How to Read a Skincare Study (Without a Science Degree)

“Clinically proven.” “In a study.” “Research shows.” Skincare marketing leans on the language of science constantly, and most of the time the study behind the claim would not survive someone actually reading it. Phrases like “clinical strength” and “clinically proven” aren’t regulated terms — they can describe a rigorous trial or ten people and a questionnaire, and nothing on the label tells you which. You don’t need a science degree to tell the difference yourself — you need a handful of consistent questions, applied every time.

The hierarchy of evidence

Not all research carries the same weight, and skincare marketing rarely tells you which rung of the ladder a claim sits on.

In vitro studies — literally “in glass” — are experiments on cells in a dish, not on skin. They can show how an ingredient behaves in isolation, but cells in a petri dish don’t behave like cells in living tissue, the ingredient may never penetrate skin in real use, and the concentrations tested are often far higher than anything in a finished product. “Shown to increase collagen production” frequently means cells bathed directly in concentrated ingredient produced more collagen — not that a cream did the same thing on a face.

Animal studies sit a rung up, showing how an ingredient behaves in a living organism, though animal skin differs enough from human skin that results don’t always translate. These are rarely cited directly in marketing but sometimes appear as “preclinical research.”

Uncontrolled human studies put a product on real people but without a comparison group. If 80% of participants report improvement, there’s no way to know whether that’s the product, the placebo effect, or the fact that people testing something new tend to expect — and therefore notice — improvement.

Controlled human studies compare a treatment group against a placebo or comparator group, which is the point where a study can actually demonstrate an effect beyond placebo. This is usually what “clinically proven” is drawing on — though it’s worth checking what exactly it was compared to.

Randomised controlled trials (RCTs) randomly assign participants to groups and, ideally, blind both participants and researchers to who’s getting what. This is the design that minimises bias most effectively, which is also why proper RCTs are expensive and skincare studies using this design tend to be smaller and shorter than the gold standard would ideally call for.

Systematic reviews and meta-analyses combine results across multiple studies and sit at the top of the hierarchy — but they’re rare in cosmetic science, simply because there often aren’t enough quality studies on a given ingredient to combine. Retinoids are one of the few skincare ingredients with enough accumulated research for this level of evidence to exist at all.

Sample size: why it matters more than the marketing lets on

This is the single most common weak point in skincare research, and it’s worth understanding why, not just knowing the rule of thumb.

Small samples are noisy. If ten people use a serum and six report smoother skin, that could be a real effect — or it could be six people who would have said the same thing about a plain moisturiser, given the placebo effect, normal skin variation over a few weeks, and the tendency to notice what you’re primed to notice. With so few participants, there isn’t enough data to distinguish a genuine pattern from statistical noise. The wider the study, the harder it becomes for chance alone to produce a misleading result.

As a rough guide: under 20 participants is preliminary and should be treated cautiously, however confident the marketing sounds. 20–50 is moderate evidence, still limited. 50–100 participants is reasonable grounds for an initial claim. 100+ starts to carry real weight, and 500+ is genuinely rare in cosmetic research but robust when it happens and the methodology holds up.

None of this means a 15-person study is worthless — it might be the only evidence that exists for a newer ingredient, and it can be a legitimate reason to be curious. It means the claim deserves proportionate confidence, not the same weight as a properly powered trial. Bakuchiol is a good working example: the headline trial behind its “retinol alternative” reputation had 44 participants — a genuinely useful data point, not a small one by skincare standards, but still worth reading with the sample size attached rather than the headline alone. A brand citing “clinical studies” for a serum, without saying how many people were in them, is usually hoping you won’t ask.

Funding: not a disqualifier, but a reason to read more carefully

Manufacturer-funded research isn’t automatically invalid — plenty of legitimate research is industry-funded, including for pharmaceuticals. But funding shapes results in predictable ways: a tendency to test under favourable conditions, a publication bias where positive results get published and unflattering ones quietly shelved, and methodological choices — the comparator used, the duration chosen, the outcome measured — that happen to favour the product being sold.

The practical effect is that a study funded by the company selling the product deserves a slightly higher bar before you take its headline claim at face value: was it published somewhere independent reviewers could scrutinise it, was the methodology disclosed in enough detail to assess, and does the actual data support the conclusion, or does the abstract oversell a modest result? Independent research — from universities or public research bodies with no stake in the outcome — carries less of this bias by default, which is one reason it’s worth checking who ran a study before repeating its conclusion. This is exactly the kind of context that gets stripped out by the time a claim reaches a product page: the bakuchiol trial mentioned above, for instance, was funded by a company with a commercial stake in the ingredient — which doesn’t make the finding wrong, but it’s the sort of detail worth knowing before treating it as neutral proof.

Other things worth checking

Was there a control group? Without one, “50% improvement” is meaningless if the placebo group improved 45%.

How were results measured? Instrument-based measurements — hydration meters, imaging-based wrinkle depth, transepidermal water loss — are more reliable than self-report, which is influenced by expectation and easy to shape through how a question is phrased. Expert grading of before/after photos sits somewhere in between, better than self-assessment but still somewhat subjective unless the graders are blinded to which photo is which.

How long did it run? Skin renewal takes weeks and collagen remodelling takes months, so a 4-week study captures a different (and often flimsier) kind of change than a 12-week one. Short studies tend to favour fast, superficial cosmetic effects over genuine structural improvement.

Was it peer-reviewed? Publication in a journal like the British Journal of Dermatology or the International Journal of Cosmetic Science means other scientists assessed the methodology before it went to print. A company website, a press release, or an unpublished “internal study” has had no such check.

The statistical sleights of hand to watch for

The percentage game. “300% improvement” sounds dramatic but means little without the baseline — going from a score of 1 to 4 on a 100-point scale is technically a 300% increase and practically nothing.

Statistical significance vs. visible significance. “Statistically significant” means a result is unlikely to be chance; it says nothing about whether the effect is large enough to notice. A statistically significant 5% improvement in hydration might be real and still be invisible in the mirror.

Cherry-picked endpoints. Studies often measure several outcomes at once. If four show nothing and one shows an effect, that one tends to be the headline. Look for the study’s stated primary endpoint — what it was actually designed to measure — rather than whatever secondary result reads best.

The subgroup shuffle. “Particularly effective in women over 50 with combination skin” can be an honest finding, or it can mean the product didn’t do much for the group as a whole and a flattering slice was found after the fact.

Reading an abstract in under two minutes

Most people won’t read a full paper, and you don’t need to — the abstract usually carries enough. In the methods, check the participant count, whether there was a control group, the study duration, and what was actually measured. In the results, look at the real numbers rather than the framing: what happened in the control group, and were the differences statistically significant. In the conclusion, check that the claims made actually match the results reported — a surprising number of abstracts oversell a modest finding in the closing sentence.

Free places to look: PubMed for abstracts on almost any topic, PubMed Central for full-text open-access papers, and Google Scholar for a broader search that sometimes turns up full text elsewhere. Searching an ingredient name alongside “clinical trial” or “randomized controlled trial” usually gets you further than the brand’s own claims page.

When the evidence genuinely isn’t there yet

For a lot of skincare ingredients, high-quality evidence simply doesn’t exist yet — not because the ingredient doesn’t work, but because nobody’s funded the trial. In that situation, calibrate confidence rather than picking a side: ingredients with in vitro evidence and a plausible mechanism are reasonable to try, at low confidence; ingredients with a few small human studies that broadly agree deserve moderate confidence; ingredients backed by large, replicated RCTs deserve high confidence. Don’t dismiss everything short of a landmark trial, but don’t hand it the same trust either.

The practical version

When a product claims to be clinically proven: ask for the study details, check its size and whether it had a control group, note who funded it, look at what was actually measured rather than what’s implied, and compare the specific claim to the specific result. Most marketing claims soften considerably under that scrutiny — usually not because anyone lied, but because “clinically tested” can legitimately describe almost anything from a rigorous trial to ten people and a questionnaire. Reading studies critically isn’t about rejecting the science behind skincare. It’s about holding the marketing to the standard the science itself already sets.