials of the finished product, ideally double-blind and compared with the same base without the active (a vehicle), are the strongest direct evidence for a specific product. Systematic reviews pool several studies, but they are only as good as the studies included.

Many cosmetic studies report how participants felt their skin looked or felt. These perception results are real experiences but are easily influenced by expectation, and the UK advertising regulator treats them differently from efficacy claims. Measurements by an assessor who does not know which side or product was used are harder to influence.

A published example: a 40-person randomised, double-blind split-face trial of a CALECIM cream found that both the active and the vehicle sides improved. Blinded investigator assessments did not differ significantly between them, although participants preferred the active side. Reading only the participant preference would give a very different impression from the full result.

Brands often run their own tests. These can be well conducted, but they are not independent, they may be unpublished, and the methods are often unavailable to check. Affiliation does not make a study wrong; it means it deserves the same scrutiny and should be described honestly.

Our evidence levels describe research on an ingredient or product type, scoped to what has been studied. They are not a guarantee for a specific product or a medical endorsement. Products we have not individually reviewed are shown as Not yet assessed rather than given a rating.