The biggest challenge companies producing made-to-measure fashion face may be their own customers, often tasked with the tricky-to-perfect and time-consuming chore of taking and submitting their own measurements.

“A lot of guys are uncomfortable with that. They’ve never touched a measuring tape in their lives and even if they do it, it would take 10 to 15 minutes,” said Jacob Wood, founder of Woodies, a direct-to-consumer startup selling customizable men’s shirts online.

A new approach to the problem turns to statistical mathematics and comes from New York-based startup Body Labs, which plans to reveal the first apparel retailers using its Body Kit technology, currently in private beta testing, later today. Now in use on the Web sites of online retailers Woodies, Original Stitch and Mesh01, the API platform appears to be a virtual fitting technology that creates an avatar based on either a body scan or a series of measurements entered by the customer. But Body Labs’ chief executive officer Bill O’Farrell argues the platform he and a team of research scientists from the Max Planck Institute for Intelligent Systems have created is fundamentally different.

“What we’re delivering is the body as a digital platform,” said O’Farrell, an adjunct professor at the Columbia Business School and co-founder of a number of startups, including one acquired by Adobe and responsible for developing its After Effects program.

Shoppers who visit Woodies can enter basic measurements like their height, weight and typical shirt size – information they’re likely to know offhand – and use the 3D BodyKit modeling tool to instantly calculate a rendering of their body alongside the corresponding measurements required to manufacture their desired shirt. But instead of relying solely on input data from customer-submitted measurements or body scans – both of which can be used with the platform depending on how a retailer chooses to integrate the technology into its Web site – Body Labs augments this information with data gathered from scans of tens of thousands of human bodies in thousands of different poses and positions. The result is, essentially, a fully articulated digital map of the body whose measurements are based on statistical calculations.

According to O’Farrell, “With each additional piece of information, we are able to do geometrically more calculations to make statistical predictions. If someone gives us six or eight or 10 inputs, we do an exceptional job of predicting body geometry, and can predict, for purposes of things like custom clothing, the full set of measurements needed to create a shirt or jacket or pair of trousers.”

Estimating sales up 15 percent over last month and a modest drop in returns and customer abandonment during the checkout path, Wood said he’s already seen an early, but noticeable, improvement in the two months his company has been testing the platform.

Both Wood and O’Farrell acknowledged that it’s still early days for the technology, with features like real-time viewing of digital garments on the 3D models still forthcoming. But Wood already sees additional possibilities, among them easier scalability into new clothing categories and new ways of thinking about garment sizes at a time when consumers are increasingly interested in mass customization.

“Based off the shirt measurements that we have, it outputs about eight measurements and with those measurements, we could build full bodies and the 16 or 20 measurements to build a suit or the nine measurements we need to build pants,” he said.

For Body Labs, made-to-measure apparel is a testing ground, one that O’Farrell said the company chose because “a lot of the apparel companies can take what we have and get a lot of mileage out of it really quickly.” But the platform is ultimately open for a variety of uses, including 3D modeling in video games, digital effects and fitness applications.

He said, “We are commercializing what we consider to be the world’s most sophisticated and comprehensive model of the human body.”

load comments
blog comments powered by Disqus