What the laboratory actually measures
The internationally recognised method for sun protection factor is an in vivo test, meaning it is performed on living human skin. A panel of volunteers, at least ten of them, has a measured quantity of sunscreen applied to marked areas on the back. The quantity is not approximate. It is two milligrams of product per square centimetre of skin, weighed and spread by a trained technician.
A solar simulator then delivers increasing doses of ultraviolet to a series of small sub patches, both on protected and on unprotected skin. Sixteen to twenty four hours later, the technician records the smallest dose that produced a just perceptible reddening. This is the minimal erythemal dose. The protected value divided by the unprotected value gives that volunteer's individual protection factor. The panel results are averaged, statistical criteria are applied, and the labelled SPF is the number that survives them.
This is a serious test. It is expensive, it is reproducible between laboratories within known tolerances, and it means the number on a UK pack has been earned rather than asserted. It also has three limitations that matter more than most people realise.
It measures redness, so it measures UVB
Skin reddening is overwhelmingly driven by UVB. SPF is therefore a UVB measurement, and it says almost nothing about UVA. A product can score a high SPF while offering poor protection against the longer wavelengths that penetrate deeper, generate free radical damage, drive photoageing and contribute to pigmentation.
This gap is why a separate UVA test exists, and why the UVA in a circle mark was introduced. That mark means the product's UVA protection factor is at least one third of its labelled SPF, measured by a separate method. A UK pack showing SPF 50 with the circled UVA mark is telling you two different things that were tested two different ways. Our explainer on UVA, UVB and PA ratings covers what the star ratings and the Japanese PA grades add to this picture.
The dose problem
Two milligrams per square centimetre is a thick layer. On an average adult body it works out at roughly thirty five millilitres, close to six or seven teaspoons, for a single full application. On a face and neck alone it is about a third of a teaspoon.
Published measurements of how much people apply in normal circumstances land consistently below this, often between a quarter and a half of the test dose. The relationship between applied thickness and delivered protection is not linear in a friendly direction. Halving the dose does not halve the SPF. It reduces it disproportionately, which is why a well applied SPF 30 outperforms a thinly smeared SPF 50 in practice.
This single fact is the most useful thing a buyer can know, and it changes what you should look for. Texture is not vanity. A formulation you will happily apply at full thickness is delivering more protection than a stiffer one you ration, whatever the numbers say. That reasoning runs through every score on this site and is set out in full on our scoring method page.
The move towards in vitro testing
Testing on human volunteers has drawbacks beyond cost. It requires deliberately exposing people to ultraviolet until their skin reddens, which is an uncomfortable thing to do routinely. A validated in vitro method, where sunscreen is applied to a roughened synthetic substrate and transmission is measured spectrophotometrically, was published as an international standard in recent years and is gradually being adopted alongside the in vivo method.
The in vitro approach removes the human exposure, improves reproducibility and makes it far easier to test many formulations during development. The practical consequence for a shopper is small at present: labels look the same and the numbers mean the same thing. Over time, wider adoption should make high UVA performance cheaper to demonstrate, which is a good outcome for the part of the spectrum that has always been under measured.
What the UK label categories mean
UK and European labelling uses four categories rather than treating every number as distinct. Low protection covers SPF 6 and 10. Medium covers 15, 20 and 25. High covers 30 and 50. Very high is reserved for 50+, which is the highest claim permitted. Nothing may be labelled above 50+, because the measurement error at that level exceeds the difference being claimed.
The practical translation is straightforward. SPF 15 filters around 93 per cent of UVB, SPF 30 around 97 per cent and SPF 50 around 98 per cent. Those differences look trivial written down and are not trivial in use, because they describe the transmitted fraction. The UVB reaching your skin under SPF 15 is roughly double what reaches it under SPF 30. Given that nobody applies the full dose, the higher number is best understood as margin for error rather than as a licence to stay out longer.
What SPF does not tell you
SPF says nothing about how long you may safely stay outside. The old rule that multiplied SPF by the time to burn without protection has no basis in the way UV exposure accumulates through a day, and it encourages exactly the behaviour that leads to burning.
It says nothing about water resistance, which is a separate test with its own criteria, covered in our water resistant sunscreen review. It says nothing about photostability, so two products with the same SPF may differ substantially by mid afternoon. It says nothing about visible light, which matters for pigmentation. And it says nothing about how the product will behave on your skin, in your climate, on a windy Tuesday when you applied it in a hurry.
Read SPF as one measurement among several. Then read the UVA mark, then the water resistance claim, then the ingredient list. For general UK advice on choosing a factor, the British Association of Dermatologists and Cancer Research UK both publish clear, non commercial guidance.