Everyone wants the shortcut. Upload a selfie, answer a few questions, and an app hands you a season in ten seconds. Winter. Autumn. Done.
I understand the appeal. I also don't believe the answer — and I say that as someone who has spent years draping faces in person. Here is the part nobody expects me to say: the longer I do this work, the less convinced I am that it can be done accurately from a photo. By anyone. Including me.
That is not modesty. It's calibration.
Your photo changes your colors before anyone sees it
A color analysis is a measurement. And a photograph is a measurement that has been altered at least four times before it reaches a human eye — or an algorithm.
Your camera guesses the light, then "fixes" it. Every phone and camera runs automatic white balance: it looks at your scene, guesses what the light source is, and shifts the colors to compensate. Think about what that means. The single most important question in your analysis is whether your coloring is warm or cool — and your camera has already made its own private decision about that and adjusted your skin accordingly. It can literally neutralize the evidence.
"Natural light" is not one light. Sunlight at noon sits around 5,500K. Golden hour drops toward 2,700K — deeply warm. Open shade can run past 10,000K — strongly blue. The same face, photographed honestly in three kinds of perfectly natural light, produces three different color stories. Add the green bounce off a lawn, the warm bounce off a brick wall or a wooden room, and the second light source from indoors, and your selfie is a committee decision, not a measurement.
Your phone throws away chroma on purpose. Photo files are compressed, and the standard trick — chroma subsampling — keeps brightness detail sharp while averaging the color detail down. Chroma is one of the three dimensions I am paid to judge. Your phone decided you wouldn't miss it.
No two screens show the same color. Phone displays can't reproduce the full range of real fabric color, and most now adapt on their own — warming at night, shifting with the room you're sitting in. So even if a photo survived the camera honestly, the version on my screen is not the version on yours, and neither one is the color of an actual drape held against actual skin.
An AI working from your selfie inherits every one of these problems, then pattern-matches the result against labeled examples. This is why the apps contradict themselves: same woman, different photo, different season. The input changed, so the output changed. That's not an analysis. That's weather.
What I'm actually reading is subtler than a photo can hold
Real draping is controlled comparison. In my studio the environment is deliberately boring: a neutral gray setting, calibrated drapes manufactured to exact specifications of hue, value, and chroma, consistent light, and your skin bare — because makeup is color sitting on top of the evidence I'm trying to read.
Then we change one variable at a time and watch your face. The changes I'm looking for are subtle and they are specific: does the skin look clearer or muddier? Do shadows under the eyes deepen or lift? Does the jawline look firmer? Do the eyes brighten, or does the color wear the woman instead of the other way around?
And here is the part a single good color can never prove: your diagnosis is a formula, not a favorite. Hue, value, and chroma all have to be right at once — and when they are, the same response shows up again and again across many different drapes, not just one flattering scarf. One good color is an anecdote. A consistent pattern across your whole formula is a verdict. When all three dimensions align, the stars line up — and you can watch it happen, live, in the mirror. That consistency is what makes the result trustworthy enough to build a wardrobe on.
No photograph preserves that. A photo freezes one lighting guess, one processing guess, one display guess — and asks me to read whispers from it.
A friendly footnote, from a recent session
One of my recent clients is an AI consultant — this is literally his field. Before his appointment, he did what his instincts told him to do: he gathered 29 photos of himself, in different colors and different backdrops, and asked AI for his palette. It told him Bright Winter.
He arrived at my studio, wrote that prediction down, and folded it up like a sealed bet. Then we draped him, properly, for an hour. When we unfolded the paper, the drapes had disagreed with the algorithm — completely.
I'll let him tell it himself:
His verdict at the end is the one I'd put on the studio wall: the trained eye has it; the algorithm, at this point, does not.
Could I guess from your photos? Probably. I won't.
Give me a set of selfies and I would have an idea. I could tell you what almost certainly won't work. I could probably narrow you to two or three candidate seasons. After years of this work, pattern recognition is real.
But a shortlist of maybes is not a diagnosis, and selling it as one would be a guess wearing a lab coat. My certification through True Colour International comes with uncompromising standards for a reason: the method only means something if the measurement is real. I am not comfortable charging you for certainty I don't have, and I'm not willing to let "close enough for online" become the standard with my name on it.
So I don't offer virtual analysis. Not because I can't use a webcam — because I won't pretend a webcam is a studio.
For women who don't build on guesses
If an app has given you a season, treat it the way a scientist treats a promising hunch: interesting, unproven. It might be right. It might be your camera's opinion.
The women I work with tend to have the highest standards — for their work, their homes, their lives. They don't build a wardrobe, a makeup bag, and years of shopping decisions on a hunch. They come in, sit in the gray room, and watch their own face deliver the verdict, drape by drape.
That's the standard. In person, calibrated, unhurried — and no guessing.
Book an in-person color analysis in Salt Lake City: Schedule — https://colorcompass.as.me
