Which Moisturiser Ingredients Are Overrated, and Which Four Actually Work

Four ingredients keep turning up on the front of moisturiser jars with more promise attached to them than the published evidence carries: collagen, activated charcoal, snail mucin and plant stem cells. None of them is a scam. Two of them do something measurable. What they share is a gap between the claim on the box and the trial data behind it, and that gap is where your money goes.

Below is what each one has actually been tested for, the four ingredients that hold up when you look for the trials, and how to read a label so you can tell the difference yourself.

What overrated means here

An ingredient earns the word when the marketing claim outruns the evidence, not when the ingredient does nothing. The test used throughout this article is narrow and checkable: has the ingredient been tested on human skin, in people, against a control, with the result published? Laboratory work on cultured cells is a reason to run a trial. It is not a reason to make a claim on a jar.

That standard is not invented for this article. Regulation (EU) 655/2013, which sets the common criteria for cosmetic claims across the European market, states plainly that “Ingredient claims referring to the properties of a specific ingredient shall not imply that the finished product has the same properties when it does not,” and that claims “shall be supported by adequate and verifiable evidence.” A jar of cream that says “with collagen” is making an ingredient claim. The rule says the jar must not imply the cream does what pure collagen does in a laboratory.

Two more things worth naming before the list. Price tracks brand positioning, not concentration. And “natural” carries no safety meaning at all, since plant extracts are among the more common causes of allergic contact dermatitis in cosmetics.

Collagen in a moisturiser

Collagen makes up at least 70% of the dry weight of human skin, according to DermNet’s review of the protein by Dr Louise Reiche. Skin ageing involves less collagen being made and more of it being broken down, so putting collagen back sounds like the obvious fix.

The obvious fix runs into the size of the molecule. DermNet’s summary of the evidence is direct: studies using oral supplements of collagen hydrolysates “have typically shown better results than large collagen molecules and topical formulations, for reducing signs of skin dryness and wrinkling.” Note what that sentence does and does not say. It does not say topical collagen is inert. It says the format with the weaker results is the one on the jar.

A PubMed search on 3 August 2026 for the phrase “collagen cream” in titles and abstracts, combined with skin, returned zero records. Not one weak trial, not one negative trial. Nothing. The trials that do exist and use the word collagen study something else: injected fillers, recombinant collagen delivered by needle, and signal peptides that are fragments of collagen rather than collagen itself. Those are different products with different evidence.

What a collagen cream reliably does is moisturise, because it is a moisturiser. Hydrated skin looks smoother and fine lines look shallower for as long as the hydration lasts. That effect is real, it is worth having, and you can buy it for a fraction of the price in a formula that does not carry the word collagen on the front.

Activated charcoal

Charcoal genuinely adsorbs. That property is why activated charcoal is given in hospital for certain poisonings, and it is the entire basis of the detox story told on face masks and cleansers.

The story does not survive contact with the search results. A PubMed search on 3 August 2026 for activated charcoal in titles and abstracts, combined with skin and filtered to randomised controlled trials, returned exactly one record: a 2005 study of an activated charcoal and silver dressing on chronic wounds. A wound dressing left in contact with a wound for days is not a cleanser rinsed off in forty seconds. Widening the search to charcoal with face masks, cleansers or acne returned 18 records, and the first ten included nanocomposites for virus adsorption, microplastics in facial cleansers and a study of cattle traders in Uganda. Not one clinical study of charcoal skincare in people.

There is a systematic review, published in 2022, of activated charcoal as a tooth-whitening agent. It covers in vitro studies only, and its concern is abrasiveness. That is the closest the literature comes to testing charcoal in a cosmetic, and it is a different body site with an unflattering finding.

The cleaning in a charcoal cleanser comes from the surfactants, which are in every cleanser. The black powder is a colour.

Snail mucin, the fairest case on this list

Snail secretion filtrate deserves better than the dismissal it usually gets in articles like this one, because it has been tested and it passed.

A two-centre, double-blind trial published in 2013 enrolled 25 patients with moderate to severe facial photodamage. Each applied an emulsion containing 8% secretion of the snail Cryptomphalus aspersa plus a 40% serum to one side of the face and placebo to the other, for 12 weeks. Periocular wrinkles on the treated side improved significantly (P=0.03), and texture improved more than placebo at 8 and 12 weeks. The split-face design is a strength: each participant is their own control.

A second randomised, placebo-controlled trial in 2022 tested a serum combining snail secretion filtrate with calendula and liquorice root extract in 66 people with mild to moderate mask-related acne. Inflammatory lesions fell more than placebo at every time point, with a coefficient of −33.89% (95% CI −65.24 to −2.53; p=0.03). Nothing else moved: no significant difference in non-inflammatory lesions, sebum, hydration, water loss or participant satisfaction. And the serum contained three active botanicals, so the result cannot be assigned to the snail.

Two small trials, one of them on a mixture, on cosmetic endpoints, several years apart. That is a real evidence base and a thin one. It supports “may help fine lines and inflammatory spots” and it does not support regeneration, repair or any of the vocabulary that sells it.

Plant stem cells

Plant cells and human cells are not interchangeable, and no amount of processing makes a plant stem cell into a skin cell. What ends up in the jar is a plant extract produced by cell culture, and some of those extracts do have antioxidant or anti-inflammatory activity in the laboratory.

A PubMed search on 3 August 2026 for plant stem cells in titles and abstracts, combined with cosmetics or skin and filtered to randomised controlled trials, returned zero records. The closest clinical study, published in 2024, enrolled 104 women through menopause and had them apply a meristem-cell serum daily for one month. It was randomised and open, meaning nobody was blinded and there was no placebo arm. It reported improvements in elasticity, moisture and collagen content. With no blinding, no placebo and a one-month run, that design cannot separate the extract from the base cream, from the ritual of applying something twice a day, or from what participants expected to see.

A 2020 review in the cosmetic science literature framed the field as promising and unproven, closing on “the challenges that we need to overcome in order to see meaningful changes in human skin.” Six years later, the randomised evidence in humans is still where that review left it.

The four that hold up

Four things have repeatedly changed measurable outcomes in people, in trials with controls, published and citable: retinoids, vitamin C, niacinamide, and the humectant-plus-occlusive base that most people ignore because it is printed at the bottom of the ingredient list. Everything below is dosed, tested and boring, which is precisely why it works.

Retinoids, and the strength question nobody answers on the box

Retinoic acid increases type I, III and VII collagen in the dermis and reorganises collagen bundles, and DermNet’s review states that in clinical trials topical retinoids were clinically effective for dermal ageing, reducing wrinkles, roughness and laxity.

The size of that effect, and its cost, comes from a 48-week double-blind study of 99 photoaged patients published in 1995. Tretinoin at 0.1% and at 0.025% both produced statistically significant improvement in facial photoaging compared with vehicle, and there was no clinically or statistically significant difference between the two concentrations. Epidermal thickening was 30% on the higher strength and 28% on the lower, against an 11% decrease on vehicle. Irritation, measured as redness and scaling, was significantly worse on 0.1%. Four times the concentration bought no extra benefit and a good deal more discomfort.

Then there is the question everyone asks about over-the-counter retinol. A randomised, double-blind trial published in JAMA Dermatology in 2022 put a 1.1% formulation of retinol and retinyl esters against prescription tretinoin 0.02% for 24 weeks in patients with moderate to severe facial photodamage. Photoaging scores at week 24 showed no significant difference (median 4 versus 5; difference −1; 95% CI −2 to 1; P=0.27), while erythema occurred about six times less often on the over-the-counter formula (11% versus 64%; P=0.01). Only 20 participants were analysed, so the confidence interval is wide and the finding needs replication before anyone treats it as settled. As a starting point it is still useful: gentler formulas are not automatically weaker ones.

Vitamin C

A double-blind, randomised trial ran a 5% vitamin C cream against its own excipient for six months on photoaged skin of the low neck and arms. Assessment by the investigating dermatologist and self-assessment by volunteers both showed significant improvement in the global score on the treated side. Silicone replicas showed a highly significant increase in the density of skin microrelief and a decrease in deep furrows, and biopsies at the end showed evidence of elastic tissue repair.

Two practical notes the marketing skips. The trial tested 5%, not the highest number a brand can print on a box. And vitamin C oxidises, which is why a serum that has turned deep orange has less of what you paid for than the day it was opened.

Niacinamide

Niacinamide is the ingredient with the widest set of tested endpoints and the least dramatic marketing.

In a double-blind, left-right randomised study, 50 women with facial photoaging applied 5% niacinamide to one half of the face and its vehicle to the other twice daily for 12 weeks. The treated side showed reductions in fine lines and wrinkles, hyperpigmented spots, red blotchiness and yellowing, plus improved elasticity on cutometry. For pigment specifically, a separate set of trials found niacinamide inhibited melanosome transfer by 35–68% in a coculture model, and significantly reduced hyperpigmentation and increased skin lightness against vehicle after four weeks in 18 people, with a further 120-participant arm on facial tanning.

Oily skin is the one place the answer is qualified. Testing 2% niacinamide, a Japanese trial of 100 participants found significantly lowered sebum excretion rate at weeks 2 and 4, while a 30-participant study in Caucasian volunteers found casual sebum levels fell over six weeks but the excretion rate did not. Same ingredient, same concentration, different populations, different results. That is what an honest sebum claim looks like.

The base does most of the work

Search interest in “non-active” ingredients is usually an attempt to find out what the filler does. The answer is that in a moisturiser, the filler is the product.

A randomised, double-blind crossover study in 51 volunteers with dry skin separated an emollient into its parts and measured each. Petrolatum was the component that improved barrier function by reducing water loss through the skin. Glycerol was the component that improved hydration. Neither did the other’s job, and the full formula did both.

Glycerol has been tested on its own in disease, too. In a placebo-controlled, double-blind study, 24 patients with atopic dermatitis used a 20% glycerol preparation or its vehicle twice daily for four weeks. Hydration and barrier function improved significantly against the glycerol-free placebo, while redness, SCORAD and local severity scores did not differ. It moisturises. It does not treat the eczema.

Hyaluronic acid belongs here rather than among the miracle actives, and molecular weight turns out to matter. A double-blind randomised trial in 36 nursing-home residents aged 60 to 80 applied low-molecular-weight hyaluronic acid, high-molecular-weight hyaluronic acid or vehicle to three separate sites on the leg for four weeks. The low-weight version gave higher skin capacitance than the high-weight version (56.37 versus 52.37 AU, p=0.004) and than vehicle (56.37 versus 49.01 AU, p<0.001). Water loss and symptom scores showed no difference between any of the three. Labels almost never state which molecular weight is in the jar.

Ceramides earn their place alongside a treatment rather than instead of one. In an eight-week split-face randomised trial, 40 patients with mild to moderate acne used 5% benzoyl peroxide and 0.1% adapalene on the whole face, then a ceramide-and-niacinamide moisturiser on one side and a plain hydrophilic cream on the other. The ceramide side showed significantly greater improvement in inflammatory, non-inflammatory and total lesion counts by week 8.

What gets misrepresented about fine lines

Fine lines attract more claim inflation than any other concern, and the pattern is consistent enough to name.

The first move is treating hydration as correction. A humectant plumps the surface, fine lines look shallower within an hour, and the effect ends when the skin dries. Nothing was corrected. The second is transferring a laboratory result to the jar, which is what the EU truthfulness rule prohibits, and what “with collagen” does. The third is quoting a concentration without a study, since 20% of an ingredient tested at 5% is not four times the evidence.

The fourth is the quietest. Photographs taken under different lighting, or with the face relaxed rather than expressive, will show a difference that no ingredient produced. Ask what the comparison was. Trials that mean something use a control arm, a blinded assessor, and often the other half of the same face.

How to read a moisturiser label

Under Regulation (EU) 1223/2009, Article 19(1)(g), “The list of ingredients shall be established in descending order of weight of the ingredients at the time they are added to the cosmetic product. Ingredients in concentrations of less than 1 % may be listed in any order after those in concentrations of more than 1 %.”

Everything follows from that sentence. Order is informative at the top of the list and meaningless at the bottom, and the actual percentages are never declared. An ingredient sitting after the preservatives is present below 1%, whatever the front of the box implies about it. A hero ingredient printed in large type and listed second-to-last is doing very little.

So: read the first five ingredients, because those are the product. Look for a humectant such as glycerin and an occlusive such as petrolatum or dimethicone near the top. Treat a stated concentration as a claim you can check against a trial, and an unstated one as no information at all. If a formula lists fragrance and your skin reacts to things, choose one that does not.

Risks, and who should not use these

The four that work are not risk-free, and the trials that proved they work also recorded what they cost.

Topical retinoids can cause retinoid dermatitis with redness, peeling and dry skin, irritant contact dermatitis, sun sensitivity and, rarely, allergic contact dermatitis. They can make eczema worse because they dry the skin. Some topical retinoids are contraindicated in pregnancy, and DermNet states that women of childbearing age must use effective contraception because of the teratogenic effects of retinoids; tazarotene is pregnancy category X in the United States. They are generally not recommended for young children, and adapalene is the least irritating of them while tretinoin is the most. In Dubai sun, the photosensitivity point is not academic: a retinoid at night needs sun protection during the day.

Vitamin C in the low-pH formulations that stay stable can sting on compromised skin. Niacinamide is among the better-tolerated actives and still causes flushing in some people. Ceramide and glycerol formulas are the gentlest of the group, which is why they are what most trials use as the base under everything else.

Two rules cover most trouble. Add one new active at a time, so that a reaction tells you what caused it. And do not apply an acid or a retinoid to skin that is broken, weeping or infected.

When this needs a doctor rather than a different cream

Some skin problems are not ingredient problems, and swapping products delays the answer. See a doctor if a rash spreads, weeps, crusts or becomes painful; if a patch of skin is hot and tender with fever, which can indicate infection; if a mole changes in size, shape, colour or border; if a spot bleeds, will not heal within a few weeks, or keeps returning in the same place; or if acne is leaving scars, since scarring is preventable and permanent once formed.

Swelling of the lips, tongue or throat, difficulty breathing or swallowing, a widespread hive-like rash with faintness or collapse, or vomiting and cramps after a product goes on the skin are not reactions to manage at home. In the UAE, call 998 for an ambulance. Anaphylaxis is treated with adrenaline, and someone feeling faint should lie down with their legs raised until help arrives.