SPF, broad spectrum, and why New Zealand made sunscreen testing mandatory

Published 6 August 2026Zane Hitchcox, editor

Almost every argument about sunscreen labelling collapses into two questions that are usually confused with each other. Is the number measuring the right thing? And is the number true? New Zealand spent a decade arguing about the second while barely mentioning the first.

What the number on the front actually measures

Sun protection factor is a ratio, not a percentage and not a duration. DermNet, New Zealand's dermatology reference, defines it as the ratio of the ultraviolet dose needed to produce a barely detectable sunburn on skin treated with the product to the dose needed on untreated skin.¹ A product that scores 30 means that, in the test, protected skin took thirty times the dose to redden.

The test is done on people. Under the Australian and New Zealand standard, small areas of a volunteer's back are irradiated with a solar simulator, unprotected and protected, and the ratio of the two doses is calculated; the procedure is run on ten individual volunteers and the individual results are averaged.¹³ The product is spread at 2 milligrams per square centimetre. That density is not a guideline for consumers. It is the fixed condition under which the number was generated, and it is the single most important thing to know about SPF.

Redness is also a specific endpoint. Erythema is driven mainly by UVB, so SPF is fundamentally a UVB measure. It says nothing on its own about UVA, which penetrates deeper and is implicated in photoageing and in melanoma risk. That is what the separate "broad spectrum" test exists to cover, and why a high SPF with no broad-spectrum claim is a half-finished product.

The 2 mg per square centimetre problem

Nobody applies 2 mg/cm². A literature review by Petersen and Wulf in Photodermatology, Photoimmunology & Photomedicine in 2014 found real-world application amounts of roughly 0.39 to 1.0 mg/cm², and concluded that current sunscreen labelling "overrates the protective effect of a given sunscreen when the reality of sunscreen use is taken into account".³ For an average adult body, 2 mg/cm² works out at about 35 mL of lotion, roughly seven teaspoons, of which about one teaspoon is for the face, ears and neck.¹

Applying less does not reduce protection proportionally. Faurschou and Wulf tested an SPF 4 sunscreen at four densities on the backs of twenty volunteers and found the relationship between quantity and delivered SPF was best described by exponential growth, so that halving the dose approximates taking the square root of the SPF and quartering it approximates the fourth root.² That study is small, used one low-SPF product, and its extrapolation to modern SPF 50+ formulations has never been confirmed in vivo, so it should be read as the direction of the effect rather than a conversion table.

Why the SPF curve flattens, and why that does not make the difference meaningless

The share of UVB a sunscreen stops is one minus the reciprocal of the SPF. SPF 15 blocks about 93 per cent, SPF 30 about 96.7 per cent, and SPF 50 about 98 per cent; DermNet gives the rounded figures as roughly 94 per cent for SPF 15, 97 for SPF 30 and 98 for SPF 50+.¹ Stated that way the step from 30 to 50 looks trivial, which is the standard argument for ignoring high numbers.

The argument is arithmetically true and practically misleading, because what reaches the skin is the transmitted fraction, not the blocked one. SPF 30 lets through 1/30 of the dose, or 3.3 per cent; SPF 50 lets through 1/50, or 2.0 per cent. The higher product transmits about forty per cent less. Whether that difference is detectable on a real person is an empirical question, and it has been tested. A randomised, double-blind, split-face trial of 199 adults skiing in Vail, Colorado applied SPF 100+ to one side of the face and SPF 50+ to the other; after an average of six hours of exposure, blinded assessors scored 55.3 per cent of participants as more sunburned on the SPF 50+ side and 5 per cent as more sunburned on the SPF 100+ side.⁴ The trial's own stated limitation is that a single day's exposure does not establish longer-term benefit, and several of its authors were employed by a sunscreen manufacturer.

The disease endpoint has been tested only once at randomised-trial standard. The Nambour Skin Cancer Prevention Trial in Queensland assigned 1,621 adults to daily or discretionary sunscreen use and followed them for ten years after the intervention stopped. Eleven new primary melanomas occurred in the daily group against twenty-two in the discretionary group; the overall reduction did not reach conventional significance (hazard ratio 0.50, 95% CI 0.24 to 1.02, p = 0.051), while the reduction in invasive melanoma did (three cases against eleven, hazard ratio 0.27, 95% CI 0.08 to 0.97).⁵ Those are very small numbers of events, and the authors' own conclusion is appropriately cautious: melanoma "may be preventable" by regular sunscreen use in adults.

"Broad spectrum" means three different things

Every major market has a UVA test. They do not agree, and a bottle carrying the phrase in Auckland has passed a different examination from one carrying it in Los Angeles.

Under the Australian and New Zealand standard a product must satisfy two conditions: an in vitro UVA protection factor of at least one third of the labelled SPF, and a critical wavelength of at least 370 nm.¹ The critical wavelength has a precise definition, given in the American rule as the wavelength at which the integral of the spectral absorbance curve reaches ninety per cent of the integral across 290–400 nm, so it describes the shape of the protection curve rather than its height.⁹ The European Commission asks for the same pair of conditions — a UVA protection factor of one third of the sun protection factor, and a critical wavelength of 370 nm — but it asks in a Recommendation, 2006/647/EC, which is not binding law on manufacturers the way the New Zealand standard now is.⁸ The United States asks for one of the two: under 21 CFR 201.327, a mean critical wavelength of 370 nm or greater is classified as broad spectrum, with no requirement that UVA protection scale with SPF.⁹

Market	UVA-PF ≥ 1/3 of SPF	Critical wavelength ≥ 370 nm
Australia / New Zealand	Required	Required
European Union (recommendation)	Required	Required
United States	Not required	Required

The practical consequence is that an American "broad spectrum SPF 50" can pair very high UVB protection with UVA protection that does not rise to meet it, and still be labelled correctly. Laboratory work published in Photodermatology, Photoimmunology & Photomedicine in 2022, testing 51 sunscreens bought on the United States market, found average unweighted UVA protection at just 24 per cent of the labelled SPF.⁷ That study measured absorbance on plates rather than sunburn on people, which is a real limitation, but the gap it describes is the gap the American rule permits.

What "SPF 50+" is allowed to mean, and what "4 hours" is allowed to mean

Two label phrases carry more information than they appear to. Under the Australian and New Zealand standard, SPF claims are restricted to a fixed set of values and each is a band, not a measurement: a mean test result of 30 to 39 is labelled SPF 30, 40 to 49 is labelled SPF 40, 50 to 60 is labelled SPF 50, and only a mean result of 60 or higher may be labelled SPF 50+.¹³ "50+" is therefore a floor of 60, not a ceiling of 50, and two products both labelled SPF 50+ may have tested at 61 and at 95.

Water resistance is measured, not asserted. The volunteers' treated skin is immersed in water for a set period, dried in air, and the SPF is then determined as in the static test; the rating that may go on the label is the total time the test areas actually spent immersed, and under the joint standard products at SPF 30 and above may claim up to four hours.¹³ DermNet records the current requirement as a duration-qualified claim up to a maximum of four hours, with an SPF of at least 8 retained after immersion.¹ The United States permits only "Water Resistant (40 minutes)" or "Water Resistant (80 minutes)", and expressly makes "sunblock", "waterproof" and "sweatproof" misbranding terms.⁹ The joint Australasian standard rules the same three words unacceptable.¹³ They are among the clearest cases of a phrase being banned rather than merely discouraged, which is unusual on a cosmetic pack; most of the words on the front of a bottle have no legal definition at all.

New Zealand made it compulsory on 8 September 2022

Until 2022 the joint standard was voluntary in New Zealand. Sunscreen is a cosmetic there rather than a medicine, which is the underlying reason the obligation had to be created by a separate statute rather than inherited from a therapeutic-goods regime as it was across the Tasman; who regulates cosmetics in New Zealand is a genuinely unusual answer. The Sunscreen (Product Safety Standard) Act 2022 fixed that by making the Australian and New Zealand sunscreen standard a mandatory product safety standard under the Fair Trading Act, enforced by the Commerce Commission.¹⁰¹¹ The obligations took effect on 8 September 2022, with a twelve-month transition for stock manufactured or imported before 8 March 2022, and penalties of up to NZ$600,000 for a company and NZ$200,000 for an individual.¹²¹¹ Consumer NZ, reporting the Act becoming law, gives the same NZ$600,000 maximum for companies.¹⁸

The Act covers products whose primary purpose is protection from ultraviolet radiation, insect repellents labelled SPF 4 or higher, and skincare products such as moisturisers with an SPF above 15, excluding products intended for the lips and products sold primarily as a tinted base or foundation.¹⁰ The Commerce Commission's guidance to manufacturers and importers is that a sunscreen "should be independently tested to ensure they comply with the Safety Standard before being supplied in New Zealand" — the legal duty is compliance with the standard, and testing is how a supplier discharges it.¹⁰

Two wrinkles are worth stating plainly, because the official record is not tidy. The Commerce Commission names the instrument as the Sunscreen (Product Safety Standard) Act 2022 and the standard as AS/NZ 2604:2012; Product Safety New Zealand, the Ministry of Business, Innovation and Employment's own consumer-products site, names the same instrument as the "Sunscreen (Product Safety Standard) Act 2021", requires products to meet "AS/NZ 2604:2012", and then describes the transition in terms of the "AS/NZ 2604:2021 standard" in the next sentence.¹⁰¹¹ A 2021 edition of the standard does exist and is the edition Australian testing now works to.¹⁴ A supplier reading only the government pages cannot tell from them which edition binds it.

The testing that caused it

The Act did not arrive from nowhere. Consumer NZ ran a sunscreen testing programme through the period when compliance was voluntary and kept finding shortfalls. In testing reported in November 2019, twenty sunscreens were assessed and nine failed to meet their SPF label claim, with three more failing broad-spectrum requirements; Consumer NZ's then chief executive Sue Chetwin said at the time that "New Zealand has one of the highest rates of skin cancer and melanoma in the world but the sunscreen standard remains voluntary".²¹ In 2020 the organisation commissioned blind testing of twelve sunscreens at an accredited laboratory in Germany; half of the twelve failed to meet their SPF claims, and in April 2021 Consumer NZ asked the Commerce Commission to investigate seven of the products. Measured results included one labelled SPF 50+ returning 30.4 and two labelled SPF 30 returning 22.5 and 20.2.¹⁹ A further round reported in December 2021 tested nine products and found three short — every one of them a repeat failure from the previous year, with two of the three also missing the broad-spectrum requirement.²⁰

The enforcement outcome is the part usually left out. In late 2022 the Australian manufacturer Ego Pharmaceuticals was fined NZ$280,000 over unsubstantiated SPF claims for two Sunsense products labelled SPF 50+, following a Commerce Commission investigation prompted by a Consumer NZ complaint; Consumer NZ's own 2018 and 2019 testing had found both products delivering moderate protection rather than the very high protection the labels claimed.¹⁸²⁴ The company had relied on certificates from AMA Laboratories, a United States testing house whose owner admitted fabricating test reports over three decades and was sentenced in May 2022 to five years' imprisonment.¹⁸ That case is the strongest available argument for independent testing being a legal duty rather than a commercial choice, and it is also a reminder that the certificate is only as good as the laboratory behind it.

Across the Tasman, CHOICE tested twenty sunscreens to AS/NZS 2604:2021, which works through the international methods ISO 24444 and ISO 24443, at an external laboratory approved by Australia's Therapeutic Goods Administration, using a panel of ten adult volunteers; it reported in June 2025 that sixteen did not meet their SPF claims.¹⁴¹⁵ Eleven of the products are sold in New Zealand.¹⁶ The largest single miss was a product labelled SPF 50+ returning SPF 4, which CHOICE re-tested on a different batch at a German laboratory and obtained SPF 5. The manufacturer initially suggested human error was likely, then commissioned tests from several independent laboratories and withdrew the product, stating that across eight tests it had "returned SPF data of 4, 10, 21, 26, 33, 60, 61, and 64".¹⁴ Those eight numbers are the single most useful data point in this article, and they point in two directions at once: a product that can read 4 and 64 is either wildly non-uniform or being measured by a method that can be wildly wrong.

The complication: the reference test is noisy

A mandatory standard is only as good as the measurement underneath it, and the in vivo SPF method has a reproducibility problem that is well documented and rarely reported alongside the failure headlines. A multicentre study published in the International Journal of Cosmetic Science in 2025 coded 32 products into eight groups of four and sent each group to four of twelve pre-qualified in vivo SPF testing laboratories. The expected SPF values were withheld; the laboratories were told only whether the true value was above or below 25. Its conclusion was that interlaboratory variability of SPF results is considerable and likely greater than the public and regulatory bodies expect, that variability scales with the SPF of the product, and that a laboratory's repeatability with itself is much better than its agreement with other laboratories.⁶

This is why manufacturers' rebuttals cannot simply be dismissed. When Consumer NZ reported failures in December 2021, one brand produced a 2021 retest from a United Kingdom laboratory supporting its SPF 30 claim and supplied the report; another said its supplier had retested but did not supply the report; a third said its product had been tested in Australia and declined to provide the report.²⁰ Two accredited laboratories disagreeing about the same product is an expected outcome of a method with this much spread. A single laboratory returning SPF 4 against a claim of 50+, confirmed at 5 by a second laboratory on a different batch, is not. Distinguishing the two cases requires knowing the size of the ordinary disagreement, which is precisely what the ring studies were done to establish.

There is a third possibility, and by 2025 it had a name attached to it. In July 2025 Consumer NZ reported an ABC News investigation into the sunscreen testing laboratory Princeton Consumer Research, which found that at least half the sunscreens failing the CHOICE test had also been tested by that laboratory, and that its results had come back significantly higher; industry experts told the ABC they were troubled by how little variation there was between the laboratory's individual subject results, which is odd in a test run on human skin. The laboratory's technical director conceded the uniformity was "uncommon" but said it was "not inherently implausible".¹⁷ The retest that supported one brand's SPF 30 claim in Consumer NZ's December 2021 round was carried out, in 1News's words, at "UK lab PCR Corp".²⁰ Method spread, formulation failure and a laboratory producing improbably tidy numbers all look identical from outside, and a certificate does not distinguish them. Consumer NZ's standing request since the Act took effect is not that the standard is wrong but that the Commission run a regular monitoring programme, because as the law stands compliance is tested before supply and not again.¹⁷¹⁰

Why New Zealand, specifically

The claim that New Zealand has the world's highest melanoma rate is repeated constantly, including by people arguing for this legislation, and in the most recent global estimates it is not quite right. In GLOBOCAN 2024, New Zealand's age-standardised incidence of melanoma of the skin is 29.2 per 100,000 (34.6 in males, 24.3 in females), with 2,847 new cases estimated for the year, making melanoma the country's fifth most common cancer.²² The world rate is 3.1 per 100,000, and the Australia–New Zealand region leads every other region on the planet; the highest individual country for incidence is Australia, at 38.5 per 100,000.²³ So New Zealand's incidence sits at roughly nine times the global rate but behind its neighbour's. The claim survives in a different form, and a worse one: on the same estimates New Zealand has the highest melanoma mortality rate of any country in the world, at 3.2 deaths per 100,000.²³ Australia has the higher rate of getting melanoma; New Zealand has the higher rate of dying of it. Either version is an overwhelming argument for the legislation.

What follows from all of this for a reader holding a bottle is narrower than the marketing suggests, and it is a point about physics and measurement rather than about which product to buy. The number was produced under a thicker film than the one going on, the phrase "broad spectrum" is jurisdictional, "50+" is a floor of 60, and "4 hours" is an immersion time. None of that makes the number worthless. It makes it a measurement with stated conditions, which is more than most of what is printed on cosmetic packaging can claim, and considerably more than most topical products can demonstrate at all. Sunscreen is also the rare case where the ingredient doing the work is legally required to be identifiable in the list on the back, which is worth reading; the ordering rules apply here as they do everywhere else.