One size fits most (at best), and a women’s sizing system created during the Depression doesn’t gibe in an e-commerce business that prizes personalization.
That’s why startups are turning to new technologies like 3D scanning and machine learning software to produce customized clothing that can be made for the masses.
“People want to buy a size,” says Patrick Donnelly, director of marketing at Fit3D, a San Mateo, CA-based maker of e-commerce software. “But we have the technology to make clothes fit people. It’s not a size 7; it’s a size [person’s name.]”
Fit3D has a body scanner that provides data to the apparel and garment industries through a subsidiary company called BodyBlock AI. Donnelly says the company has created more than 700,000 avatars from body scans of people in in the U.S. and 45 other countries. One-third of the scans are male, and the remainder are women whose ages range from 25 to 80.
“Brands typically choose to make their signature pattern on a size medium, and then scale that in Excel to other sizes,” says Donnelly, who’s also in charge of marketing at BodyBlockAI. “But you can’t scale a medium-sized person into an extra-large person.”
About 68 percent of the U.S. population is considered overweight, Donnelly adds, but only 23 percent of apparel currently made can fit them.
Given that there is no standard to women’s sizes—an 8 in one brand is not the same as an 8 in another—a lot of clothing gets returned. Shoppers now buy multiple sizes of the same item online, have them shipped to their door, keep the one that fits, and send all the rest back to the retailer.
All that buying and returning adds up, costing $351 billion in lost sales in 2017, about 10 percent of total retail sales, according to consultancy Appriss Retail.
It turns out, we’ve known that poor fit means outsized returns for nearly 100 years. The Mail Order Association of America was noting women were returning clothing because of poor fit in the 1930s, according to a Slate article.
“The Department of Agriculture measured only women in their 20s on the East Coast who were white,” says Merin Guthrie, founder and CEO of Kit in Houston.
Guthrie, an amateur sewer, got interested in the apparel business after she designed and sewed bridesmaids dresses for a friend. “This was the first time that these women had had a dress made for them,” she says. “I have vintage dresses of my grandmothers and they fit great. The experience that my grandmother had available to her we just don’t have any more.”
Those women asked Guthrie about handmade dresses made for work, eventually leading her in 2015 to found Kit, an analytics-based clothing label, with her college roommate, a former Army officer. (That’s where the name “Kit” comes from.) Guthrie spent the first year taking measurements of any consenting women that came in her path.
“Any bachelorette party I went to, ‘Step right up. I want to measure you.’ ” she says. “I was looking for a diversity of builds. In exchange for leading a strategy planning process for women in a roller derby league in Austin, I was measuring women on skates. I had a bikini waxer I really liked—you really trust a woman giving you a bikini wax—I trained her and she measured a bunch of women for me.”
Those measurements gave Guthrie the foundation upon which to build Kit’s “fit algorithm.” Guthrie says she spent months with the mantra of: “Oh, it’s so nice to meet you. Can I measure you?”
Kit has no sizes. In its “off the rack” category, the site offers a variety of dress and blouse styles, with a few skirt options, which are available in about 20 kinds of fabric. The shopper chooses, say, the A line sheath or a safari dress and then can pick from around eight to 10 fabrics. Depending on the garment, Kit asks for a preference on sleeve or hem length, pointed versus mandarin collar, and the like.
“A 5’7” woman with a long torso and pear-shaped figure, who weighs 140 pounds, creates a fit profile, and we then are able to go into our data sets and find the perfect avatar for her: someone who is that same height, weight, and build,” Guthrie says.
Instead of sizes, Kit works with five core body types: hourglass, straight, pear, apple, and busty. The algorithm is continually updated through additional measurements taken for Kit’s custom orders, she adds.
Guthrie says the company has raised about $140,000 from friends and family and is now looking to move into a hybrid production/brick-and-mortar store aimed at generating foot traffic and brand visibility.
Kit currently is housed in an apparel factory staffed with resettled refugees that come from countries with strong clothing businesses. “In a weird way, the refugee resettlement system is re-shoring that talent,” Guthrie says. “We’re leveraging this workforce and then we want to create training programs of our own.”
Guthrie adds that as Kit grows, some of these processes will likely have to automated—like using a laser cutter to cut garments instead of doing it by hand.
BodyBlock AI doesn’t manufacture clothes, but rather works with retailers and brands such as Stitch Fix to connect shoppers with items that fit and won’t be returned.
The company works with its customers in two ways. Shoppers browsing clothing click on a “Find My Fit” box and enter basic information like gender, height, age, and weight, which will be analyzed against the company’s bank of 3D body scans. The algorithm offers the shopper a display of scans of several body types and the shopper picks the one that most resembles them. Then, the algorithm recommends the appropriate size for the article of clothing.
For three particular clothing brands—shirtmaker Stantt, apparel line Sene, and denim maker Unspun—BodyBlock AI has an arrangement by which shoppers can get their own body scans done (Fit3D says it has scanners in health clubs in the U.S. and 48 other countries) and create an account with BodyBlock AI. Shoppers then log onto their account with the retailers and fit their fit.
Either way, BodyBlock AI says its technology can give those retailers and brands hard-to-come-by market intelligence such as which sizes people are shopping for, who made purchases and who didn’t, and the measurements of those individuals. The idea, the company says, is to provide deeper insight than just which SKUs are selling or not.
“No one has perfected online analytics from the metric of body shape,” Donnelly says. “They know what goes in their Starbucks coffee or [their] marital status but don’t know their true waist size. We’re setting a new standard for brands.”