News

Google makes us wonder: can you solve fashion with math?

If you’ve visited the fashion-loving part of the internet at all in the past week, then you’ve undoubtedly heard of Google’s new fashion venture, Boutiques.com. As the name would indicate, the site’s intent is to gather fashion from around the web together in one easily shopable place that allows celebrities, designers, fashion people and anyone else who might have some spare time to create their own “boutique” that is tailored to their personal style and preferences.

Google is one of the most successful companies on the face of the planet, period, so I had high hopes that its nerds would come up with some sort of magic algorithm to recommend great things to me after I filled out the site’s length style questionnaire. But other than successfully banishing all brown from my results, the site’s recommendations were middling at best: my first round of options yielded items as seemingly random as a red silk Alexander McQueen gown, a pair of white tennis shoes and a $50 pleather purse. Not only are none of those options relevant to my shopping life, but the McQueen dress was the only thing I’d be caught anywhere near. Can Google really create an algorithm that replicates personal style?

I’ve always been of the belief that some people are fashion people and some people aren’t for a very simple reason: taste and style are intuitive and irrational. Somewhere deep in my right brain lies a section that starts to hum and whir and light up like Christmas when I see a piece of clothing that I love, and it has done so since I was in preschool. I struggle daily with the task of accurately articulating why it is that I like something, and it requires much more than the satisfaction of a few key characteristics for a piece to gain my favor. Or at least I think it does; Google’s betting that I’m wrong. I was more than willing to take the bet and set up an account.

The process of setting up a boutique and teaching Google’s software what you like is time-intensive, to say the least. Before the initial questions about your shape, pattern and color preferences, you’re taken through a series of pictures and asked to pick which outfits best reflect your personal style. Based on that, Google thinks I’m “edgy,” which apparently means I like black and Alexander Wang. I didn’t need a questionare to figure that out, but at least the website was accurate thus far. Although the time spent getting to that conclusion seemed unnecessary – I have the opportunity to sit around and diddle a stupid fashion website for hours because it’s ostensibly work-related, but how many people out there with real jobs (let alone kids to raise) want to spend their time teaching Google about their favorite brands, and then have those obvious conclusions parroted back at them like the information is somehow novel or revelatory?

The process’s second part involved telling Boutiques.com what shapes, colors and patterns I prefer. Despite my philosophical problems with choosing types of dresses that I “love” and “hate” (I love anything that’s designed correctly and tailored perfectly), I’ll admit that I have shapes toward which I tend to gravitate – tunics, skinny pants, booties, maxi dresses. The same goes for colors, and with as much information as I gave the website before it even showed me any clothes, I was hopeful that the picks would be good and that maybe I’d even find something I loved that I hadn’t seen before.

Soon, though, my hopes were dashed. As it turned out, telling Google that I liked black and high heels not only didn’t turn up a particularly good crop of shoes fitting that description, but it also turned up almost as many results that had seemingly nought to do with my stated preferences. And once the suggestions appear, you’re still not done. Next, you’re tasked with rating the items based on price, appearance and brand, which then helps tailor your future results to your taste. The only information you’re given on which to make your rating is a single picture and the item’s price. Most results also include a brand name, but enough lacked that key piece of information for it to be noticeable and annoying.

But even if you slog your way through all the choosing and voting and rating that has to be done for your results to improve (which they eventually do, although not as much as I would have liked), is this system really a superior way to shop than browsing categories at large online retailers like Net-a-Porter or Saks? The visual presentation and art direction aren’t even average when compared to most of the high-end shopping sites on the Internet, and instead of offering a well-edited collection of interesting, relevant pieces, Google wants you to do all the editing. And if you already have a strong sense of what you like and dislike, as the website requires in order to work correctly, then wouldn’t it be easier just to cut out the Boutiques.com middleman altogether?

With other luxury aggregators available to those who’d like to look for an item quickly or browse items across several platforms, I don’t see any particular advantage to using Boutiques.com over sites like ShopStyle, which provides a more straightforward approach to trawling the internet for particular types of products. Instead of trying to predict your taste, ShopStyle just lets you choose your parameters for any given search, which seems like a more logical way to tackle the ever-mercurial clothing tastes of the shopping public. The black shoe I love today might not have anything to do with the black shoe I want in a month, but if I’ve already trained Google to give results based on that first shoe, where does that leave my Boutiques.com account when I’m on to the next thing?

The specter that Google could eventually cast over fashion on the Internet is daunting to say the least, and when the company sets its mind to something, failure is not usually an option. What worries me, though, is that this entire venture might be based on a false premise. Many things about human behavior are predictable, but as far as tastes and trends go, fashion does a formidable job of keeping rationality at bay. But maybe that’s just the surface, and the things that we love really do fall on some sort of linear path. Do you think that Google’s engineers can accurately predict what color sweater you’ll want to buy next winter?

guest

19 Comments
Most Voted
Newest Oldest
Inline Feedbacks
View all comments