Illustration of a man using a phone as a scanner for custom clothing

Finding a shirt cut exactly for your body shape used to require a visit to the tailor, or hours flipping through the racks and then trying to get the key for a change room. Now, all you need is an app. (Illustration by Matthew Billington)

Features | From Pivot Magazine

Sizing matters

Online shopping means it was hard to nail the perfect fit. Now, AI is helping fashion e-tailers determine not only your taste, but your exact size. It’s the new bespoke.

A Facebook IconFacebook A Twitter IconTwitter A Linkedin IconLinkedin An Email IconEmail

Finding a shirt cut exactly for your body shape used to require a visit to the tailor, or hours flipping through the racks and then trying to get the key for a change room. Now, all you need is an app. If you download MTailor, the one co-founded by Stanford graduate Miles Penn, you prop your phone against a wall and slowly spin in front of it for 15 seconds as it scans your body size (wearing a tight top, or no top at all, is key). A bespoke shirt is then made to measure in your colour and style preferences and, for an average of $90 including shipping, arrives on your doorstep within two weeks.

The app has a 4.8-out-of-5 approval rating on iTunes and, according to Penn, is 20 per cent more accurate than a professional tailor, based on an in-house experiment the company conducted in which tailors competed against phones. In addition to accuracy, one of the major advantages of the app is that it can “measure more people in one day than an in-person tailor could in a lifetime,” Penn says.

Given the growth in the online fashion market, Penn’s estimate might soon be put to the test. Globally, e-commerce revenues are expected to increase 48 per cent over the next four years, to US$713 billion in 2022. Meanwhile, traditional retailers are dying. In 2017 in the U.S. alone, almost 7,000 stores closed, 13 per cent more than were shuttered during the recession of 2008. Custom fashion was long thought to be immune, but now online start-ups are harnessing technology such as AI, virtual reality and data analysis to create customized garments at a reasonable cost, without ever pulling shoppers out of the house.

Montreal-based Frank and Oak isn’t yet asking you to use your phone as a scanner, but it’s using technology to redefine the shopping experience. The brand sells trendy casual and office wear, and though it has a handful of retail stores, a significant portion of its business is based on a clothing subscription plan that delivers a box of three or so items to customers each month. At the outset of the subscription, shoppers fill out a questionnaire—including their size, what colours they like—and then receive personalized recommendations every month.

“The app can measure more people in one day than an in-person tailor could in a lifetime,” says MTailor co-founder Miles Penn.

The items (priced somewhere between the Gap and Banana Republic) are chosen predominantly by an algorithm that studies the customer’s answers, compares them to similar customer profiles, and picks garments that profile type might like. The items are then mailed to shoppers to try on at home. What they like, they pay for; everything else is mailed back at no charge. (If they return everything, they pay a $25 “styling fee.”)

With every purchase, return, complaint about sizing, click and like, AI builds a more detailed portrait of the customer. “It’s exactly like working with a personal shopper,” says Frank and Oak co-founder and CEO Ethan Song. “The recommendations get stronger as the relationship goes on. Except the AI allows us to offer the customized service on a much larger scale, at a much lower price.”

San Francisco–based undergarment maker ThirdLove has also begun using AI-guided data analysis to customize its products. By studying the bust measurements of its shoppers, the lingerie company realized many women required specialized sizing, including half-sized cups, they hadn’t been offering. Nor were those sizing options being offered by traditional retailers such as Victoria’s Secret, which currently sells roughly half of women’s underwear in America.

So in 2017, ThirdLove launched a beta-program, adding 25 bra sizes to bring its total number of options to 70. “That sold out in eight days,” says co-founder Heidi Zak, a former Google executive. Upon the program’s official launch, “there were 1.3 million women on ThirdLove’s waitlist for these new sizes.”

The granular customer profiles AI provides have also helped companies like ThirdLove and Frank and Oak improve their back-end efficiency. “With data, we see what’s trending in terms of purchases, likes and other feedback, so we know exactly how much of what products we should order,” says Song. Likewise, “ThirdLove is able to offer more sizes than most brands because we are online-only,” says Zak. Rather than investing in mall space, they can put more money into the quality of their inventory. This in turn has fuelled expansion at both companies. Earlier this year, ThirdLove was described as one of America’s fastest growing companies by Forbes magazine, due to its 400 per cent average annual revenue growth in each of its five years of operation (sales at Victoria’s Secret, meanwhile, have been declining since 2016). And at Frank and Oak, between 2012, when it launched, and 2015, revenues exploded by 18,000 per cent.

Portuguese footwear start-up Undandy takes customization one step further. It has built an online 3-D modelling tool that allows men (just men, for now) to design their own custom brogues, boots and sneakers. The interface offers a staggering 156 billion combinations of material and aesthetic choices. (The shoes themselves, notably, are handmade by European cobblers, the same way they’ve been made for generations.)

A custom pair of brogues costs about $245, and anything that doesn’t fit is altered free of charge until it’s right. “The next plan,” says co-founder Rafic Daud, “is that we go into 3-D printing, so we’ll be able to build a custom cast around a customer’s foot, to create the best fit.” Now, if someone could write an algorithm to solve the eternal problem: how to match your shoes with your belt.