Who writes this and how we calculate it

Last updated: August 31, 2026

Personal color analysis is advice: someone reads your face and tells you what suits you. Before you follow advice, you have the right to know who gave it and how it was reached. This page answers both.

Who writes this

The content and the method behind TOP are the responsibility of the organization itself — TOP — The Outfit Planner — not of an individual contributor. The color guides are written by the product team that builds the app: the same people who write, test and maintain the code that computes the result.

We are not certified image consultants, and we claim no degree, certification or credential in color analysis. What we do have is a written, measured and auditable method — described below, without euphemism.

The method, in four steps

TOP asks for 5 to 8 photos of you under different lighting and returns a color season. This is the path between the two:

1. The photo is measured, not judged. A vision model does not pick your season. It only reports numbers: the color of your skin, hair and eyes in each photo. Before those colors are read, the photo is calibrated against achromatic anchors — sclera, teeth, a white or grey object in the scene — to estimate the light cast and subtract it. Phone photos always carry a cast; ignoring it means reading the lamp, not the person.
2. Several measurements, not one. Measurement has measurable noise, so each analysis runs multiple independent readings and aggregates them by median, photo by photo. A single outlier reading does not change your result.
3. A rule decides, not a hunch. The measurements feed a deterministic function. It converts the colors into a perceptual space and evaluates four axes: depth (how light or dark the skin/hair/eyes set is), temperature (how warm or cool it leans), contrast (the lightness gap between skin, hair and eyes) and chroma (how vivid or muted the color is). The verdict comes from the table, not from the model's opinion.
4. When it cannot be called, we do not call it. Temperature has an explicit neutral band. Anyone who falls inside it is not forced onto a side: they get Neutral — and the guidance switches to depth instead of temperature. A deep neutral is well served by rich colors from both sides. Forcing a season on someone sitting on the line is the mistake that costs the most money to someone about to buy clothes.

Why the same face always gets the same result

Given the same measurements, the rule always returns the same season — it is pure and involves no randomness. For people near a boundary there are two extra defenses: a re-analysis only flips your verdict if the new measurement crosses the boundary with margin (hysteresis), and the confidence shown drops when you are close to the edge. In other words: uncertainty shows up on screen instead of becoming a different answer on every attempt.

What the app does with the result

Your coloring feeds a palette and a per-garment match score, computed by color distance in a perceptual space — the same interpretation used by the style consultant, by the "why this score" explanation and by the metal suggestion. There is a single source: if the reading changes, it changes everywhere at once.

The limits — spelled out

It is guidance, not a report. TOP's result is a styling suggestion to help you choose clothes. It diagnoses nothing and has no medical purpose whatsoever.
It does not replace an in-person professional analysis. A consultation with drape fabrics, controlled light and a professional in front of you sees things a photo does not carry. If this matters to you, book one.
The discipline itself is not exact science. The seasonal system has no consolidated empirical validation, and human analysts frequently disagree with one another about the same person. No method — ours or anyone's — is right every time.
Photos have a physical limit. Without a light reference in the scene, color temperature is the hardest part to recover. That is exactly why we ask for several photos under different lighting, and why the answer "Neutral" exists.
We measure how often we are wrong. The calculation is evaluated against a set of photos labeled by a specialist, and every change to the table only ships if it beats the previous one on that test. That improves the method; it does not make it infallible.

Corrections and contact

Think the result does not represent you? You can redo the analysis with new photos inside the app, and you can reach us through support. Errors reported by users have already changed the method more than once. For what we do with your photos, see the privacy policy.

This page describes the method in force as of the date above. When the calculation changes, this page changes with it.