Color analysis quiz: which of the 12 seasons are you?

Updated November 4, 2026 · How we measure this

What this quiz can tell you, and what it can’t

You answer ten questions about your hair, eyes, skin, jewelry and what people compliment you in. Each answer nudges three scales: temperature (cool to warm), value (light to deep) and clarity (soft to bright). Where you end up picks the nearest of twelve seasons and names the next nearest. There is no email step and no account in front of the result.

What you get is a sensible first guess. You don’t get a measurement, because the quiz only knows what you believe about your own coloring. So its result opens with “Your answers lean”, not “You are”. It rarely claims more than that. We ran 1,500 random walks through the questions, and 1,500 of them ended on “A close call”, 0 on a lean, and none on “A clear result”. A questionnaire that sounded more certain would be overselling itself.

So how accurate is it? We can’t give you a number, and nobody should accept one without a study. Measuring accuracy would take a group of people, a quiz result for each, something independent to compare it with, and a published write-up. We haven’t run that test. Some quiz sites quote a percentage anyway, and the ones we read show no study behind it. Even the thing you’d compare against is soft: the history of the field records a lack of standard training for analysts and calls the seasons “to some degree, arbitrary”.

Why another quiz gave you a different answer

Two quizzes, two names. It’s frustrating, and the causes are ordinary. None of them means that you are hard to type or that you did it wrong.

One answer can tip it

Seasons sit next to each other on the same three scales, so someone near a border can cross it on a single reply. Here is one made-up person: ash blonde, soft hazel eyes, olive skin, some sun, compliments on dusty colors, and a preference for gold. Run through the real maths, that lands on Soft Autumn with Soft Summer close behind.

Look at the second row. Nobody changed their mind about jewelry there. The answer only went from gold to “not sure”, and because gold had been holding this person on the warm side of neutral, Soft Autumn became Soft Summer. The fourth row is the reassuring one: a warmer white nudges the scales and leaves the season alone.

We measured how common this is. Across those random walks, changing one answer to another option changed the top season 47% of the time, and in all but 3 of the 1,500 walks at least one single change would have done it. Random answers include combinations few people would give, like very pale blonde hair with nearly black eyes and deep golden skin, so read that as how touchy the arithmetic is, not as how often real people flip. It still explains the experience.

Quizzes don’t ask the same questions or sort into the same boxes

MySeasonColors sorts people into 16 seasons from 10 questions. FaceRater uses one photo and five questions for 12 seasons. This quiz uses 10 questions for 12. House of Colour, which only works in person, says eye and vein color play no part in its method, and we ask about both. A “Soft Summer” in one system may not match that name in another, and a four-season result can only be coarser than a twelve-season one.

You answered for a slightly different person

Your memory of your natural hair, this month’s tan, what a friend said last week and the lamp you checked your wrist under are all part of your reply. Words do damage as well. Your idea of “soft” may be someone else’s “medium”. A stylist at House of Colour, who sells in-person sessions and so has a stake in this, says one client got three different answers from AI tools. Take that as one anecdote from someone with a reason to tell it.

Take it apart yourself

Below are the same ten questions and the same maths the quiz above uses, with the cover off. T, V and C show how far the chosen answer pulls temperature, value and clarity, after its weight. Plus means warmer, deeper or brighter. Change any answer and the three scales move, with hollow dots marking where they were.

Try this: set the jewelry answer to each option in turn. Then do the same to the veins question. You’ll feel the difference in how much each one is allowed to matter.

What each question is really testing

• Natural hair. The strongest signal for depth, and loud on temperature too. It asks for your color before dye, as you remember it from your twenties if it has changed, because the idea is your own coloring. Grey gets half the usual weight, since it says less about what you started with.
• Eyes. Mostly a clarity question. Clear blue and soft blue-grey are similar in temperature and opposite in clarity, and that distinction is the whole point of asking.
• Sun. A loose hint about depth and warmth. A golden tan leans warm. Burning leans light. It counts for less than a plain color question, because tanning depends on skin biology as much as color, as the Fitzpatrick phototypes show.
• Veins. Weighted low on purpose. Blue and green move temperature only, at 0.36 after weighting. Physics has more to say about vein color than undertone does, and the vein test page shows the reasoning.
• Jewelry. A real temperature signal if you’ve actually compared the two. Rose gold leans slightly warm and slightly soft. “Not sure” counts for nothing, which is why it can tip a close result.
• Which white. Two axes at once. Bright white reads cool and clear, cream reads warm. If you own only one white shirt, this answer tells us about your shopping.
• Compliments. The most human question and a slippery one. People compliment what’s new, what fits, and what they like themselves. It still moves all three scales, because it’s the only question about color on you, as seen by somebody else.
• Hair against skin. Contrast, feeding clarity. Be careful with this one if you have deep skin and dark hair. Those two sit close together in lightness whatever your coloring, so “not much” can undersell you and pull you toward the soft seasons. The camera scan avoids this by also comparing the whites of your eyes with your iris, a contrast every face has.
• A black top. Striking leans bright and deep, tired leans soft and light. “It’s fine” barely counts, since most people say that.
• Skin in daylight. Your own description of your skin, including olive and two deep options. Judging your own undertone is hard, which MySeasonColors names as a quiz’s weak spot too.

Some of these matter far more than others to the final name:

Natural hair moves the name 63% of the time and veins only 20%. So if you redo one thing, redo your hair answer honestly. Anyone who has been coloring it for years should try both the shade they remember and the one they think is closest.

How ten answers become one season

1. Each answer carries a pull on one or more scales, from -1 to 1, and a weight for how much that kind of answer can say. “I can’t tell” has a weight of zero.
2. For each scale, the quiz takes the weighted average of everything that spoke to it and stretches it, by 1.5 on warmth and 1.35 on the other two, capped at the ends, since an average of honest answers rarely reaches an extreme.
3. It also keeps a confidence per scale, based on how much was said about it. That confidence is capped at 0.75, because no reply actually sees your skin.
4. Each of the twelve seasons has a fixed home on the three scales. The nearest becomes your result and the next becomes your runner-up. A scale the quiz is unsure of counts for less in that distance.

The three scales are the old ones from Munsell’s hue, lightness and strength, with hue reduced to warm and cool. Everything on top of them is judgment: the pulls and the weights come from how draping traditions describe each season, and nothing was fitted to data. We checked that all twelve seasons can be reached by a plausible set of answers, and that’s the extent of the validation. Our full approach is on the methodology page.

What to do with your result

Treat the two names as a shortlist. Open both palettes from the palettes page, read how the twelve fit together in our guide to seasonal color analysis, and notice which one you feel less surprised by. Then test them, since your opinion of a color on a screen doesn’t tell you much about that color near your face.

The result has a button for this, “Check it with my eyes”. It keeps your answers and asks for your camera. Say no, and you can upload a photo or use our example person instead. The tool reads your iris, skin and hair from the picture, then shows pairs of colors beside your photo for five to eight rounds and asks which looks better. Your answers count for less than a camera read does, and a few consistent choices can overturn both. At the end you get the full result: season, runner-up, six colors, one to avoid, your metals and a share card. The Full Color Report is the paid step after that.

You can also check the result with the scan, which reads your face from your camera, or use the season test, which skips the questions about you and asks you to look at your face instead. If you have doubts about the whole idea, they’re reasonable, and we go through what holds up there.

When a quiz is the better choice

Sometimes the quiz beats the scan. A dim room at night will mislead any camera, while your memory of your natural coloring works anywhere. You might be on a train, or have dyed hair and heavy makeup on today, or simply want a shortlist first. It’s also a reasonable way to spot a mismatch: if the quiz and a camera scan land on neighbors, you’ve learned which border you live near. If they land far apart, one of them was misled, and the color rounds are the fairest way to find out which.

Related guides and tools

• GuideSeasonal color analysisThe history, the three axes and the 12 seasons, with what the evidence does and does not back.
• Guide12 seasons chartAn original chart of all 12 seasons, their axes, their aliases and how they connect.
• GuideIs color analysis real?What color science supports, what nobody has tested and how to check a result yourself.
• ToolWhat season am I?Three axes narrow down your season, then color pairs beside your photo settle it.