Stitch Fix Uses Algorithms, Machine Learning To Dress Its Customers The next time you see a well-dressed individual walking down the street, stop to consider this: that spiffy outfit might not be the result of an impeccable fashion sense — it might be the work of a computer algorithm.
At least, that could be the case if the fashionista is a customer of one of several services that offer fashion delivered on demand, such as San Francisco-based startup Stitch Fix.
Using data analysis software and machine learning to match users with personalized clothing choices, Stitch Fix is ushering the fashion industry into the age of Big Data. For customers who don’t pry too closely into the startup’s inner workings, the service is intended to feel like magic.
“All they’re seeing is they order a box of clothes, and presto — it appears,” said Eric Colson, Stitch Fix’s chief algorithms officer. Companies in a variety of industries are relying more heavily on data to provide personalized recommendations — think Netflix using algorithms to find movies or TV shows users might like, or Amazon suggesting additional purchases based on what’s in someone’s cart.
Stitch Fix, which in September expanded into men’s fashion as part of its ongoing effort to revolutionize the clothing industry, uses that same technology to deliver curated boxes of clothing to customers’ doors. The first time users try the 5-year-old service, they answer a few dozens questions about their size, style and the body parts they like to flaunt. Stitch Fix takes that data and plugs it into its algorithms, which come up with a list of clothing options.
Then a human stylist reviews those choices, selects five items, and ships them. The customer has the option to buy the items or return them, free of charge. If the customer buys nothing, he or she pays a $20 styling fee per box. Customers can schedule regular deliveries or order one box at a time. When signing up for the service, they choose from a range of clothing price options — including “the cheaper, the better.” Stitch Fix’s software learns more about each customer every time he or she receives a shipment.
The company asks what the customer liked or disliked about each item — using natural language processing to decode their written answers — and applies that data to the next shipment. “I’ve seen the things that come in my box start to adapt to more what my personal style is,” said Kelly Walker, a music teacher at Willow Glen High School in San Jose who orders a Stitch Fix box every month. Walker has been using Stitch Fix regularly for about a year and a half, and said she looks forward to receiving her box of goodies every month. She doesn’t think much about the technology that drives the service, mostly because she and her human Stitch Fix stylist exchange personal notes on a regular basis.
During an interview on a recent Friday morning, Walker happened to be wearing a black and white top from Stitch Fix that she loves. But the company doesn’t always get it right. “One time definitely there was a sweater in there that I was like, ‘umm, this isn’t really my style,'” Walker said. “It was very loose and very baggy.” Stitch Fix learns a lot about a customer by analyzing what he or she returns, Colson said, especially because many people aren’t good at articulating what they want the first time around. “They may say they’re preppy, but it turns out they’re more of a classic or casual style,” he said. “People may think they’re a medium, but the medium — that “M” label — has a huge spectrum associated with it.