How to read an INCI list

Published 6 August 2026Zane Hitchcox, editor

The names are not there to alarm anyone. They exist so that a dermatologist in Hamilton and a formulator in Hamburg mean the same substance by the same word. Who solved that problem, and how the law then borrowed the solution, explains almost every strange feature of the panel on the back of the bottle.

The names are written by a trade association, not a regulator

INCI stands for International Nomenclature of Cosmetic Ingredients. The names are developed by the International Nomenclature Committee and published by the Personal Care Products Council, a United States industry association, in the International Cosmetic Ingredient Dictionary and Handbook. The Council states plainly that an INCI name carries no regulatory weight of its own: the assignment of one "does not imply that the ingredient is safe, or that its use in a cosmetic product complies with the laws and regulations of the United States or other global regions."¹⁴ It is a filing system, and it should be read as one.

Regulators then adopt the filing system by reference, and they do it differently. In the European Union, Article 19(1)(g) of Regulation (EC) No 1223/2009 requires a list of ingredients preceded by the word "ingredients", and Article 19(6) requires each one to be expressed using the common ingredient name set out in the glossary compiled under Article 33.¹ That glossary is a live instrument. The version most commonly cited in trade writing, Commission Decision (EU) 2019/701,⁴ was repealed by Implementing Decision (EU) 2022/677,³ which was in turn repealed by Implementing Decision (EU) 2025/1175 of 16 June 2025. The 2025 glossary runs to 30,418 numbered entries and became mandatory on 30 July 2026.² Anyone citing the 2019 decision today is citing a text that has been superseded twice.

The United States took a different route and has stayed on it. Under 21 CFR 701.3(c), an ingredient must be identified by a name the FDA Commissioner has established in § 701.30, and in the absence of one, by the name adopted in a fixed list of compendia headed by the CTFA Cosmetic Ingredient Dictionary, Second Edition, 1977.⁷ FDA's own labeling guide confirms that the 1977 second edition remains the currently recognised edition and is "recognized only in part, i.e., not all names listed in the second edition have been adopted", that the third edition of 1982 and its 1985 supplement "have not yet been recognized", and that the agency has told the association it will not act against products labelled in accordance with those newer editions while review continues.⁸ The practical effect is that United States ingredient naming runs on enforcement discretion layered over a book published nearly fifty years ago.

That divergence is visible on the packaging. FDA does not accept "Aqua" or "Parfum" on their own, because the Fair Packaging and Labeling Act requires common or usual names; the agency does not object to the international term in parentheses afterwards, and offers "Water (Aqua)", "Fragrance (Parfum)", "Honey (Mel)" and "Sweet Almond (Prunus Amygdalus Dulcis) Oil" as acceptable forms.⁹ So "Aqua" is water. The Latin is a nomenclature convention rather than a grade of water, and in one large market it is not sufficient on its own.

Descending order, and the exact line where it stops

The ordering rule is the single most useful thing on the panel, and it is narrower than most readers assume. Regulation (EC) No 1223/2009 sets it out in two sentences: the list "shall be established in descending order of weight of the ingredients at the time they are added to the cosmetic product", and "ingredients in concentrations of less than 1 % may be listed in any order after those in concentrations of more than 1 %."¹

The United States reaches the same place by a different construction. 21 CFR 701.3(a) requires descending order of predominance, and 701.3(f) offers an alternative: ingredients other than colour additives above one percent in descending order, then those at not more than one percent in any order, then colour additives in any order. The same paragraph allows the sub-one-percent group to be folded back into the descending list if a manufacturer prefers.⁷

New Zealand's rule is the crispest of the three and the least known. The Cosmetic Products Group Standard 2020, Schedule 1, condition 1(2), requires the ingredients other than colour additives at "1% or more in descending order by volume or mass", then those at "less than 1% in any order", then colour additives in any order.¹² Two details are worth noticing. New Zealand permits ordering by volume as an alternative to mass, which the European text does not; for a formula mixing dense aqueous phases with light oils, that can change the order. And the New Zealand and United States texts both close the boundary at exactly one percent, while the European wording pairs "less than 1 %" with "more than 1 %", leaving an ingredient sitting at precisely one percent unaddressed by either clause.

None of this is a loophole; it is the shape of the instrument. The list answers one question well and another not at all. It tells you, reliably, which ingredients are present at one percent or more and how they rank among themselves. It tells you nothing about the size of any gap. A product whose second ingredient is at 8 percent and one whose second ingredient is at 1.2 percent produce identical labels.

A preservative with a legal maximum is a ruler

Because the sub-one-percent group is contiguous and sits after the larger ingredients, any ingredient you know to be capped below one percent marks the boundary. Phenoxyethanol is the most common such marker. Its maximum authorised concentration as a preservative is 1.0 percent, at entry 29 of Schedule 7 of the New Zealand Group Standard and at entry 29 of Annex V to the European regulation.¹²¹ A product cannot legally contain more. So everything printed after phenoxyethanol is at one percent or less, in all three jurisdictions and under every ordering option they allow.

Two limits on the trick, stated at the same time as the trick. It is a ceiling, not a reading: an ingredient just below the marker might be at 0.9 percent or at 0.0001 percent, and the label is identical either way. And it says nothing about ingredients listed before the marker, because sub-one-percent ingredients may be placed anywhere within their own group, including ahead of the preservative. The ruler tells you where the floor begins. It does not measure anything above it.

Botanicals get a Latin binomial and a plant part

Plant-derived ingredients are named genus, species, plant part, and form. The 2025 glossary contains BUTYROSPERMUM PARKII BUTTER, HELIANTHUS ANNUUS SEED OIL, LEPTOSPERMUM SCOPARIUM BRANCH/LEAF OIL, KUNZEA ERICOIDES LEAF OIL, PHORMIUM TENAX SEED OIL and PSEUDOWINTERA COLORATA LEAF EXTRACT.² The convention is why lists read as more chemical than they are, and why they are more precise than the front of the pack. "Harakeke" on a carton could be anything from the plant. PHORMIUM TENAX SEED OIL is a pressed seed oil and PHORMIUM TENAX LEAF EXTRACT is not; they are different materials with different composition, and the research attaches to one and not the other.

What the binomial does not tell you is the solvent used, the ratio of plant to extract, whether the material was standardised to any marker compound, or the chemotype of the population it came from. For essential oils, all four of those change the chemistry substantially.

The glossary also freezes botanical names at the moment they were entered. Kawakawa appears as MACROPIPER EXCELSUM LEAF EXTRACT and MACROPIPER EXCELSUM LEAF POWDER; there is no Piper excelsum entry in the glossary,² even though World Flora Online's current classification places Macropiper excelsum (G.Forst.) Miq. under the accepted name Piper excelsum G.Forst.¹⁸ A reader searching a label under the accepted botanical name will not find the ingredient. This matters for anything sold on a single named plant, where the name is doing most of the persuading.

Colourants run on a separate system

Colourants other than hair colourants may be listed in any order after everything else, and the Colour Index nomenclature must be used where applicable. For decorative products sold in several shades, every colourant used across the range may be listed provided the words "may contain" or the symbol "+/-" are added.¹ The United States has an equivalent at 21 CFR 701.3(g), which permits an ingredient not present in the product to be declared after "may contain" where it is a colour additive added to some batches for colour matching, or where a single declaration is shared across a line of shaded products.⁷ New Zealand simply places colour additives last in any order.¹²

So a number after "may contain" is not a statement that the substance is in the tube in your hand. It is a statement about a range. CI 75810, to take the green that turns up on botanical-positioned products, is a chlorophyllin copper complex, listed at entry 116 of Schedule 6 of the New Zealand Group Standard and also permitted as the food colours E 140 and E 141.¹² FDA does not accept a bare CI number as a substitute on a United States label; the common or usual name is required, and the agency does not object to the Colour Index number in parentheses after it, as in "FD&C Yellow No. 5 (CI 19140)".⁹

Parfum, aroma, and the allergens that must be named anyway

"Parfum" is a permitted collective term for an entire composition. Article 19(1)(g) provides that perfume and aromatic compositions and their raw materials shall be referred to by the terms "parfum" or "aroma", with named substances added on top where Annex III requires it.¹ New Zealand's Schedule 1, condition 1(2)(d) and (e), allows "flavour", "aroma", "fragrance" or "parfum", or alternatively the individual ingredients.¹² 21 CFR 701.3(a) allows fragrance or flavor to be listed as such.⁷ One word can therefore stand for dozens of substances in every one of these markets.

The exception is the named-allergen list, and this is where the three jurisdictions have come apart. Recital 5 of Commission Regulation (EU) 2023/1545 records that 24 fragrance allergens, at entries 45 and 67 to 92 of Annex III, were required to be individually labelled. The Scientific Committee on Consumer Safety identified 56 further allergens that had clearly caused allergies in humans, and the regulation added them, at thresholds of 0.001 percent in leave-on products and 0.01 percent in rinse-off products. Products not complying could be placed on the Union market until 31 July 2026 and may be made available until 31 July 2028.⁵ The first of those dates fell a week before this article was published, so a European product entering the market now carries roughly eighty possible named allergens rather than twenty-four.

New Zealand has not followed. The Group Standard as consolidated on 1 January 2026 carries exactly 24 individually labelled fragrance allergens, at Schedule 5 entries 45 and 67 to 92, at the same 0.001 and 0.01 percent thresholds.¹² The United States has no such requirement at all yet. Section 609 of the Federal Food, Drug, and Cosmetic Act, added by the Modernization of Cosmetics Regulation Act and codified at 21 U.S.C. 364e, requires the responsible person to identify each fragrance allergen on the label, but leaves the identification of those substances to a regulation FDA was to propose no later than 18 months after 29 December 2022.¹⁰ A search of Federal Register documents from FDA mentioning fragrance allergens returns no such proposed rule as of 6 August 2026.¹¹

Those named substances are not a bureaucratic curiosity. In Brumley and colleagues' retrospective review in Dermatitis in 2024, of 135 children patch tested at one North American centre between 2020 and 2023, the highest-ranking allergens by significance-prevalence index were linalool hydroperoxides, propylene glycol, limonene hydroperoxides, fragrance mix I and lanolin.¹⁶ Two of the top three are oxidation products of substances the label does name, which is the useful and uncomfortable part: the label prints "Linalool", while the sensitising species is what linalool becomes on exposure to air.

The regulations do try to reach the oxidation problem, but not through the ingredient list. Schedule 5 of the New Zealand standard caps peroxide value at less than 10 mmoles per litre for terpenes and terpenoids generally, and at less than 20 mmoles per litre for the limonene isomers at entries 88, 167 and 168.¹² No label prints a peroxide value, and no shopper can measure one. That is covered further in the piece on fragrance and essential oils.

Four ways a fully compliant list still misleads

An active below the one-percent line. Under the European claims regulation, Commission Regulation (EU) No 655/2013, an ingredient claim must satisfy the criterion of truthfulness: "if it is claimed on the product that it contains a specific ingredient, the ingredient shall be deliberately present", and "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."⁶ Deliberately present is the whole test. There is no minimum. A trial-supported concentration and a token one are equally compliant, and the ingredient list cannot distinguish them, which is why an active listed after the preservative should be read as a claim about the formulator's marketing rather than about the formula.

A hero botanical at trace level. The same arithmetic, applied to the plant on the front of the box. This is the commonest reason the back of a bottle disappoints a reader who has just been persuaded by the front, and it is not detectable from the list alone. The European regulation does give the public a route: Article 21 requires the responsible person to make the qualitative and quantitative composition easily accessible. The next sentence closes most of it again, limiting the quantitative information that must be public to hazardous substances under the classification regulation.¹ The percentage of the hero botanical is not disclosable, and asking will not produce it.

Proprietary blend names. A trademarked complex cannot appear in the list in place of its components. Article 19(6) requires glossary names,¹ and FDA's guide is explicit that the compounds of a proprietary mixture "must be integrated into the product formulation and declared individually in order of decreasing predominance".⁸ The trade name therefore lives on the front of the pack while the same materials appear on the back under ordinary names, often several positions apart and often below the one-percent line. Matching the two is the reader's job and nobody's legal obligation.

What is allowed to be absent. 21 CFR 701.3(l) exempts incidental ingredients present at insignificant levels with no technical or functional effect, including substances carried in as components of other ingredients, and processing aids removed before packaging. Separately, 701.3(a) permits the phrase "and other ingredients" at the end of the declaration in place of an ingredient FDA has accepted as exempt from public disclosure as a trade secret.⁷ New Zealand adds its own wrinkle: condition 1(11) of the Group Standard waives the domestic ingredient rules if the labelling complies with the current requirements of Australia, the United States, Canada, the United Kingdom or the European Union.¹² A New Zealand shopper may therefore be reading a list built to any one of six rulebooks.

What the front of the pack is subject to is a general standard rather than a format. Article 20 of the European regulation forbids text, names, trade marks and images that imply characteristics a product does not have, and Regulation (EU) No 655/2013 supplies the six common criteria that give it content.¹⁶ In New Zealand the Environmental Protection Authority, which administers the Group Standard, points elsewhere for claims: to Medsafe and the Medicines Act 1981 for anything therapeutic, and to Commerce Commission guidance on making accurate claims.¹³ The division of labour is covered in the article on who regulates cosmetics in New Zealand. A general standard is enforced case by case, after the fact, on complaint. A format rule is enforced by the printer. This is why the back of the pack is more informative than the front even though the front is where the promises are.

What a label survey can and cannot establish

What is actually on moisturiser labels has been counted. A cross-sectional analysis in JAMA Dermatology took the hundred best-selling whole-body moisturisers at each of three United States online retailers in August 2016, 174 unique products once duplicates were removed. Only 12 percent were free of ingredients in the North American Contact Dermatitis Group series; the three commonest were fragrance mix, paraben mix and tocopherol. Of the products claiming to be fragrance free, 45 percent contained at least one fragrance cross-reactor or botanical ingredient.¹⁵ That last figure is a fact about labels, not about reactions, and it belongs with the words on cosmetic labels that have no legal definition and with what "natural" and "clean" are permitted to mean.

Case evidence goes the other way and shows why the list is worth reading anyway. Aerts and colleagues, in Contact Dermatitis in 2016, reported 13 patients from two Belgian university patch-test clinics with allergic contact dermatitis caused by ethylhexylglycerin over the period 1990 to 2015; the culprit products included a high number of facial creams, sun creams and deodorants marketed as "hypo-allergenic" and "preservative-free".¹⁷ The front of those packs was reassuring and the back of them was accurate. Only one of the two was useful.

Both papers point the same way. The ingredient list is not a safety guarantee or a quality signal. It is the only place on the package where a substance can be identified by a name a clinician can patch test against, which is narrower and more useful.