Jessie Zeng was scrolling through her Instagram feed in 2016 when
a post caught her eye. It was a photo of Gigi Hadid sporting a pair of
pearl-studded jeans at a Paris Fashion Week event. Below the image
were more than 50,000 comments, nearly all of them asking: Where
can I buy those jeans?
A self-described fashion obsessive who has been sharing her own
looks with her 30,000-plus Instagram followers since her undergrad
days at MIT, Zeng scoured the web for the answer, only to find it was
nowhere: The jeans had been custom-made for Hadid.
While trends used to be set twice a year in the pages of Vogue and on
Paris runways, now they sprout up daily from the Instagram feeds of
people like Hadid, the American model with 43. 4 million followers.
Retailers know fashion’s center of gravity has shifted, but they haven’t
been able to capitalize on it.
But Zeng, 26, believed social media—Instagram, in particular—
was powerful enough to alter that status quo. “For the first time ever,
there is an entire feedback loop existing on a platform where people tell
you exactly what they want to buy, and you can create it for them,” she
says. The consumer dollars are clearly there. Fashion and apparel
e-commerce sales in the U.S. are heading to $171 billion in 2022, says
eMarketer, up from $104 billion in 2018.
Choosy, the data-driven shopping platform Zeng launched earlier this
year, is her attempt to harness that loop. “It takes us about three days
from seeing it on a celebrity on Instagram to having a sample made up,”
says Zeng. “We do anywhere from 30 to 60 designs every month.” And
because everything is made to order, items are stitched and shipped to
consumers within two weeks of purchase, mitigating the inventory problem plaguing larger fast-fashion brands. It’s a hybrid of Stitch Fix, the
personal styling delivery service, and bespoke clothing, says Paula Rosen-blum, managing partner and retail technology analyst at RSR Research.
In figuring out how to speed up the clothing supply chain, Zeng
had the benefit of family expertise. Her uncle owns textile factories in
China, where she was raised. After an eight-month stint trading currencies fresh out of college, she moved near Beijing and spent two years
managing a few of her family’s factories, giving her an up-close look at
operations. “Without those years of experience, it would have been
completely impossible to do this,” she says. Choosy works with a network of about 200 small, agile textile factories, making it possible to
produce a diverse range of items in small quantities.
In figuring out what to make, she had the benefit of her friend
Sharon Qian, who was working on her PhD in applied math at Harvard.
Zeng persuaded Qian to join her startup, and Qian wrote algorithms to
analyze the comments under Instagram photos and rank the popularity of
specific items using natural language processing. This enables Choosy to
scan millions of comments for what Zeng calls “buying intent.” By coupling that intent with existing sales data, the machine learns the attributes
of items that sell well—in terms of celebrity connections, silhouettes,
colors, and styles—and assigns a relevancy score to rank future photos.
Today, Zeng works with 35 employees in a downtown New York City
office, as well as 15 people in China who focus on supply chain management. “Choosy really represents social commerce 2.0,” says Charlotte
Ross, an associate at New Enterprise Associates, which raised Choosy’s
seed round of $5.4 million. “Customers are the ones deciding what gets
made. It’s the way people will ultimately want to shop.” —JANE PORTER