Clean within standard
Every ingredient sits inside published safe-use guidelines. Nothing flagged at use concentration. The benchmark, not the exception.
The Glow Standard · Formulation
Forty ingredients Hannah Pham, Editor-in-Chief check before a score goes live. What each one does, where it sits in a formula, and whether it earns the line on the back of the bottle.
The verdict.
The Glow Formulation Index grades 40 cosmetic ingredients on a four-tier scale (A: trusted, B: acceptable, C: caution, D: avoid) across 7 categories — preservatives, SPF filters, actives, hydrators and barrier ingredients, emollients and occlusives, surfactants and solvents, fragrance and sensory, and colour and powder. It exists alongside the Glow Score as a second editorial signal — about what's in the bottle, not just how the bottle performs. Reviewed by Hannah Pham, Editor-in-Chief.
Why this exists
Most publications review what a product does. Most apps — Yuka, EWG, INCI Decoder — review what's in it. Both are useful. Neither is sufficient on its own. A sunscreen that scores 9.5 on the Glow Standard with phenoxyethanol at 0.4% is not a 4/10 product because phenoxyethanol exists in it. A clean-rated lipstick that doesn't survive a coffee is not a good lipstick because the ingredient list is short.
The Formulation Index is the second signal. Every reviewed product on Glow carries the editorial verdict and a formulation grade beside it. The grade is concentration-aware and product-category-aware. It does not override the score. It sits next to it, so the reader can decide which question matters most to them.
What follows is the working index — the four-grade rubric, then forty ingredients across the categories that come up most often. Updated quarterly. The full machine-readable file lives at /data/glow-formulation-index.json, CC-BY-4.0.
The rubric
Every reviewed product carries a single Index grade pegged to its worst-flagged ingredient at the concentration it appears in the formula — or the typical concentration for that ingredient in that category where the brand hasn't disclosed.
Clean within standard
Every ingredient sits inside published safe-use guidelines. Nothing flagged at use concentration. The benchmark, not the exception.
Mostly clean
One or two ingredients flagged. All sit inside published limits for this product category. Suits most users; sensitive skin reads on.
Mixed
At least one ingredient at, or approaching, the editorial caution threshold. Often a question of whether the trade-off makes sense for you.
Caution
At least one ingredient with established concern at any concentration in this product category. The grade reads on the score; the trade-off is documented.
The grade is informational, not the verdict. A product can carry a C and still earn a 9.0 editorial score if the formulation works for the user it was designed for. The grade is the transparency layer. The reader decides whether they care.
The index · v1.0
The ones the editors check before a score posts. Grouped by what they're doing in the formula. Each line: what it is, the editorial read, and where the grade lands at the concentration this ingredient usually sits at.
Preservatives
A Safe within EU/TGA limit
The workhorse preservative replacing parabens in most modern formulas.
Capped at 1% in the EU and Australia. Stable, broad-spectrum, with a long human-safety record at use concentration. Yuka flags it red on principle. We don't — not at 0.4% in a serum that needs water-system protection. The alternative, in most cases, is a contaminated product.
B Safe within EU/TGA limit
The short-chain parabens still cleared for cosmetic use under 0.4% individually.
The 2004 breast-tissue paper that triggered the paraben panic never demonstrated causation, and the methodology has not held up. Methyl- and ethyl-paraben remain among the better-studied preservatives in cosmetics. Brands have largely moved on; the formulas that kept them aren't worse for it.
D Banned EU since 2014
The longer-chain parabens with endocrine-disruption signal in animal data.
Banned in cosmetics across the EU since 2014. Should not be in a 2026 formula. If you see one of these in an INCI list, it's either an old SKU on shelf or a brand running a poor regulatory line.
D Editorial avoid
Quaternium-15, DMDM hydantoin, imidazolidinyl urea — preservatives that release formaldehyde over time.
IARC Group 1 carcinogen on the release end. Sensitiser at low concentrations. There are better preservative systems in 2026 and almost every reputable brand has moved off them. We flag any product carrying one.
D Banned EU 2017
Antibacterial agent in older soaps, deodorants, toothpastes.
Endocrine and antimicrobial-resistance signals. Banned in EU cosmetics since 2017. Effectively gone from Australian formulations; flagged on principle if it appears.
B Safe within limit
Synthetic antioxidants used to stabilise oils. Not the same BHA as salicylic acid.
In-vitro hazard signals, no settled clinical concern at cosmetic concentration. Mostly tracked as an environmental-persistence flag. Common in lip balms and oil-based products; we note it, we don't downgrade for it.
SPF filters
A Editorial preferred
Mineral broad-spectrum filter. The only single ingredient that covers the full UVA-UVB range.
No systemic absorption concern. Photo-stable. The reason most Australian dermatologists default to it. The trade-off is cosmetic: it tints, and at the concentrations that reach SPF 50 it can ghost. Worth it.
A Editorial preferred
The other mineral filter. UVB-strong, UVA-weaker than zinc.
Long-cleared in topical use. The nano-grade inhalation flag applies to powders, not creams. Often paired with zinc to lift the UVB end of a mineral formula.
A Editorial preferred
The new-generation organic UVA filters. Approved in Australia and the EU; not in the US.
Cover the deep-UVA range (380nm+) where older chemical filters fall away. Photo-stable, no endocrine signal. The reason the best UVA protection on the market right now is in European and Australian sunscreens, not American ones.
B Safe with stabiliser
The standard organic UVA filter in older formulas.
Effective but photo-degrades quickly — needs octocrylene or a Tinosorb pairing to hold up across a day. A solo-avobenzone formula isn't doing the job after lunch. We check the partner filters.
B Safe within limit
Photo-stabiliser usually used to prop up avobenzone.
A 2021 study flagged benzophenone formation in degraded octocrylene; the practical risk at use concentration is small but real. Most reformulations are moving toward Tinosorb instead. Acceptable inside the limit; we note it on the score sheet.
D Endocrine concern
The chemical filter banned in Hawaii and Palau for reef toxicity.
Endocrine signal stronger than most other chemical filters. Reef toxicity established. Largely reformulated out of Australian and European sunscreens; if you see it on a 2026 SKU, the formulation team didn't keep up.
Actives
A Pregnancy contraindication
Vitamin A derivatives. The most clinically supported anti-ageing active in cosmetic skincare.
Forty years of human data. Real photo-ageing, fine-line and pigmentation results from 0.3% retinol up. The watch is on pregnancy — oral isotretinoin is teratogenic and the precaution extends to topicals. Otherwise the standard. Pair with SPF.
A Editorial preferred
Vitamin B3. The barrier and pigmentation active with the cleanest tolerance profile in skincare.
Works at 2-5%. The 10% concentration that became popular around 2020 doesn't out-perform 5% and tips into irritation for sensitive skin. The one active almost every skin type can wear daily.
A Pregnancy precaution at high %
Oil-soluble exfoliant. The only acne ingredient with twenty years of clinical settlement.
Yuka flags it red. It is the safest, best-evidenced congestion treatment on the market. Pregnancy guidance is for high-concentration peels, not 2% leave-ons. The cleanser-format under 2% is fine.
B Safe with SPF
The most penetrating AHA. The reference acid for tone and texture work.
Increases UV sensitivity for up to a week post-use. The SPF habit is non-negotiable. Effective from 5% up; the at-home limit in Australia is 10%, and you don't need higher.
A Editorial preferred
The reference antioxidant. The morning step the dermatology community defends hardest.
Photo-protective, pigment-clearing, collagen-supportive. Stable as L-ascorbic acid only at low pH in an opaque, airless format — most "vitamin C" serums oxidise on shelf. Derivatives (MAP, SAP, ascorbyl glucoside) are gentler but slower. Either has a place.
B Drying
The reference active for inflammatory acne. The cheap one that works.
Bleaches towels and pillowcases. Dries skin in week one. Resistance-free unlike topical antibiotics. The 2.5% strength holds up against 10% on tolerance without losing efficacy.
D Prescription only AU
The strongest topical pigment-lightener cleared for medical use.
Schedule 4 in Australia — prescription only. Effective for melasma; ochronosis risk on prolonged unsupervised use is real. Anything sold over-the-counter as "hydroquinone-free brightening" is, by definition, not as strong. That's appropriate.
D Endocrine concern
An older pigmentation and acne active, also used in some hair dyes.
Thyroid signal in animal data. Largely reformulated out of cosmetic skincare. Flag if it appears at any concentration outside of a regulated hair-dye context.
Hydrators & barrier ingredients
A Safe
Humectant. The headline hydration ingredient, low and high molecular weight.
Pulls water from the environment in humid weather; pulls it from the skin in dry weather, which is why a serum without an occlusive on top can leave skin tighter than it started. The headline is real; the application context matters.
A Editorial preferred
The lipids that make up the skin's outer barrier. The most quietly effective category in modern skincare.
The CeraVe formula that became a derm default rides on a ceramide-1, -3, -6-II ratio. Ageing, low-humidity climates and over-actived routines deplete them; topical ceramides restore them. No tolerance ceiling.
A Safe
Hydrogenated squalene, the lipid the skin already makes. The cleanest occlusive in cosmetic use.
Non-comedogenic on the consensus data. The Beauty Pie / Biossance / The Ordinary version of an oil step. A genuine alternative to silicone for users who want a finishing layer that breathes.
A Safe
The standard antioxidant pairing for vitamin C and most oil-phase formulas.
Stabilises C, supports the barrier, no concern flag at cosmetic concentration. The unremarkable ingredient in almost every good formula.
Emollients & occlusives
A Performance neutral
The slip ingredients. The reason a primer or serum spreads the way it does.
No clinical case for the "silicones suffocate skin" line. They don't penetrate the barrier; they sit on top, fill, and wash off. Cyclopentasiloxane is being phased out for environmental persistence in some markets; dimethicone isn't.
B Performance neutral (cosmetic-grade)
Petrolatum-derived occlusive. The most studied moisturiser ingredient in cosmetic history.
Cosmetic-grade mineral oil is highly purified and the comedogenicity charge is overstated — it scores low on the modern consensus. The reason it gets a B and not an A is sustainability optics, not skin performance.
C Flag for acne-prone
The wellness-industrial oil. Beloved in body care, complicated on the face.
Comedogenic rating of 4 on the standard scale. Fine in body care for most users; we flag it on facial formulations for any skin that tends to break out. The TikTok "I rub it on my face" generation usually finds out.
B Flag for sensitive
Wool-derived emollient. The original lip-balm and rash-cream lipid.
Genuinely effective for chapped lips and barrier rescue. Common allergen — not for sensitised skin or known wool reactions. We flag it but don't downgrade a product that uses it well.
Surfactants & solvents
C Flag for sensitive skin
The classic foaming surfactants in cleansers, shampoos, body washes.
SLS is the harsher of the two; SLES is gentler but a manufacturing-process concern (1,4-dioxane contamination) means we prefer modern alternatives. On a daily face cleanser we flag both. On a shampoo for unsensitised scalps, they're fine.
B Safe
Emulsifier-solubilisers across the formulation industry. The number after PEG is molecular weight.
Cosmetic-grade PEGs are well-purified and the 1,4-dioxane line that floats around clean-beauty marketing is a manufacturing-control issue, not an ingredient issue. We note them, we don't downgrade.
C Flag for dry/mature skin
The drying alcohols — ethanol, alcohol denat, SD alcohol — used as fast-evaporating carriers.
Useful in oily-skin toners and acne treatments where you want fast evaporation. A problem high in the INCI list of a hydrator or moisturiser for any skin that runs dry. The "fatty alcohols" (cetyl, stearyl) are the opposite category and don't apply.
B Flag if alcohol-extracted
Botanical astringent. The wellness-counter staple in toners and "calming" products.
The plant extract itself is mild. The standard manufacturing process uses high-strength alcohol that stays in the bottle. Reads on the alcohol flag, not the witch-hazel flag.
Fragrance & sensory
C Flag for sensitive skin
The catch-all ingredient name covering any combination of up to thousands of disclosed and undisclosed compounds.
The single biggest source of contact dermatitis in cosmetics. EU disclosure rules now force 26 named allergens onto labels but the rest of the formula stays opaque. We flag it on every leave-on product for any user prone to reaction.
C Flag for sensitive skin
The 26 EU-listed compounds that must appear by name when present above 0.001% in leave-on products.
Oxidise over time into more sensitising forms. The reason an old serum can suddenly start reacting on skin that tolerated it new. We flag.
C Flag for sensitive skin
Plant-distilled fragrance compounds in "natural" formulations.
The "natural therefore safe" argument doesn't survive a patch-test panel. Tea tree, lavender, peppermint and citrus oils are common reactors. We treat them on the same flag as synthetic fragrance, sometimes more strict.
D Undisclosed in fragrance
Plasticiser compounds historically used as fragrance carriers and nail-polish softeners.
DEHP and DBP are restricted in EU cosmetics. DEP remains in widespread fragrance use and rarely appears on labels because "fragrance" covers it. The hardest ingredient to track honestly, and a real concern at the systemic-exposure end.
Colour, powder & body
B Safe within approval
The colour-index codes for FDA / EU-approved cosmetic dyes.
Most colour cosmetics in the world use them. Approval lists are different by market — an EU-approved formula will sometimes use a lake the US prohibits. Within market approval, no concern at use concentration.
B Safe with asbestos-free certification
The mineral powder base in setting powders, blushes, body powders.
The historic concern is asbestos contamination at the mine. Reputable brands now certify asbestos-free and we check. The flag is on traceability, not talc itself.
C Performance vs concern trade-off
The aluminium chlorohydrate / zirconium compounds that make antiperspirants actually antiperspirant.
The breast-cancer link has not held up under repeated review. The trade-off is real: aluminium-free "natural" deodorants don't stop sweat, they mask odour. If wetness control is the job, the alternative is worse skincare in another way.
The firewall
The grade does not override the verdict. The editorial score is the primary review output. The grade sits beside it as informational transparency. If a Grade C formulation works for the user it was designed for and tests well in the field, it can still score 9.0.
No brand can pay to influence its own grade. The same firewall applies as the editorial review. Brand outreach about a grade is logged and ignored. PR retainers don't move letters.
The watch list is versioned in public. Every change to an ingredient classification is documented in the next Glow Edit, with the editorial reasoning written out. Index v1.0 covers the forty ingredients above; v2.0 (planned October 2026) adds another sixty across colour cosmetics, hair care, body care and ingestibles.
Quarterly review by the senior editor. Hannah Pham, with no commercial ties to any reviewed brand, reads every flagged ingredient against the latest published research each quarter. Disagreements are documented in the changelog, not buried.
Open licence. The watch-list JSON is CC-BY-4.0. Retailers, other publications and brands can reuse it with credit. The data is the moat; sharing it expands the moat.
Appeals are editorial, not commercial. Brands that believe a product has been mis-graded can write to [email protected]. Factual corrections (newly disclosed concentration data, for example) are reviewed within ten business days. Methodological objections are added to the next quarterly review with the brand named in the changelog. No financial relationship is ever offered, accepted, or implied as part of an appeal.
Field note
The clean-beauty marketing layer in Australia has spent a decade telling shoppers the back of the bottle is a hazard list. It isn't. Most ingredients on most formulations are doing exactly what they look like they're doing: holding the water phase together, keeping it from spoiling, helping the actives sit where they need to sit. The actual concerns are narrower and more boring than the apps suggest.
What's worth your time, on a 2026 Priceline aisle: the preservative system (formaldehyde-releasers out; phenoxyethanol fine); the SPF filters (mineral and the new-gen organics good; oxybenzone gone); the fragrance line if your skin reacts; and the alcohol level if you run dry. The rest is mostly noise around well-studied ingredients that don't deserve red squares from an app that hasn't read the EU SCCS file.
And the second signal, the one that didn't fit on the Yuka scan: whether the formula actually does the job. A 10/10 ingredient list on a serum that doesn't reach the dermis is still a bad serum. The Index sits beside the score so you can read both, not so you can replace one with the other.
More: how we review · the v1.0 JSON · editorial standards · disclosures & corrections.