Pick up a pair of jeans marked “size 10” at practically any North American clothing store, and there’s a good chance that the same waistband would have been referred to as a “14” in the 1970s. This change was not formally announced. Season by season, brand by brand, it simply happened as merchants discovered that consumers make larger purchases when they feel good about the number on the tag. There was no conspiracy behind vanity sizing. It was a business plan. a quiet, industry-wide consensus to put emotions ahead of metrics.
It’s difficult to ignore the numbers behind this. An American size 8 waist measurement has increased by almost four inches over the past fifty years, according to research cited by the fit-tech platform Genlook. Four inches. That is a structural change in the meaning of labels, not a rounding error. Furthermore, the downstream expenses are now truly astounding. An estimated £190 billion in returns are handled by fashion retailers worldwide each year, primarily due to inconsistent sizing, with customers citing ill fit as the primary cause.
The industry may have known for years that this was unsustainable but lacked a way to address it. AI sizing engines, which don’t give a damn if a customer feels better being told they’re a medium, may be a covert manifestation of that mechanism. The fit of the garment is important to them.
Businesses like True Fit, 3DLook, and EasySize have been developing tools that predict fit at the individual level using fabric behavior data, smartphone body scans, and purchase histories. These are not gimmicks for virtual try-ons. This technology’s more advanced layer operates further upstream, even before clothing is delivered to stores.
Fit Collective, a UK startup that has raised £3 million in pre-seed funding, has developed a software dashboard that extracts sizing insights from manufacturer logs, returns data, and fabric behavior. The dashboard then informs production teams of the precise areas where a garment is failing before the next run starts.
Phoebe Gormley, the founder of Fit Collective, discovered something startling when she began examining real clothing data. She examined 179 women’s shirts at a high-street store that were all labeled as size 12 and discovered that the smallest and largest measurements varied by 66 centimeters.
The same number can mean completely different things depending on the shirt you choose, so it’s not a subtle variation. It makes sense that women return clothing at rates of 40 to 50 percent, sometimes as high as 60 percent, for high-end womenswear as opposed to about 15 percent for menswear.
One lenient size block at a time, the fashion industry seems to have created this issue gradually while secretly hoping that someone else would find a solution. It’s possible that AI is driving the problem because it is now impossible to justify the financial cost of ignoring it. A feel-good label can no longer compensate for the cumulative effects of inventory write-downs, lost customer trust, and return logistics.
However, it would be unrealistic to expect technology to solve a problem this deeply ingrained in the fashion industry. When Paul Alger of the UK Fashion and Textile Association points out that fit is intrinsically subjective and that body measurements rarely match a number on a label, he makes a valid point. Individuals are not all alike. Different people have different preferences. One customer’s tight fit might be the wrong size for another. Although the data can help you get closer to accuracy, it most likely won’t be able to do so completely.
It’s becoming more and more obvious that AI sizing engines are moving the conversation in a significant way, away from labels that are meant to flatter and toward dimensions that are meant to fit. That change isn’t ideological for North American retailers keeping an eye on their profit margins and return rates. It’s useful. The era of vanity sizing might not come to an abrupt end. It is more likely to fade out in the same manner that it came in: quietly, season by season, as the data makes it more difficult to defend the previous strategy.
