Methodology

How we evaluate evidence.

521 citations across 12 databases, each one a specific source you can check. We grade every study by what it actually found, then sort the evidence into tiers. Below is the framework, the source list, and the rules we follow.

What our citations show

305
Positive efficacy

Single trials or meta-analyses concluding the ingredient works clinically.

46
Safety / regulatory

Cosmetic Ingredient Review (CIR), Scientific Committee on Consumer Safety (SCCS), or Korean Ministry of Food and Drug Safety (MFDS) documents. Safe to use at typical concentrations.

155
Mechanism only

Shows how the ingredient works in lab tests (in vitro), not proof it works on real skin.

15
Mixed / inconclusive

Reviews that explicitly say "trials needed" or "benefit unclear". We list them honestly.

We cite specific papers for all 111 ingredients we track.

Strong
15+ studies or large cohort

Multiple randomized controlled trials (RCTs), a systematic review, or a large longitudinal cohort. Mechanism confirmed at molecular level. Effect size meaningful at typical concentrations.

E.G. Salicylic Acid · Azelaic Acid · Tranexamic Acid
Moderate
5–14 studies

Several controlled trials, but smaller samples or mixed results. Mechanism plausible and partially confirmed.

E.G. Snail Mucin · Centella · Mugwort
Anecdotal
No clinical proof yet

Used in tradition or reported anecdotally to help, but we don't yet have controlled clinical evidence. May work for some people; we track these as the research catches up.

E.G. Yuzu Seed Extract

Beyond Count: The Concentration Gate

While an ingredient might have 50+ studies (Strong Tier), it only works if it's present in the product at the correct concentration. Korean Pharmacists (Yaksa) frequently highlight that many "cosmetic" products use trace amounts ("fairy dust") for marketing, while "Clinical Grade" products meet the specific PPM (Parts Per Million) thresholds proven in clinical papers.

Products with the ✨ badge have either disclosed concentration matching clinical papers or have been verified as pharmacy-grade staples.

Our 12 sources

We pull from databases that cover skincare research and regulators worldwide, well beyond Western dermatology. Each name links to the underlying database or regulator's own site:

🌐
PubMed
US National Library of Medicine, the biomedical literature index
476
🌐
Cochrane
Independent systematic reviews of clinical evidence
6
🇰🇷
KCI
Korea’s national index of peer-reviewed academic journals
24
🇰🇷
KoreaMed
Korean medical journal aggregator that indexes Annals of Dermatology
5
🇯🇵
J-Stage
Japan’s national database of scientific & medical journals
17
🇨🇳
CNKI
China’s largest academic literature database
3
🇨🇳
Wanfang
Chinese medical, dissertation & trial registry that complements CNKI
2
🇨🇳
VIP
Chinese scientific journals, the strongest derm coverage of the three
4
🇫🇷
SFD
French dermatology society, publisher of Annales de Dermatologie et de Vénéréologie
7
🇰🇷
MFDS
Korean regulator that approves cosmetic functional actives
17
🇪🇺
SCCS
EU regulator’s cosmetic-safety panel, with opinions on actives + restrictions
5
🇺🇸
CIR
Independent US cosmetic-ingredient safety review panel
12

Counts grow as we add citations on a rolling basis. A dash (—) means the database is part of our planned coverage but we haven't yet sourced a specific citation from it. Chinese (CNKI, Wanfang) and French (SFD / Annales de Dermatologie) coverage is in active expansion for K-beauty botanicals and pigment-correction research.

How an ingredient gets in the database

  1. Two editors independently identify candidate ingredients from journal scans + reader requests.
  2. We pull every relevant citation (RCTs, in-vitro, in-vivo) and tag by study quality.
  3. An ingredient needs at least 3 peer-reviewed sourcesacross at least two databases before it's added.
  4. Tier is assigned by total study count + presence of RCTs + mechanism confirmation.
  5. Every claim on every page must link back to one of the cited papers. No exceptions.

How tiers can change

An emerging ingredient moves to "moderate" once it crosses 5 peer-reviewed studies including at least one controlled trial. Moderate moves to "strong" when one of two thresholds is met: 15+ studies including multiple RCTs, or a single large longitudinal cohort study (something like 50+ subjects followed for 12+ months with controlled measurement against baseline). One solid real-world cohort can carry as much weight as a stack of small trials when the methodology is sound. We re-review every ingredient in the database quarterly.

What we don't do

  • We don't accept industry-funded studies as the sole source for a tier upgrade.
  • We don't cite blogs, Reddit threads, or product marketing pages.
  • We don't cite an LLM as a source. Ever.