The answer, first
Black, at 28.8% of everything people wear. White is second at 19.2%. Nothing else breaks 9%.
832 distinct garments read off 876 real New York accounts across 4,464 posts. Deduplicated: the same jacket worn in six posts counts once. Bars scaled to the largest value.
Most of a wardrobe has no color in it
Add up black, white, grey, beige, charcoal and cream and you get 58.3%. A clear majority of what people wear has no chromatic color at all. Add denim and two thirds of everything sits in black, white, grey or blue jeans.
That is the number worth sitting with if you plan ranges. The colorway decisions that feel most creative are being made over a third of the wardrobe, not the whole of it. Everything from pink through to the long tail of oranges, yellows and purples fights over the remaining space.
The split inside a category is where it gets useful
An aggregate palette is interesting. A palette per garment type is actionable, and the two do not look the same.
Black takes over half of jackets and over half of eyewear. Shirts and shoes flip to white. Denim owns exactly one category, and it is pants. So "black is 29% of the wardrobe" is true and slightly misleading: black is where outerwear lives, white is where the top and the sneaker live, and if you are picking one colorway for a jacket drop the aggregate number is not the one you want.
The black itself is also specific. It reads as a washed, worn-in near-black across the sample, nothing like the flat black of a swatch book. Know an audience's palette at that level and colorway decisions stop being taste and start being measurement.
How this differs from survey data and sales data
Three methods answer three different questions, and it matters which one you are reading.
Stated preference. Ask people and they tell you what they think they wear. A 2022 Statista survey of US consumers found neutrals, black, white and grey, the runaway most popular color scheme, with Gen Z overwhelmingly in the same place. Directionally right, and it cannot tell you that jackets and shoes disagree.
Sales. Retail data records what got bought. It includes the returns, the impulse buy worn twice, and the sizes that sat on the shelf. It is the best signal for what a store moved and a weak signal for what a person wears.
Observed wear. What was actually on bodies. That is what we measure, and as far as we can find it is the only public dataset of it at garment level.
The striking part is how closely the first and third agree. When Tinder analysed 12,000 profile photos of 18 to 40 year olds in New York, Los Angeles and Atlanta, 30.6% of women and 32.3% of men were wearing black. That was 2016, a different population, and a photo people chose deliberately to attract someone. We got 28.8% in 2026 from unposed everyday posts. Two methods a decade apart, landing within two points.
That convergence is the useful part. It says the neutral dominance is real and stable, not a quirk of one sample. Which means the interesting question was never "what color do people wear," it is the one underneath: which garment, which colorway, on which people.
How we measured it
Briefly, because the full method is published separately with its error bars.
We reached 876 New York accounts by chain referral, moving through people's real connections rather than searching hashtags, so the sample is not just whoever performs the city loudest. Every visible garment across 4,464 posts was found and cut out by a computer vision model we fine-tuned, on the main subject and on anyone else in frame. Only 2% of the frames were outfit posts, which is the whole point: nobody has to be posting about clothes for us to read what they are wearing. Crops were then clustered so the same item worn repeatedly counts once, and each was read for color.
What this data does not tell you
Two limits, stated plainly, because a color chart with no error bars is decoration.
First, we require three near-duplicates before a garment becomes a node in the map. That removes 18% of crops, and it removes them unevenly: dresses survive at 56.8% against sunglasses at 95.5%. A dress nobody else in New York owns cannot find two twins. A plain white tee always can. So this chart slightly under-reports exactly the garments that make a wardrobe distinctive, and the true long tail is a little fatter than 1.4%.
Second, this is one city in one season. Palettes move. The number we would defend is the shape, neutrals taking a clear majority and one category disagreeing with the aggregate, not the second decimal place on beige.
Why we published the numbers
Most color reporting in fashion is either a survey of what people say or a trend forecast of what someone thinks should happen next. Both are useful. Neither is a measurement of what got worn.
We build that measurement as infrastructure, per audience, and point it at whatever cohort a brand cares about: a city, an age band, or a competitor's customers. The New York numbers above are one output of it. The same pipeline aimed at your audience returns their palette, their garments, their brands, with the evidence frames attached to every figure.