Why algorithms struggle to disrupt the fashion world

When I first met an American tech company in dress and heels, it seemed to me that I unwittingly wore flashy clothes. A cozy fleece uniform and a comfortable trainer are the default uniforms in Northern California. Next time I wore jeans.

But just because San Franceskan doesn’t dress doesn’t mean the tech industry isn’t interested in fashion. Mark Zuckerberg’s gray Brunello Cucinelli t-shirt and cashmere sweater worn by venture capitalists, though Normcore, are a unique fashion choice. San Francisco creates its own global fashion brand with Allbirds wool trainers.

But what the industry isn’t doing is somehow toppling the fashion industry itself. Ecommerce puts rocket boosters in vogue fast For sale, FinTech allowed fast online payments and social media provided specific advertising. But the business of choosing and designing clothing is almost always out of the control of the technology industry.

I don’t want to prove this: virtual reality eyewear maker Magic Leap has promised to revolutionize the way clothes are tried. Collapse under the weight of his own drum. Google, in collaboration with e-commerce company Zalando, tested a design based on artificial intelligence through Project Muze, but the results were not exciting. According to the technology blog, the design with which he appeared was “Sammy” Endgadget. AI-based custom fashion is still largely sleek.

The company that embodies the difficulty of applying data-based decision-making to human subjective preferences is StitchFix, an algorithm-based email clothing company. Founded ten years ago in the San Francisco shopping district, it was created when delivery boxes were in vogue. Vegetables, juices, socks and razors were sold in e-commerce with subscription.

Stitch Fix stands out because it focuses on data. In 2012, former Netflix data scientist Eric Colson was named CEO of Algorithm and was probably the first person to earn that title. The company has been armed with keywords such as network effects and proprietary algorithms. He promised to bring science to the art of fashion.

Fears that Amazon could overwhelm the business model meant a volatile start to stock prices. But for the benefit of Stitch Fix, Amazon has constantly struggled to sell fashion. Think of the AI-powered camera, Echo Look. This was supposed to provide style recommendations, but he was repeatedly criticized for being badly dressed. Even a new StyleSnap search tool allows users to upload photos and suggest similar clothing. High-tech press. The chosen volume can make your site look like a cluttered sale.

The failure of Amazon combined with the boom in online shopping of the pandemic era very good for stitch correction. There are more users who visit this place because they are reluctant to go to the store and are willing to find comfortable clothes in a trapped state. In the three months to May, sales rose 44% year-on-year. It tries to allow users to buy items directly.

As it grows, Stitch Fix seems to be excited to emphasize both the human element of the business and the magic of the algorithm. In August, Elizabeth Spalding was appointed new CEO. statement The stylist “plays a very active role in the formation of machine learning models with the data science team.” Note that the 2017 list document mentions the word algorithm. 76 times .. It was mentioned only once in a phone call with investors this summer. The number of human stylists has kept pace with the growth of users, doubling since 2017.

The problem with collecting multiple data points is knowing what to do with them. At the time of the deepest lockout, I ordered a stitch correction box and filled out a long online style contest about the stores and styles I like. Paying a $ 20 style fee, storing your favorite clothes, and returning the rest seemed like an efficient way to shop. But what came was an assortment of unfortunately weird clothes, mostly brands I had never heard of. I wanted Kate Moss in the office. I got a real minor at a garden party.

The taste is abstract and difficult to identify, regardless of the amount of data. The visible image of a floral shirt is not the same as the one you draw. Nor does it match the initial results shown by Amazon, Google, or StitchFix.

According to Stitch Fix, the more you ask, the better your experience will be, as the stylist and algorithm will get to know you better. Who knows, maybe the second or third box is much more appropriate. Again, there may be some things that the algorithm cannot do.


Why algorithms fight to disrupt the fashion world Link source Why algorithms fight to disrupt the fashion world

Source link

Leave a Comment

Your email address will not be published. Required fields are marked *