Home Guides What a skin score can tell you
What a skin score can actually tell you
The short answer
A skin score is one number a program computes from one photograph, and a skin age test is the same machinery with a different label — one photograph, one number, printed in years instead of points. Either can tell you exactly one thing honestly: which way you have moved compared with your own earlier photograph taken in the same light. Neither can tell you how your skin compares with anyone else's, and neither is a fact about you — it is a reading taken under particular conditions, two of which (the light and the distance) usually move more between two photos than skin does in a month. The question worth asking any scorer, including ours, is not how right the number is but how far it wobbles when nothing has changed. We looked for a published answer to that from every app in our comparison table and did not find one.
What a skin score is computed from
Every product in this category does the same three things: take a photograph, reduce it to a handful of quantities, and combine those into one headline number. The interesting part is never the arithmetic at the end — it is which quantities went in, and whether they are things a camera can honestly report on. Ours are five, they are deliberately general, and none of them is the name of a condition:
That vocabulary is a design constraint rather than a style choice. Pick clinical words and the number starts implying it knows something it does not; pick five plain ones and the worst a reader can do is over-trust a decimal place.
The question nobody in this category answers
Photograph the same face twice, ten minutes apart, having changed nothing at all. How far apart are the two numbers? That figure — call it the wobble — is the thing that decides whether a score can register anything real, because a change smaller than the wobble is indistinguishable from noise. It is cheap to measure and nobody publishes it.
We do not have that figure either, and this page is not going to pretend otherwise. What exists is the method: the Same-Light Study was registered in public — what will be reported, and what will never be reported — before a single reading was collected. Collection is in progress and there are no results on this site yet. When there are, they go up whatever they say.
Why the same face scores differently twice in ten minutes
Because the light moved, and a camera has no way to tell the difference between a face that changed and a room that did. Our own lighting checker reads two numbers off the frame — average brightness on a 0–255 scale, and how much variation there is between light and dark pixels — and applies four published thresholds to them:
| Reading | Threshold | What it means |
|---|---|---|
| Brightness under 42 | Too dark | Not enough light on your face for two photos to be comparable |
| Brightness 42–62 | Usable, but only from the same spot | Works if you come back to exactly the same place every time |
| Brightness over 218 | Overexposed | Highlights are clipped and that detail is gone for good |
| Detail under 12 | Too flat | Usually too far away, out of focus, or pointed at a wall |
Those four numbers are first-party: they are the constants in the free checker on this site and they are held identical to the ones in the iPhone app. You can point the checker at your kitchen window, then your bathroom ceiling light, and watch a number that has nothing to do with your skin move by more than a month of skincare plausibly could. That is the confound, and it is measurable in about a minute.
Skin age tests: the same machinery, a different label
A skin age test takes the same photograph, computes the same kind of quantities, and prints the answer in years instead of points. Nothing about the underlying reading is more solid for being expressed as an age — but the label changes how it is read, because a number of years sounds like a fact about you rather than a measurement taken in one room at one time of day.
It is worth knowing where the plausibility of these numbers usually comes from. The popular story is that skin renews itself on a 28-day cycle, so a photograph is a snapshot of a tidy, well-understood process. That figure has been questioned since 1972, when a re-examination in the British Journal of Dermatology put full epidermal renewal nearer 40 to 56 days — and the range varies by person and by age. A test whose output shifts between Monday and Tuesday is not tracking a process that takes six to eight weeks; it is tracking your lighting.
We do not produce a skin age number and are not planning to. What SkinLately gives is a reading compared only with your own earlier readings, which is a smaller claim and one a camera is actually in a position to support. The timescales page has the longer version of why six weeks is the floor for judging anything.
What a score is genuinely good for
- A direction, not a verdict. Up or down against your own Day 1, read at a date you chose in advance rather than on the morning you feel like checking.
- A reason to keep the photo series. The number is a handle on a series of pictures; the pictures are the evidence. The side-by-side maker is where a month of them becomes one comparison you can actually hold still.
- A prompt to hold the conditions. The discipline a daily score imposes — same spot, same light, same time — is worth more than the score is.
What this page is not
- Not a scorer. There is no "scan my skin" button on this website and there is not going to be one; the browser tools have no server to send a photograph to. The demo runs on sample data and says so on every screen.
- Not a claim about how right any number is. Ours included. That figure does not exist yet, here or anywhere else we could find.
- Not medical. Nothing here names a condition, gives a severity or rates how anyone looks.
Frequently asked
What is a good skin score?
There is no such thing as a good one in the abstract, and any product that tells you otherwise has invented a population to compare you with. A score is a reference point: useful next to your own reading from four weeks ago, meaningless on its own. Two people with the same number have learned nothing about each other.
Why did my score change when my skin did not?
Almost always the light, and it is not a small effect. Brightness and detail in a frame move dramatically between a window at 8am, a window at 6pm and a bathroom ceiling light, and everything computed from those pixels moves with them. Measure your own three lights — it is free and takes a minute, and it is the fastest way to see the size of the problem.
Is a skin score the same as a skin age test?
Mechanically, yes: both take one photograph and return one number from it. The difference is the label on the output, and the label is the part that does the damage — an age reads as a fact about you rather than as a reading taken under particular conditions. The section above has the longer version.
Does SkinLately publish a figure for how right its number is?
No, and we are not going to invent one. What can honestly be measured is consistency — how far a reading moves when nothing about your skin has changed — and the Same-Light Study is the method we registered in public to measure it. The figure goes on this site when the study finishes, whatever it says. Until then there is no number here, and a page that gave you one would be making it up.
So is a score worth anything at all?
Yes, under one condition: that it is compared with your own earlier reading taken in the same light. That is a narrow claim and it is the whole product. A number that only ever answers "compared with your Day 1, which way has this gone?" cannot rank you against anyone, and there is nothing in it to over-read.
Why not just build a scorer on this website?
Because a browser tool that scored your face would have to send the photograph somewhere, and the two photo tools here are built so that there is nowhere to send it. The reading lives in the iPhone app. What is on the web is an interactive demo on sample data, which is labelled as such on every screen.
Where these numbers come from
We cite, we do not endorse. None of the sources below has any relationship with SkinLately, and none of them has reviewed this page.
- American Academy of Dermatology, "How to maximize results from anti-aging skin care products" — most products need at least six weeks, and sometimes three months, before there is anything to judge. aad.org (retrieved 15 August 2026)
- K. M. Halprin, "Epidermal ‘turnover time’ — a re-examination", British Journal of Dermatology, 1972 — full renewal nearer 40–56 days than the popular 28. A number that swings between Monday and Tuesday is not tracking that. onlinelibrary.wiley.com (retrieved 15 August 2026)
- The apps referred to on this page are the 8 in our comparison table, each checked against a source it controls on 2026-08-15 — for example App Store listing and joinskor.com. We looked for a published repeat-measurement figure in each of those sources and did not find one — including for our own product, which has none either.