If you've typed a product into EWG's Skin Deep database, watched it return a number from 1 to 10, and felt either relief or alarm — you've used one of the most influential tools in clean beauty. Whether that number answers the question you're actually asking turns on one distinction: hazard versus risk.
Quick Answer
Skin Deep scores the potential hazard of ingredients, not the risk of using a product. EWG states this itself — its ratings don't account for exposure level or individual susceptibility, which are the factors that determine real-world risk.
So it's genuinely useful for looking up ingredients and spotting personal allergens, and incomplete as a standalone safety verdict — because a high hazard score doesn't mean high risk at the dose actually used.
In this article
01 — The tool
What Skin Deep Actually Is
A free online database from the Environmental Working Group, a non-profit research and advocacy organisation. Look up a product or ingredient and it returns a hazard score from 1 to 10, alongside a data-availability indicator showing how much research sits behind the rating.
Its appeal is obvious: fast, free, and it gives a single number where the underlying science is genuinely complicated. That's also the problem — a single number invites people to read it as a verdict, which isn't what it's built to be.
02 — The core distinction
Hazard Versus Risk
Hazard is the capacity of a substance to cause harm under some conditions. Risk is the probability of harm given the actual exposure — how much, how often, by what route, and to whom.
And the decisive point comes from EWG itself. Its own materials explain that Skin Deep ratings reflect potential hazards but don't account for exposure level or individual susceptibility — the factors that determine actual risk. The database is transparent that it scores the left-hand box, not the result.
This is the principle running through this whole series: the dose makes the poison. Anything is a hazard at a high enough dose, water included. What protects you is that real exposure usually sits far below where hazard becomes risk.
Myth
A high Skin Deep score means the product is dangerous and should be thrown out.
Fact
It flags a potential hazard, not a demonstrated risk. Because the score excludes exposure, a high-hazard ingredient present in trace amounts may pose very little real-world risk.
03 — The method
Where The Method Strains
| Feature | Why it can distort the picture |
|---|---|
| Hazard-only basis | Omits exposure entirely, so it can't express real-world risk for a finished product. |
| Data-gap sensitivity | Ingredients with little published data can be rated cautiously — a "limited data" 7 means far less than a "robust data" 3. |
| Penetration-enhancer weighting | Products score higher for containing penetration enhancers, largely regardless of what is being delivered. |
| Naming inconsistencies | Chemically identical materials have been reported scoring differently under different names or listings. |
A note on the examples that circulate. The petrolatum case — the raw ingredient rated as a moderate concern while products that are essentially pure petrolatum score low — is widely cited, and it does illustrate how listing rather than chemistry can move a number. But scores change over time as EWG updates its database, so check any specific example before repeating it rather than relying on a screenshot from some years ago.
04 — The criticism
What Critics Say — And How Good That Evidence Is
Toxicologists, cosmetic chemists and science communicators have criticised Skin Deep for weighting hazard over risk-based assessment. That critique is coherent on its own terms — it follows directly from what the score does and doesn't include.
Holding our own evidence to the same standard
The figure most often quoted in this debate is a 2009 survey of Society of Toxicology members, in which a majority of those familiar with EWG felt it overstated risk. It gets cited constantly — including, until this revision, by us.
It's worth being straight about its limits. It is sixteen years old. It was conducted by an organisation whose funding and framing drew questions at the time. And the headline sample size refers to the whole survey, not to the subset who actually rated EWG.
None of that makes it worthless — but an article arguing that people should check what a number really measures has to apply that standard to its own citations too. The hazard-versus-risk critique stands on the logic, not on one old survey.
05 — The other side
The Case For EWG
Most critiques of Skin Deep stop at the criticism. That's a weaker article than one that states the opposing case properly, so here it is at its strongest.
Regulatory safety assessment leans heavily on industry-supplied data. Manufacturers generate and submit the studies. That's normal practice and not in itself sinister, but it does mean the system's independence isn't total — and EWG's existence as an outside check has value precisely because of that.
Some ingredients genuinely are under-studied. Long-term data, cumulative exposure across many products, and effects on vulnerable groups are all thinner than most people assume. A data gap isn't proof of safety any more than it's proof of harm — and a system that treats "not yet studied" as equivalent to "shown to be safe" has a real weakness.
Precaution has legitimate value where data is absent. Reasonable people weigh unknowns differently, and someone who prefers to avoid an ingredient with limited data isn't being irrational — they're applying a different threshold, which is a values judgement rather than a scientific error.
Where that leaves the score
The hazard-versus-risk criticism holds — a hazard score cannot answer a risk question, and EWG says as much itself. But "the score has a real limitation" is a different claim from "the organisation is worthless," and only the first is supportable. Skin Deep is a hazard-flagging tool used correctly, and a fear-generating one used as a verdict. The distinction is in how it's read, not in whether it should exist.
06 — Fair credit
What It's Genuinely Good For
- Looking up what's in a product — a fast, searchable index of ingredients and compositions.
- Spotting personal allergens — if you know you react to something, this helps you find it.
- Comparing formulations at a glance.
- Pushing transparency — its existence has pressured brands toward fuller disclosure, which benefits everyone.
Used as a reference library, Skin Deep earns its place. The trouble starts only when a hazard score is treated as the last word.
07 — Practical guidance
How To Use It Without The Fear
| Instead of… | Try… |
|---|---|
| Treating the score as "safe versus toxic" | Reading it as a hazard flag that still needs exposure context |
| Reacting to the number alone | Checking the data-availability rating alongside it |
| Trusting the headline rating | Looking at which specific ingredient is driving it |
| Using it as your only source | Cross-checking against expert assessments that include exposure |
For readers in India specifically: cosmetic ingredients are regulated here under the CDSCO framework and the Cosmetics Rules 2020, which sit alongside the EU and US systems most international databases reference. A product compliant in one market isn't automatically compliant in another — worth knowing when a database's verdict and a label's claims disagree.
FAQ
Questions, Answered
The Bottom Line
Skin Deep is a hazard flag, not a safety verdict — and EWG says so itself. Read it as one input, check the data-availability rating, and look at which ingredient is driving the number. The answer isn't fear and it isn't blind trust in any source, including this one — which is why we've flagged the limits of our own citations above, and of our own study below.
Sources, by type
Peer-reviewed literature
- Grandjean P. Paracelsus revisited: the dose concept in a complex world. Basic & Clinical Pharmacology & Toxicology. 2016;119(2):126–132.
- Frontiers in Toxicology. The impact of perfumes and cosmetic products on human health: a narrative review. (2025 — please verify DOI and publication status before publishing)
Regulatory and expert-panel sources
- Scientific Committee on Consumer Safety. Notes of Guidance for the Testing of Cosmetic Ingredients and their Safety Evaluation — Margin of Safety methodology.
- Cosmetic Ingredient Review Expert Panel. Safety assessments of cosmetic ingredients. cir-safety.org.
- European Commission. CosIng — Cosmetic Ingredient Database, Regulation (EC) No 1223/2009.
- Central Drugs Standard Control Organisation. Cosmetics Rules, 2020 (India).
Primary source under discussion
- Environmental Working Group. Understanding Skin Deep® ratings; Skin Deep® FAQ and methodology. ewg.org/skindeep.
Survey and commentary — not peer-reviewed
- Society of Toxicology member survey, STATS / George Mason University, 2009. Sixteen years old; see the caution in section 04.
- Romanowski P. Chemists Corner — critiques of Skin Deep scoring methodology. (industry-side commentary)
- Novakovich J. The Eco Well — hazard-versus-risk critique. (independent science communication)
- Wong M. Lab Muffin Beauty Science — hazard versus risk and the Skin Deep database. (independent science communication)
Note to editor — a strong article undermined by its own badge. The credential row read "12 Peer-Reviewed References." Working through the list, roughly two are peer-reviewed papers. The remainder are EWG's own FAQ, the CosIng database, an SCCS guidance document, a market-research survey, a CIR panel page, a video lecture transcript and three blog posts. Those are legitimate sources for this argument — Lab Muffin and The Eco Well are genuinely good on this topic — but they are not peer-reviewed references. An article whose thesis is "check what the number actually measures" cannot carry a badge that misstates its own evidence base, and a critic would find it in about ninety seconds. The badge is removed and sources are now grouped honestly by type, which is more useful to a reader anyway. ⚠️ This is the third consecutive article in this batch with an inflated reference count — retinol claimed "12+", glycolic carried "10+" in one place and "30+" in another, and this claimed "12". That is a template problem rather than three separate slips, and it's worth a sitewide check of every credential badge. The 2009 toxicology survey was leaned on twice as the main "experts disagree" evidence; it is sixteen years old, was run by an organisation whose funding drew questions, and the quoted sample size refers to the whole survey rather than the subset who rated EWG. It is now stated with those limits — which strengthens the piece, since the hazard-versus-risk critique rests on logic rather than on that survey. The main addition is a steelman section giving EWG's best case: regulatory assessment leans on industry-supplied data, some ingredients genuinely are under-studied, and a data gap is not proof of safety. Including it makes the article more credible, not less. Structural fixes: three stacked answer boxes at the top consolidated to one — including one titled "Google Featured Snippet Answer", which showed readers the SEO scaffolding; a duplicate H1 removed, since Shopify renders the post title already; and viewbox= corrected to viewBox in the SVG, which is invalid lowercase since XML attributes are case-sensitive. 🔴 Slug conflicts found: the inline links and the further-reading list use different URLs for the same three articles — chemical-free-skincare versus chemical-free-skincare-scientific-impossibility, are-preservatives-in-skincare-bad versus preservatives-in-skincare, and silicones-dimethicone-for-skin-safety (in ingredient-directory) versus silicones-skincare-evidence (in skin-science-journal). Internal links have been removed pending canonical confirmation; please restore them once the paths are settled.
