When bra e-tailer ThirdLove was founded in 2013, it was hyper-focused on organizing its product range around fit, rather than numeric size — so much so that its Fit Finder matched customers with bra recommendations without telling women their size.

But a few months in, ThirdLove scrapped that. Customers, who were so accustomed to shopping by size, hated it. “Women kept asking and needed to know. But you think you’re a 34D, and we tell you you’re a 32E in our Fit Finder — but they are the same cup size,” said company co-founder and co-CEO Heidi Zak. “It’s very confusing.”

What they were familiar with? Old-school quizzes. ThirdLove’s Fit Finder uses an online questionnaire to make recommendations based on style, fit challenges and breast shape, in addition to size. ThirdLove offers 70 sizes, including half-sizes, while the average bra brand, Zak said, has about 30 or 35. “It comes down to a more personalized service that is starting someone in the right direction, and not just by indicating size on the hangtag,” she said. “That really is the future.”

As concepts like data algorithms become more familiar to online shoppers, the idea of “fit” has expanded beyond a numerical size into more personalized notions of shape and fit intent. By harnessing a feedback loop from customers, largely through “quizzes,” brands are able to suggest a size, style and brand that is most likely to fit, regardless of the numerical label on the tag.

“The biggest thing is not just about having better sizes or more sizes, or sizes that are clear,” Zak said, “but about using data from real people to design product.”

Online personal styling service Stitch Fix also uses a quiz-based Style Profile. Male clients, for example, select from four different body shapes (slim, average, athletic or husky), and add details including fit challenges and “fit intent,” meaning how they like their clothes to fit.

MensStyleProfile-BodyShape-IllusA step in Stitch Fix Men’s Style Profile quiz

Because Stitch Fix offers a range of brands (most not designed in-house), it must make recommendations without the benefit of standardized sizing — a challenge that has vexed the fashion industry for decades.

But rather than creating a sort of internal universal sizing code across multiple brands, it “let the data speak for itself,” said Stitch Fix Men general manager Chris Phillips, meaning that it knows if one brand’s medium tends to run large and if another brand’s medium runs small, based on customer feedback. Additionally, Stitch Fix takes key points of measurement in every garment it carries in men’s and has begun doing this with its women’s business, Phillips said.

Stitch Fix clients report that fit is the most important attribute of the clothes they wear, and Phillips has found that often there was greater dissatisfaction among clients toward the edges of the fit scale, meaning those wearing an XS or an XXL.

“In the market, everyone is trying to build product based off the middle or the average, but they aren’t thinking about use-cases,” Phillips said. Thus, for Stitch Fix’s exclusive brands, the pattern makers began to take shape into account, rather than grading on a strictly linear scale. A men’s shirt in XXL, for example, might need more fabric in the mid-section or maybe a loosened arm, which takes shape into account more than is standard for the industry.

After making these changes, client success rates grew significantly, so Stitch Fix shared its learnings w­ith the brands it carries, many of which changed their fit specs based on the insights.

When it launched two years ago, Stitch Fix’s men’s business had 160 fit and size combinations; by this fall, it will have more than 800 fit and size combinations, including a new 28-inch inseam for men.

Stitch Fix has learned similar insights in its women’s business, which launched in 2011 and uses a similar Style Profile that takes body shape into account. It found that 25 percent of women who were shopping for plus were shorter than 5-foot-3, revealing a need for shorter pants in plus sizes, and that more than 15 percent of plus clients only needed a plus size on the top or the bottom. As many as 15 Stitch Fix partner brands are now making plus sizes for the first time in partnership with the company’s design team.

“These are examples of fit and size combinations that reduce the need for a standardized approach and move toward personalization,” Phillips said.

He added that anyone who has a favorite pair of jeans inherently knows that there is a deep correlation between “style” and “fit,” and that a shopper’s style ratings will improve when fit and style are in concert. In other words, if a customer mainly prefers dresses because they are more flattering, “let’s figure out why, and then how that might work for pants and tops,” Phillips said. “We’re not buying inventory to house in a store or put online, hoping someone will find the fit and size they love. We can buy based on different fit and size, because we know our clients. In [traditional] retail, they just know that people are shopping and have transactional data, but not why someone left the store.”

Retailers with a strong brick-and-mortar presence are taking note.

Since last year, Target has been working with Savitude, a software that uses a quiz-style approach to match a shopper’s body shape and proportion to a garment’s silhouette and design details (with 739 various possible results). The shopper begins by identifying as one of nine different body shapes and can then narrow down the search using additional visual cues, such as hem length or neckline, which essentially creates a personalized product assortment within the retailer’s online store.

Early in her career as a fashion designer, Savitude CEO Camilla Olson observed that when something didn’t look good or hang right, clients would say, “It doesn’t fit me.”

But Olson thinks that the fashion industry has a fit and language problem. “‘Fit’ has been over-used to represent a major reason for returns, but since only 20 percent of women have an hourglass figure, improved measurements can only solve a small percentage of the industry’s problem — that’s why we focus on shape and proportion.”

Early tests of Savitude showed an increase in engagement, with a 91 percent completion rate and a lift in sales. And by clueing in retailers to what clients need, the software stands to decrease unproductive inventory and increase profitability.

“When I first arrived at [retail accelerator] Target + TechStars, I was insulted, because they said, ‘You are the McDonald’s of fit tech — you are just good enough that people will do it and get results,” Olson said. “But I embraced it, because they are absolutely right. We don’t need high precision — the results do the job that is necessary, and we have tested it.”

And, in a nod to ThirdLove’s early days, Savitude also does not specifically call out what size or shape someone is. “We identify what you are, but don’t tell you because we don’t want to label our customers,” she said. “We just want her to be able to buy better clothes — no one needs another label.

Image via Stitch Fix