Does color analysis actually work?

The perceptual part, and you can test it yourself in about ten seconds. Stand at a window, hold a mustard top under your chin, then a dusty pink one, and watch the skin under your eyes. Shadows deepen or lift; redness around the nose gets louder or quieter; the whole face looks either rested or slightly unwell. Nothing about your skin changed — the comparison did.

That is simultaneous contrast: a color shifts the appearance of whatever sits beside it, and your collar sits beside your face. It is ordinary color science, it is repeatable, and it is the entire foundation the industry stands on.

Two smaller things are also real. Consistent color makes a wardrobe combine better, because pieces chosen from one palette agree with each other by construction. And there is a confidence effect — people who believe they are wearing their colors carry themselves differently. We would rather name that than pretend it is measurement.

The taxonomy. There is no ground truth for “what season is this person” — no register, no instrument, no consensus test that settles it. The number itself is a choice: Carole Jackson popularized four seasons, the modern standard is 12, and some systems run 16. All of them are drawing lines through the same continuous space of human coloring.

So the honest framing is that a season is a useful label for a region of that space, not a fact about you. It earns its keep by making decisions easier, not by being provably correct. That is also why we say plainly, on every page, that this is style guidance rather than any kind of clinical or dermatological assessment.

Because four things vary, and all of them move the answer:

Lighting. The largest single factor. North-facing daylight, a warm bulb and a phone screen produce three different faces. An analyst working in a controlled room and an app working from your bathroom selfie are not looking at the same input.

The system. A four-season analyst has to round you somewhere; a 16-season analyst splits hairs a 12-season one would not. Different maps of the same territory produce different place names.

The day. Fake tan, a flush, a fresh dye job, foundation half a shade off — any of them can tip a borderline read.

Genuine overlap. Neighbouring seasons share most of their palette. When you sit on a border, two competent analysts looking at the same face can land either side of it and both be right about what they saw.

A disagreement between sisters is a normal outcome, not a scandal. A disagreement between opposite corners of the system — light and warm versus deep and cool — is a sign that one of the reads was done in bad light, or that the two analyses disagreed about your undertone in the first place.

Sister seasons are seasons that agree on two of the three axes and part company on the third. They look alike in a mirror because most of what drives the look — the depth, or the clarity, or the warmth — is identical; only one dial has moved.

These are the pairs people mistype themselves as most often, each one straddling a border between two families:

Being between two seasons is the common case, not a failure of the analysis. The practical move is to treat the colors both palettes contain as your safe ground, and test the ones only one palette has in daylight, on your own face — the 12 season pages list every palette with its hex values so you can do exactly that.

They win on different things, and we would rather say so than pretend otherwise. Draping beats us on control and on the experience of watching your own face change. We beat draping on cost, speed and repeatability. If you have the budget and a good analyst nearby, in-person is a genuinely lovely thing to do.

The fair summary: in-person is the better experience and the better controlled read; online is the better deliverable per dollar, and it is repeatable. Neither is a laboratory result, because there is no laboratory version of this.

Four things it is regularly claimed to do and does not:

Whether you are being draped, testing yourself, or uploading a selfie, the input decides the answer. Five things do almost all of the work:

Natural daylight, facing a window. Not a bathroom bulb, not overhead office light, not a ring light. Indirect daylight on your face, with the window in front of you rather than behind.

No filter, no beauty mode. Smoothing and “natural” filters shift skin hue, which is the one thing being measured. Turn them off, including the ones your phone applies by default.

Hair back, ears visible. Hair color crowds the read, especially if it is dyed. Pulling it back lets the skin be the subject.

Minimal makeup.Foundation is the problem — it is a layer of somebody else’s undertone. Bare skin, or as close as you are comfortable with.

Plain top, plain background. A bright wall or shirt bounces its own color onto your face and will fight the analysis.

Lighting is unforgiving enough that we plan for it: every Palettewell report includes one free redo with a different photo. If the first read feels wrong, take a better-lit one and run it again — that is what the redo is for, and it is not an extra charge.