Let's write about how to predict fashion trends! What every fashionista wants to know!

I found this weirdly fascinating short fashion echo essay in The New Yorker, reported by Lauren Collins.

Maine Writer:  So, fashionista trends are reduced to dull data entries? Let me get this right?  If wearing sneakers on a fashion runway is determined to be trending down by 34 percent, this means the women who dare to wear ankle length cocktail dresses should not also wear sneakers at the same time?  I don't get it.  Haha.....I am of the age when the glamorous Elizabeth Taylor and Lauren Bacall were the data trendies.

Tagwalk Takes on the Hemline Index

The cerulean sleeve and the smoky eye have long been the province of whim, but Alexandra Van Houtte (the catwalk’s Bill James) is changing all that with fashion-data analysis.

Fashion is an emphatically subjective industry. “Pink is the navy blue of India” (Diana Vreeland). “There is a famine of beauty, honey” (André Leon Talley). “Sweatpants are a sign of defeat” (Karl Lagerfeld). “An evening dress that reveals a woman’s ankles while walking is the most disgusting thing I have ever seen” (Valentino). A whim or a feeling can seal the fate of a cerulean sweater, a smoky eye, or a mutton sleeve.

Yet, just as sabermetrics transformed baseball, data is coming for fashion, supplementing the hemline index—the theory that skirt lengths rise and fall with the stock market—with data lakes, traffic-share analyses, and lots of graphs. The other day, Alexandra Van Houtte, the Bill James of the catwalk, was sitting in a conference room in the Ninth Arrondissement of Paris. Her company, Tagwalk, is known as “Google for fashion.” But, instead of typing “weird rash” or “post office hours” into its free search engine, you can search every runway look by season, city, designer, color, model, and trend. “People are always scared of data, but data is inspiring to creativity,” Van Houtte said.

Van Houtte—in Chloé sneakers, soft black pants, and a long black coat embroidered with frangipani and chrysanthemums—founded the company in 2016. At the time, she was assisting a stylist, who was preparing for a shoot with Angelina Jolie and needed to see black dresses: every single one on the market. Van Houtte got to work, clicking through slide shows, screenshotting looks, compiling them into PDFs, and compressing the documents, then waking up the next day and doing it all over again. “It was massive torture,” she recalled. Now a buyer or a designer or an editor can enter “black dress” into Tagwalk and, within seconds, find seventeen thousand four hundred and ninety-seven of them, from a 2019 fish-net number by Dolce & Gabbana to a one-shouldered mini from Balenciaga’s recent show in L.A.

Tagwalk’s business model relies on two sources of revenue: smaller brands can pay to have their images appear on the site, while big brands buy trend reports that tell them who’s looking at what where. “I don’t care who’s cool, and it’s not my place to care,” Van Houtte said. The day’s order of business was to finalize a book—the first volume in a series—of takeaways from the Spring/Summer 2024 collections, which had concluded in Paris just a few days earlier. Van Houtte and a trio of analysts gathered around a table and opened their laptops.

“You have these other trend reports that are, like, ‘Orange is the color. We think leopard’s going to be a really big thing,’ ” Van Houtte said. “There’s this thing where data’s becoming cool, so sometimes people are just throwing numbers everywhere, like, ‘Data, but make it fashion!’ ” She pulled up an Instagram post, which claimed that a certain brand had enjoyed “140% more press” during Paris Fashion Week. “Compared to what? Compared to when?” Van Houtte said. “I really do not know how this was calculated.”

A draft of the book appeared on a wall-mounted screen. “O.K., let’s go!” she said.

A slide ranking the “20 Hottest Brands” according to Tagwalk users’ searches filled the screen. Miu Miu was No. 1, followed by Chanel, Gucci, Bottega Veneta, and Valentino. The team moved on to trends, which the site tracks using descriptive tags. Each look can have up to thirty, including americana, androgynous, babydoll, bantu knot, chapka, chelsea boots, dominatrix, earmuffs, equestrian, face jewels, gavroche cap, minaudiere, nomadic, oceanic, pastoral, polka dot, rave, round buckle, schoolgirl, spaghetti strap, taupe, visible underwear, and zig zag.

Van Houtte’s team cycled through a series of slides. glittery was surging (+783 per cent), as were ladylike (+235 per cent) and skirt (+30 per cent), while technical, animal print, bodycon, and pink were all trending down. Next came a pie graph (N.B.: it’s called a “Camembert” in French), which examined the colors of the season.

“There’s no more green!” she exclaimed. “It’s crazy.”

A two-page spread dealt with accessories, but she thought it needed more of a narrative.“If I was to buy this book, I’d be, like, ‘I spent this money, but you’re telling me that models carry bags and wear jewelry, which isn’t really groundbreaking information,’ ” Van Houtte said. They decided to go more granular. “O.K., a hundred and fifty per cent more bags were carried by hand?” she said. “That’s interesting.”

She had a hunch that she’d seen fewer sneakers on the runway, but she wanted to know whether the data bore it out. The team went searching in their spreadsheets, and a few minutes later returned with confirmation: sneakers were down thirty-two per cent.

Van Houtte looked at them excitedly. “Guys, this is like gold!” ♦

Published in the print edition of the January 1 & 8, 2024, issue, with the headline “Designer Data.”