Back Fence

Taking the “I” Out of Identity, One Stitch Fix Box at a Time

Can personal taste be quantified?

By Kathryn Robinson August 18, 2016 Published in the September 2016 issue of Seattle Met

Algorithim method final 2 2  hpl7fa

Image: Adam Hancher

Is there a more accurate billboard for who we are than how we dress?

I wondered as I held up a boxy lace-inset white top I’d just unfolded from its neat tissue sheath, out of a Stitch Fix box that I’d plucked off my front porch with Christmas morning glee. Well…I wouldn’t have picked it out, I thought, pulling it on, but…wow, okay, now facing the mirror. I look like a refrigerator.

Not that this was a surprise, mind you. Spurred by the thousandth friend who had recommended the increasingly ubiquitous clothing subscription service as the busy person’s replacement for shopping, I had signed on with Stitch Fix a few months back. I’d filled out the questionnaire (sizes, style preferences, price ranges, which body parts I flaunt, which I blame on DNA) then waited each month for my personal stylist to select and send five items for me to try. A prepaid return bag minimized the stakes to a $20 style fee, which I decided was a small price to pay for a nudge out of my fashion rut.

The first month’s box had been a noble miss: shoes right for my age but not my style, a dress right for my style but not my midsection. I rejected another item as pricier than it looked; another as too close to something I had. Finally the layering piece—a color-splashed kimono-ish thing—which I artily flung over the dress and showed off to my teenage daughter.

“Okay, wow…no,” she said, trying for diplomacy but landing on horror. “They’re weird together!” With no salesperson in the room to arbitrate, I folded everything into the return bag.

We’re so sorry nothing worked! came Stitch Fix’s prompt reply, which, honestly, surprised me. They really thought I was expecting something to work? Already? With upward of a trillion variables of size, color, price tag, practicality, and personal style—why would anyone expect to be able to nail a stranger in five tries? Considering that a Seattle entrepreneur was recently able to raise nearly $2 million to fund the successful startup Fitcode—whose entire business model consists of the evidently onerous task of finding women jeans that fit them—it shocks me more, frankly, that Stitch Fix or its ilk ever works.

Because even if they get the fit right, there’s the considerably thornier business of getting, you know, the fit right. An individual’s brand identity. Her vibe. Sure, there’s art in it: Stylists working from their homes hand compose each box, in part from subjective cues like clients’ Pinterest pages. But they’re working off clothing lists generated by algorithms, the holy grail of marketing formulas, which predict what a consumer is likely to want based on click histories. Stitch Fix’s algorithm guy came from Netflix—Netflix, which knows what shows we’ve watched and which we’ve stopped watching, at precisely which minute—so all the cells of data about our own buying habits and others’ proven preferences combine to form a giant uber brain that generates, as one board member told a journalist, “a predictive heat score for every single item against every single user.” 

So much for the delightfully idiosyncratic fingerprint of my personal fashion statement. “Everyone thinks that their fashion is so unique—that it’s a very fuzzy, subjective thing,” Stitch Fix founder Katrina Lake told Forbes soon after launching. “In reality, the fundamental things are quantitative.” 

And you thought Amazon’s “You May Also Like” book suggestions felt demeaning. It’s bummer enough to be marketed to as an automaton, one who having liked The Light Between Oceans may also like The Orchardist. (For the record, Mr. Amazon, I bought The Light Between Oceans but I couldn’t finish it—and it was nothing like The Orchardist, which I loved.) Subscription services go further, resting on the deflating premise that the aesthetic fruit of our personality, that marker of self we call taste, can be quantified, even outsourced.   

But here’s the thing: It sort of can. In his recent book You May Also Like, Tom Vanderbilt goes on a hunt for the headwaters of personal taste and finds them shrouded in mystery. A person’s taste isn’t congenital, it isn’t consistent, it’s not even coherent. Indeed the only two factors he finds that predictably influence taste might have been lifted wholesale from the clothing subscription playbook: We’re drawn to things we are familiar with, and to things that others are drawn to.

Great. How much longer, I wondered, till popular things are the only things? Till no woman is wearing anything but nautical stripes and boyfriend jeans, and until the way we clothe ourselves—no longer a billboard for who we are—has become more like a predictive analytic of our demographic?

Something may have cracked in me. All I know is I heard the exhilarating thud of Stitch Fix No. 3 on my porch, I ran out and sliced open the box, and I pulled out two tanks, a maxi dress, a work blouse, and a silver bracelet. I tried on all five and turned in front of the mirror. All of them fit.

And none of them fit.

I folded all five into the return bag, now oh for three, a practical part of me disappointed. But a better part of me—the fuzzy, subjective part who’s still pretty sure that kimono went with that dress—well…what’s the opposite of disappointed?

Because she’s that.

Filed under
Show Comments