How to Read Your Skin Score Report
Every number on the report is a severity measurement, not a grade. Lower is better, and a 2.4 is an ordinary, healthy result.
Lower Is Better: These Are Severity Scores
This is the single thing worth getting straight before anything else. Each score runs 0 to 10 and measures how much of a concern is present. Zero means the scan found none of it. Ten would mean the most severe reading the scale allows.
So a report full of ones and twos is a good report. Across the faces this scanner has measured, overall scores cluster between roughly 1 and 3 — that band is normal skin, not a failing mark. If you have been conditioned by fitness and sleep apps to read a low number as bad, this will feel wrong for a moment.
The overall figure is a weighted blend of the five traits rather than an average: oiliness and acne each contribute a quarter, dryness and pigmentation a fifth each, and sensitivity a tenth. It exists to rank zones against each other, not to grade your face.
The Five Zones
Your face is divided into the T-zone (forehead and nose), the left cheek, the right cheek, the chin and jawline, and the under-eye area. Every pixel of measurable skin belongs to exactly one of them, with the lips and both eyes excluded.
Left and right are the image's left and right, not yours. The cheek labelled "left" is the one on the left of the photo — which, because the camera faces you, is the cheek on your right side. This trips people up constantly when they are trying to match a reading to a spot they can feel.
The zones are not equivalent to each other, and the scoring accounts for that. The T-zone has many times the density of oil glands that the cheeks do, and its skin sits over nostril creases and larger pores that make texture measurements read higher by default. Dryness and pigmentation are calibrated per zone for exactly this reason, so a T-zone reading of 2 and a cheek reading of 2 mean comparable things.
Under-eye oiliness is always reported as zero. There are effectively no sebaceous glands there, so any oil the scan seemed to find would be light reflecting off thin skin.
The Five Traits
Oiliness measures surface sebum and shine. It is the most stable of the five and the one a single photo captures most reliably.
Dryness measures surface texture irregularity associated with dehydration and flaking. Its baseline is strongly shaped by the geometry of each zone, so it is offset-corrected per region.
Acne is measured differently from the rest. A heuristic pass finds candidate lesions, and every candidate crop is then classified by a dedicated model across four severity levels, whose result replaces the heuristic score entirely. If a zone is obscured by facial hair, acne there is reported as unmeasured rather than estimated.
Pigmentation measures fine-grained variation in lightness, weighted by an estimate of your skin tone and attenuated when the variation shows no colour support — melanin is chromatic, whereas pores, shadow and sensor grain are not. It is the least reliable of the five, and we would rather say that here than let you over-read it.
Sensitivity measures redness and visible irritation. It carries the smallest weight in the overall figure because a photograph captures it least well.
The Confidence Number
Every report carries a confidence value between 0.40 and 0.95. It is not a measure of how bad your skin is — it is a measure of how much the scanner trusts its own read of this particular photo.
It is built from image quality and blur, how much the five zones disagree with each other, how reliable each detector was on this image, and what share of the axes could be measured at all rather than coming back unmeasured. A typical good scan lands around 0.71 to 0.81.
When confidence drops far enough, the consequences are real rather than cosmetic: the recommendations fall back to a maintenance routine instead of introducing any active ingredient. The scanner is built to say "I am not sure" and act on it, which is the whole reason the number is shown to you rather than kept internal.
Unmeasured Is Not Zero
An axis marked unmeasured means the scan declined to produce a number, and that is different in kind from a score of 0. Zero is a measurement: the scanner looked and found none of that concern. Unmeasured means it could not look properly.
The common cause is obstruction — hair across a zone, or facial hair over the chin and jawline making acne unreadable there. A consequence worth knowing: many scans have no chin acne reading at all, and that is the intended behaviour rather than a fault.
Unmeasured axes also feed back into the confidence number, so a report with several of them will show lower confidence as well.
Why Your Report May Differ From What You See
The number shown to you is not always the raw measurement. From your second scan onward, a single unusual reading is damped against your recent history, so a one-off outlier shows only partially until a second scan confirms it. The raw value is what gets stored; the damped value is what gets displayed.
This is deliberate. Photo-to-photo noise is real, and a scanner that swung your routine around on the strength of one grainy Tuesday photo would be worse than useless. If a change is genuine, it persists, and the second scan brings it through in full.
It also means your first scan after a scoring-engine update is shown raw with no deltas, because it has no comparable history to damp against.
Questions People Ask
Bad. Every score is a severity measurement from 0 to 10, so lower is better and 0 means the scan found none of that concern. Most healthy faces land between roughly 1 and 3 overall.
Some asymmetry is genuine — sleeping side, sun exposure through a car window, how you hold your phone. Some of it is the photo: if the light came from one side, the shadowed cheek reads darker and lower-contrast. If the difference is small and swaps sides between scans, it is measurement noise. If it is consistent across several scans, it is probably real. Remember that the cheek labelled "left" is on the left of the image, which is the right side of your face.
How much the scanner trusts its reading of that specific photo, on a 0.40 to 0.95 scale — not how healthy your skin is. It combines image quality, blur, how much the zones disagree, detector reliability and how many axes could be measured. Good scans usually sit around 0.71 to 0.81, and a low enough value makes the recommendations drop back to maintenance only.
Because the scan declined to guess. Usually something was covering that region — hair, a hand, or facial hair over the chin and jawline, which makes acne specifically unreadable there. Unmeasured is not the same as a score of zero, and it lowers the report's confidence value too.
Most often the zone is marked unmeasured rather than zero — check the label. Otherwise, the acne classifier works from lesion crops found in the photo, and a single small or flat mark in poor light may not survive that step. It is also the axis most affected by image grain, so a scan taken in a dark room is the least likely to find it.
Photo-to-photo variation is normal, which is why a single outlier reading is damped against your recent scans before it is displayed. Lighting, camera grain and the exact angle all move the numbers a little even on identical skin.