.svg)
One problem,
since day one.
For more than ten years Virtusize has worked on one question and nothing else: will this fit me? More than 500 fashion brands now use us to answer it.
Two beliefs about
size and fit.
A recommendation needs the garment's real measurements
Most size advice comes from a brand's general size chart, or a category average. But two shirts labelled M in the same catalogue can differ by several centimetres. We measure every product individually, in every size. The recommendation is then right for that garment, not right on average.
Fit should be shown, not stated
A size recommendation is just a label. It says nothing about sleeve length, shoulder width or where the hem falls, and the same jacket is right for one shopper and too loose for another. In a shop, you try it on. Online, the shopper should see the garment on their own proportions and choose. We never ask anyone to measure themselves, because that is where people give up.
We moved the company to where the data was.
Most fit models are built on Western shoppers first and adjusted for Asia later. Ours was built on Asia first. This is why our accuracy stays the same when a brand enters a new market, in either direction.
Virtusize started in Stockholm. We moved to Tokyo because the data we needed was published in Japan and almost nowhere else.
Ten years ago, almost every Japanese fashion website already listed the measurements of each product, in every size it was made in. Outside Japan and Korea, almost no one did this. European and American sites are only starting to do it now.
Ten years of building there means our model learned from real product measurements, not from size charts.
Virtusize is now the most widely used fit service in Japan and Korea. We expanded across Asia, and then into Europe and North America.
| SIZE | LENGTH | CHEST WIDTH | SHOULDER |
|---|---|---|---|
| S | 65 | 52 | 46 |
| M | 68 | 55 | 49 |
| L | 71 | 58 | 52 |
| XL | 74 | 61 | 55 |
| SIZE | CHEST, TO FIT | WAIST, TO FIT |
|---|---|---|
| S | 91–96 | 76–81 |
| M | 97–102 | 81–86 |
| L | 103–108 | 86–91 |
| XL | 109–114 | 91–97 |
We did not build this model in one go. We collected it over ten years.
More than 500 brands have added data to it, including adidas, Ralph Lauren, Bottega Veneta, Levi's, Gap, Asics, Acne Studios and L.L.Bean. A model trained on one retailer only learns that retailer's customers. Ours has seen what shoppers of every size bought, and what they kept, in every market we operate in.
Each brand runs on its own version. Enterprise clients start from a category model, then receive logic tuned to their catalogue, which can be adjusted down to a single product. The data is shared. The recommendation is not.
Any vendor can be accurate in one market. Staying accurate in every market takes ten years. No enterprise client has left Virtusize since 2022.
Before Virtusize, fit was only measured when it failed. We defined Size Match Rate, and it is now the standard in this category. We publish how we calculate it, including the orders we exclude, and we exclude very few. A vendor that excludes many reports a better number for the same performance, so ask every vendor for the calculation, including us. We would rather be measured than believed.
Virtusize reduced our size-related returns by 27% year-on-year, across both apparel and footwear. Their continuous improvements to the size recommendation logic give us full confidence in the service.
Naoki Ito, Ecommerce Division, Under Armour JapanOne system, every market, every category.
Brands come to us to solve fit once, in every country they sell in, instead of buying a different tool for each region and hoping the numbers agree. Three things make that possible.
No single global average
We train a separate body model for each region, on that region's own data. We do not take one global average and adjust it. Two models are live today, and we add more as the data allows.
Every category, one engine
Apparel, footwear, bags and kidswear all use the same model. Fit works across the whole catalogue, not only part of it.
Certified globally, compliant locally
One security and privacy review covers every market a brand sells in. You do not have to repeat it region by region. ISO 27001 certified, with GDPR, APPI, PIPA, PIPL, CCPA and CPRA met locally, and ICP certification in China. Shopper data stays in the region where it was collected.
We built this because we had lived the problem.
Between us we have run product pages, built the platforms behind them and answered for the returns. The same thing kept appearing: size charts nobody trusted, returns nobody could stop, and customers who quietly never came back. We came here to fix it, not manage it.
No country's bodies are the standard, and a team that only knows one market will only ever solve one. Eleven nationalities work across the company, in the markets our clients sell in.
United States, Australia, Sweden, Ukraine, Japan, Korea, China, Mongolia, Indonesia, India and the Philippines.
Everyone else gives a generic size. We show the actual fit.
The shopper gives us four things about themselves. We draw the garment on a body built from those proportions, so they can judge the fit for themselves before they order.
Shoppers who can see the fit buy with more confidence, send less back, and return to the brand more often.
