Glamingo Blog

You have seen it on the box: “clinically tested,” “94% of women agreed,” “dermatologist proven.” And if you have been using skincare long enough, you have also noticed that none of it reliably predicts whether a product will do anything for your skin. That is not a coincidence. The standards governing how cosmetic brands test and report their clinical evidence are, at best, inconsistent — and at worst, designed to produce a favourable number with the least rigorous effort possible.

Think about the last serum you bought because of a claim like that. Maybe it was a vitamin C with “92% of participants reported brighter skin in four weeks.” Maybe it was a moisturiser “backed by 8 weeks of clinical testing.” You used it. You waited. Your skin did approximately what it was already doing. And yet the box sits on your shelf, its confident statistics intact, because there is no mechanism that requires those numbers to mean anything beyond the fact that they were gathered. That gap — between what the claim sounds like and what the claim is actually worth — is the thing this article is here to explain.

The Verdict Upfront: ‘Clinically Tested’ Is a Marketing Statement, Not a Quality Standard

What the phrase actually means — and what it is legally required to prove

Here is the short version: “clinically tested” means the brand conducted some form of testing in a clinical-adjacent setting. It does not specify what was tested, how many people were involved, whether there was a control group, whether the results were independently verified, or whether the testing was designed to answer the question “does this product work” rather than “can we find a number that sounds good on packaging.”

Think of cosmetic clinical trial standards like restaurant hygiene grades — except in this system, there is no independent inspector, the restaurant writes its own grade, decides which dishes get assessed and which do not, and is not required to show anyone the kitchen. A grade still appears on the window. It just does not mean what you assume it means.

That is not a cynical metaphor. It is an accurate description of how the industry currently operates. Unlike pharmaceutical drugs, cosmetics are not required to prove efficacy before going to market. A brand can develop a product, conduct a panel study with twenty participants who were asked whether their skin “felt more radiant,” get fourteen of them to say yes, and print “70% of women agreed” on the box. Technically accurate. Clinically meaningless.

How Cosmetic Clinical Trials Actually Work (and Why the Design Gaps Matter)

The difference between an RCT and a self-assessment consumer panel

There is a spectrum of trial quality in skincare research, and most cosmetic brands operate at the lower end of it. At the rigorous end sits the randomised controlled trial (RCT) — the study design in which participants are randomly assigned to either the product being tested or a control (usually a placebo or vehicle cream), neither the participants nor the researchers know who is receiving which treatment (this is the “blinding”), and outcomes are measured using objective tools rather than participant opinion. This design exists to eliminate bias. It is the minimum standard required for a drug to prove efficacy.

At the other end sits the self-assessment consumer panel, where a group of participants — often recruited specifically because they tend to respond positively to skincare products — use the product for a set period and are then asked how their skin feels. No control group. No blinding. No independent measurement. Just a curated sample of people reporting their subjective experience, with results reported as a percentage of those who agreed with a favourable statement.

The gap between these two designs is not subtle. The first is designed to establish whether a product causes an outcome. The second is designed to generate a marketable number.

Why only 0.3–3.7% of skincare studies use the trial design that would satisfy a drug regulator

This is the statistic that should permanently change how you read skincare marketing. Randomised controlled trials make up only 0.3% to 3.7% of published skincare research — meaning the overwhelming majority of cosmetic clinical literature sits below the quality threshold that any drug regulator would require for an efficacy claim. Most of what gets called “clinical evidence” in skincare would not pass review at a pharmaceutical company.

That is not a marginal problem. It means that when a brand says its product is “clinically proven,” they are almost certainly not pointing to an RCT. They are pointing to a self-assessment panel, an open-label study, an instrumental measurement study without a control group, or a trial they designed, funded, and reported themselves. The word “clinical” is doing an enormous amount of work.

The outcome measures brands choose — and the ones they quietly avoid

A systematic review of 70 skincare studies found that effectiveness and safety evidence varies significantly depending on which outcome measures were selected — and that is the operative phrase. Brands choose their outcome measures. They choose whether to measure hydration using a corneometer, or whether to ask participants if their skin “felt softer.” They choose a four-week endpoint rather than twelve weeks. They choose to report the subgroup that responded best rather than the full intention-to-treat population. The outcome measures that end up in the marketing materials are the ones that produced the most useful numbers. The ones that did not are simply not mentioned.

There is no requirement to disclose the measures that showed no effect. Industry experts have explicitly flagged that mandatory disclosure of cosmetic clinical trial results does not currently exist in most markets, meaning brands can run multiple trials, bury the unfavourable results, and publish only the study that told the story they wanted to tell. This is a known, documented problem — and one that has no current enforcement mechanism behind it.

What Singapore’s Regulatory Framework Actually Requires

HSA notification versus proof of efficacy — they are not the same thing

If you are in Singapore and you have ever assumed that the Health Sciences Authority’s oversight extends to verifying skincare efficacy claims, here is the reality check. Every cosmetic supplier must notify HSA for each product placed on the Singapore market — but this notification process does not require brands to submit clinical proof of their efficacy claims before the product can be sold. The notification system exists to ensure product safety and ingredient compliance. It does not assess whether the “firms skin in four weeks” claim has evidence behind it.

That framework is broadly consistent with most markets globally. Cosmetics are not drugs. They are regulated for safety, not efficacy. Which means the entire evidentiary structure behind a product’s performance claims exists in a regulatory vacuum — accountable primarily to consumer protection laws around misleading advertising, and only then if someone actually files a complaint.

What ‘dermatologist tested’ means when there is no standardised testing protocol behind it

“Dermatologist tested” sounds reassuring. A medical professional was involved. There was testing. What it does not tell you is whether the dermatologist designed the protocol, evaluated the results independently, or simply recruited participants and administered the product while the brand handled everything else. There is no standardised protocol that this phrase refers to. A dermatologist can test a product for tolerability — checking that it does not cause adverse reactions in a small group — and the brand can print “dermatologist tested” on the box. That test says nothing about whether the product works. It says only that a small group of people did not immediately react badly to it.

The Funder Problem: When the Brand Pays for the Study

Industry-funded trials in skincare — what the conflict of interest looks like in practice

Even when cosmetic brands do conduct well-designed trials — randomised, double-blind, placebo-controlled — funder conflict of interest is a legitimate concern that does not disappear because the methodology is sound. The funder chooses the research question. They choose the participant profile. They choose the dosage, the duration, the outcome measures, and ultimately control the decision of whether and where to publish. Industry-funded trials using rigorous RCT design do exist for ingredients like topical hyaluronic acid serums and oral nutraceuticals — but these represent the exception, not the norm, and funder conflict must always be factored into how you weigh the results.

This is not unique to the beauty industry. Pharmaceutical research has the same problem. But in pharmaceuticals, there is at minimum a regulatory framework that requires pre-registration of trial protocols — meaning the outcomes the researchers said they would measure are locked in before the study begins, and deviations must be disclosed. No equivalent mechanism exists for cosmetic trials.

How to spot a study where the outcome was probably decided before the first participant enrolled

There are patterns. If every participant saw improvement and the effect size is uniformly positive, that is unusual — real studies produce messy, distributed results with non-responders. If the study period is exactly as long as it takes for the active ingredient to plausibly show an effect, but no longer, that is a calibrated choice. If the study has no registered protocol, no control group, and was conducted entirely by the brand’s own research team, the study was designed to generate evidence, not test a hypothesis. These are not disqualifying factors on their own, but together they describe a very specific kind of research: the kind that is useful for packaging copy and not much else.

The Exceptions — Ingredients Where the Trial Evidence Is Genuinely Better

Retinoids, hyaluronic acid, and a small number of actives with independent replication

This is not a blanket indictment of all cosmetic efficacy research. A small number of ingredients have accumulated genuinely meaningful evidence — not because the industry is transparent, but because independent researchers have studied them, the mechanism of action is well understood, and the results have been replicated across multiple studies with different funders.

Vitamin A derivatives — specifically prescription-grade retinoids — represent the clearest example. The mechanism by which they accelerate skin cell turnover (what dermatologists call cellular desquamation and epidermal renewal) is well documented, the results at prescription concentrations are consistent across independent studies, and they remain the benchmark against which most other anti-ageing ingredients are compared. But even here, caution applies. A critical appraisal of existing randomised, double-blind, vehicle-controlled trials for over-the-counter retinol products found meaningful gaps in blinding quality and independent replication — meaning even for the category’s best-evidenced active, the OTC version of the story is less solid than the prescription evidence would suggest.

There is also a useful reference point from prescription retinoid research that reframes what “clinically proven improvement” actually looks like at its most rigorous. In a properly controlled trial for a prescription retinoid, a meaningful proportion of participants — in some published data, around 10% — showed only moderate improvement in fine line appearance. If that is the number for one of the most studied topical actives in the entire category, then “94% of women agreed their skin looked more radiant” on a department store serum deserves considerably more scrutiny than most of us give it.

Evidence grade: what ‘moderate’ versus ‘strong’ actually means for your routine

When you encounter the language of evidence grades — strong, moderate, limited, weak — here is what they mean in practice. Strong evidence means the finding has been replicated across multiple independent, well-designed trials and the effect is consistent. Moderate evidence means the mechanism is understood and some good studies exist, but there are gaps in methodology, sample size, or independent replication that prevent full confidence. Limited evidence means the research is early, mechanistic, or relies heavily on brand-funded trials. Weak evidence means the claim is based on expert opinion, consumer perception data, or small studies with significant design problems.

Most cosmetic efficacy claims sit at limited or weak. A small number of actives sit at moderate. Very few reach strong — and when they do, it is often for mechanisms studied in medical rather than cosmetic contexts. Some more rigorous cosmetic trials now use 3D photography for precise outcome measurement — a methodological step up from self-reported outcomes — but these remain the exception in consumer skincare research, not the standard.

How to Read a Cosmetic Efficacy Claim Like Someone Who Has Seen the Inside of a Study

Six questions to ask before a clinical claim changes what you spend

The first question is the most important: was the study randomised and controlled, or was it an open-label panel where everyone received the product? If there was no control group, there is no way to separate the product’s effect from the placebo effect, the seasonal change in skin, or the fact that participants who volunteer for skincare studies tend to take better care of their skin throughout the trial period.

The second question: who funded it? A brand-funded study conducted by the brand’s own research team is not independent evidence. It is an internally generated number. It may be accurate. It may not be. There is no way to know, because the raw data is not published.

Third: how was the outcome measured — by participants rating their own skin, or by an objective instrumental measurement? Self-reported outcomes are influenced by expectation, by the experience of using a product that feels luxurious, and by the subtle pressure of knowing you have been selected to test something. A corneometer reading or a standardised photographic assessment with blinded evaluators is a different kind of evidence.

Fourth: what was the sample size, and how was it recruited? Twenty participants recruited through a brand’s loyalty programme is a different population from 200 participants recruited from the general public across multiple centres. A comprehensive comparative review of clinical trials for FDA-cleared dermal fillers found significant inconsistencies in data reporting — demonstrating that even regulatory review does not guarantee clean or comparable underlying numbers. If a product’s marketing does not specify sample size, that absence is informative.

Fifth: what was the duration? Four weeks tells you something about early-stage hydration and tolerability. It tells you very little about whether an ingredient is delivering structural change, which typically requires a minimum of twelve weeks to assess.

Sixth: were the results statistically significant — and was the effect size actually meaningful? A statistically significant result in a skincare study can still represent a change so small that no human being would notice it in the mirror. The study found an effect. The study did not necessarily find an effect that matters.

The Ruling: Worth It, Skip It, or Investigate Further?

The ruling here is nuanced, because this is not a verdict on whether clinical claims in skincare are always meaningless — it is a verdict on what they are actually worth, read accurately. A small number of ingredients, studied independently over time, have earned moderate to strong confidence. Retinoids at prescription concentration. Hyaluronic acid for surface hydration. Broad-spectrum UV filters with SPF rating methodology that is, notably, more standardised than most cosmetic efficacy testing. These are worth your trust.

The broader category of “clinically tested” marketing language? Investigate further — every time. Not because brands are universally dishonest, but because the same ingredient can appear both “proven” and “inconclusive” in the literature depending on how the trial was designed, who funded it, and what outcome measures were chosen. The claim on the box is not the study. It is the brand’s interpretation of the study, filtered through a marketing department.

Skipping the claim entirely is not the answer either. Some products do work. Some studies are well-designed. The skill is not cynicism — it is the specific literacy to tell the difference between evidence that was built to withstand scrutiny and evidence that was built to print well on packaging. In Singapore, where the skincare market is saturated with products at every price point making broadly similar claims, that literacy is worth considerably more than any individual product you might buy with it.

The next time a skincare brand uses “clinically tested,” “9 out of 10 women agreed,” or “dermatologist proven” as a purchase signal, ask one specific question before it influences your decision: was the study randomised, blinded, and independently conducted — or did the brand design, fund, and report the results itself? If the brand’s marketing does not answer that question directly, treat the claim as unverified consumer perception data, not clinical evidence.

If this article has you thinking about treatments where efficacy standards are held to a higher bar — such as medical-grade facials, laser treatments, or professionally administered actives — Glamingo has verified providers near you with real reviews from women who have actually been through the door. Browse treatments near you on Glamingo. Search now →