How much retinol do you need? 0.075 to 0.3%

The effective retinol percentage, by the vehicle-controlled trials, is 0.075 to 0.3%; above the 0.3% ceiling, more stopped adding effect or started adding irritation.

Give it four to eight weeks: that is when the change first became measurable.

how it was measured

Measured on investigator-graded photoaging parameters versus vehicle - lines and wrinkles, pigmentation, elasticity, firmness and overall photodamage on a 0-9 scale; the first visit at which the change was measurable.PMID 19852122PMID 17515510

See it on the timeline, beside the others →

Dose line for Retinol: the trials worked from 0.075 to 0.3%; ticks at 0.04%, 0.075%, 0.1%, 0.104%, 0.247%, 0.3%, 0.5%, 1% where the named trials worked; the highest dose tested 1% (Mellody 2022): more irritation; EU cap from 2027 at 0.3%; catalogue labels at 0.5, 1, 0.3%. Logarithmic scale.0.04%, Kikuchi 20090.075%, Kikuchi 20090.1%, Tucker-Samaras 20090.104%, Pordab 20260.247%, Pordab 20260.3%, Zasada 20200.5%, Zasada 20201%, the highest dose tested in leave-on use on human skin (Mellody 2022): more irritation0.04%0.075%0.1%0.104%0.247%0.3%0.5%1%more irritationKikuchi 2009Kikuchi 2009Tucker-Samaras 2009Pordab 2026Pordab 2026Zasada 2020Zasada 2020highest tested, Mellody 2022Paula's Choice Clinical 0.3% Retinol + 2% Bakuchiol Treatment: 0.3%The Ordinary Retinol 0.5% in Squalane: 0.5% (outside the trials’ range)The Ordinary Retinol 1% in Squalane: 1% (outside the trials’ range)EU cap from 2027 0.3%0.025%

0.04%, Kikuchi 2009.

Floor 0.075%: a stabilised 0.1% retinol moisturiser beat its vehicle on lines, pigmentation and firmness at 8 weeks in 64 women (Tucker-Samaras 2009), in a pooled analysis of 6 vehicle-controlled studies, 471 women (Farris 2024); 0.075% beat vehicle on fine wrinkling in 54 Japanese women over 26 weeks while 0.04% did less (Kikuchi 2009).

The rose band is where the vehicle-controlled trials worked; each tick is a concentration a named trial used, with its author and year; the last tick is the highest dose tested in leave-on use on human skin, 1% (Mellody 2022), where the paper reports: more irritation; the dashed line is the regulator’s number (EU cap from 2027); the bottles above the line are the products in our catalogue that declare a percentage on the name, hollow where it falls outside the trials’ range. The scale is logarithmic because the values span more than twenty-fold (0.04 to 1%).

What retinol percentage is effective?

Floor 0.075%: a stabilised 0.1% retinol moisturiser beat its vehicle on lines, pigmentation and firmness at 8 weeks in 64 women (Tucker-Samaras 2009), in a pooled analysis of 6 vehicle-controlled studies, 471 women (Farris 2024); 0.075% beat vehicle on fine wrinkling in 54 Japanese women over 26 weeks while 0.04% did less (Kikuchi 2009).

0.075 to 0.3%

The trials, and what they used
  • Ceiling 0.3%: in split-face trials 0.3% and 0.5% gave the same benefit and 0.5% more irritation in 37 volunteers over 12 weeks (Zasada 2020), 0.104% and 0.247% did not differ in 28 women over 12 weeks (Pordab 2026), and 0.3% and 1% remodelled photoaged skin equally under 12-day patches in 5 people while 1% was worse tolerated by 218 (Mellody 2022).
  • EU Regulation 2024/996 caps retinol at 0.3% in face products and 0.05% in body lotions from November 2027.

Is 0.5% retinol too much?

0.5% is above the 0.3% ceiling: the trials that compared doses found that above it more stopped adding effect or started adding irritation, so a declared 0.5% reads as above the range, not as stronger. The highest dose tested in leave-on use was 1% (Mellody 2022), and the paper reports more irritation there.

The in-use study found that 0.3% retinol was better tolerated than 1% retinol, with fewer and milder adverse events reported (χ2 (1) = 23.97; p < 0.001).

The highest dose tested, in the paper’s words
  • The in-use study found that 0.3% retinol was better tolerated than 1% retinol, with fewer and milder adverse events reported (χ2 (1) = 23.97; p < 0.001).

Neither number settles what more does. The trials that compared doses put the ceiling at 0.3%: above it, more stopped adding effect or started adding irritation, which is why a declared percentage above 0.3% reads as above the range, not as stronger.

What does the retinol percentage on a label tell you?

A name that states a percentage ("Retinol 0.3%") is the only amount a cosmetic label ever declares, so it is the one number that can be read against the trials: inside 0.075 to 0.3%, above the 0.3% ceiling, or below the 0.075% floor.

A name without a percentage leaves the position. The rules make a brand list ingredients in descending order of amount down to 1%, and in any order below it, so retinol above the 1% line is at least 1% and retinol below it can be any trace the brand likes. The floor the trials worked from is 0.075%, and a position alone cannot tell you whether a label reaches it.

The retinol percentages in our catalogue

The 6 products in our catalogue that carry retinol: the number is the percentage the name declares, or the position on the label where it declares none. A rose mark is a declared percentage outside the range the trials used.

3 labels declare a percentage on the name, read against the trials:

3 labels carry it without a number, so the position is all there is:

Read on

Other actives with a dose the trials pinned:

The range column was researched and verified in September 2026 (this row last changed 22 September 2026). The number is the trials' and the reasoning is the knowledge base's; neither is advice about your skin. How we score.