Many Unhappy Returns

Can Amazon figure out how to sell clothes that fit?

Amazon CEO Jeff Bezos
Can Jeff Bezos figure out how to sell you pants that fit?

Photograph by Emmanuel Dunand/AFP/Getty Images.

Sometimes in the middle of the night, when the toddler wakes me up and I can’t get back to sleep, I go online to look for pants. Pants are my fashion Waterloo. I’ve got a slight waist and short legs, a dimensional no-pants-land that’s ill-served by most retail stores. (This is one of the reasons I like wearing pajamas.) I’ve come across a few expensive brands of jeans, chinos, and dress pants that fit me somewhat well, and my strategy has been to buy up as many of them as possible. But this isn’t easy—the stores never have more than one pair in my size—and it’s sartorially boring, not to mention expensive. That’s why, in my weaker sleepless moments, I can’t resist the lure of shopping online. From afar, the Web looks like a trouser treasure trove—there are so many styles, they’re all available in every size, and they’re often cheaper than in stores. What could go wrong?

But then my pants arrive and they’re a disaster. Even though they’re in my size, they don’t really fit. Depending on the brand and the style, a given waist and inseam length fits differently from a supposedly identically sized pair in another brand. Bad fit is a common problem in online apparel sales, perhaps the main reason that shopping for clothes online hasn’t taken off in the same way that shopping for everything else has. Most successful online apparel retailers suffer excruciatingly high rates of returned merchandise. Customers return about 35 percent of the shoes purchased from Zappos, for instance, and the store’s best customers (those who spend the most money) return half of what they buy. This suggests that a lot of people are buying stuff to try it on and send it back. Some of the returns are likely a matter of taste—those sandals don’t look as good on your feet as they did on your laptop—but many retailers say poor fit is their biggest headache.

Over the last few months, has pushed aggressively into the apparel game. The company—which also owns Zappos and the flash-sales clothing discount site MyHabit—has signed on hundreds of designer brands, and it has invested heavily in improving how it lists clothes on its site. Among other things, according to the New York Times, the firm is shooting 3,000 fashion photos a day to post online, and it has hired three women to try on and review shoes full time. As an Amazon devotee, I was happy to hear about all this. If Amazon can make it as easy to buy clothes as it has everything else, I might finally be able stop searching the world for nice pants.

But then, at 3 a.m. the other night, I logged on to Amazon and started searching for pants. Things didn’t go well. I loved the wide selection at Amazon’s clothing store, and I found several pairs of jeans and khakis that I liked. Reviews on the items, though, suggested that they weren’t “true to size,” and I’d have to choose a bigger or smaller size to get a good fit. But how much bigger or smaller? It was a guessing game. When the pants arrived—in a stylish white envelope rather than Amazon’s standard smiley-face cardboard box—it was clear that I’d guessed wrong. The pants were all either just too tight or loose.

If Amazon really wants to revolutionize online clothing sales, it would do well to come up with a fix for the size problem. When I buy a pair of pants or any other clothes online I’d like to know that they’ll fit. If Amazon can give me that certainty, it will make a killing in clothes.

But how can it do so? The fit problem has always plagued online apparel retailers, especially when it comes to women’s clothes—a far bigger market than men’s apparel, and one in which sizing is even more treacherous. Because most women’s clothes carry numerical rather than measured sizes—that is, women’s jeans are listed as a size 10, while men’s are labeled with a 30-inch waist—it’s more difficult for a woman to figure out her size in a specific brand. Numeric sizing also allows for wider variations in sizes between brands, and changes over time, too. There’s a widespread assumption that fitting problems are caused by “vanity sizing,” a phenomenon in which clothing manufacturers steadily, intentionally inflate sizes in order to make women feel better about their clothes. (“Hey, this size 2 dress fits me! I should buy it!”) But many clothing size experts say vanity sizing is a myth. Sizes change over time because of the changes in the way manufacturers make clothes, and also because people have changed over time—we’ve become bigger, so our clothes have, too.

A bunch of firms have come up with ideas for bringing sanity back to clothing sizes. At the peak of the last tech boom in 2000, I interviewed a startup that wanted everyone to get body-scanned—you’d walk into a booth at a mall, get prodded with radio waves, and then emerge with an electronic receipt of your exact dimensions. The dimensions could then be fed to online retailers, who’d be able to figure out which of their clothes would fit your profile. The company that I spoke to seems to have disappeared, but a firm called Me-Ality is still pushing the scanning idea. It has set up dozens of scanning booths across the country, but their utility is limited, since few sites list clothes by your scanned size. (You can visit the company’s site for a shopping guide based on your measurements.)

Another startup, back then, wanted to use your old clothes to find new ones. You’d tell the company which brands and styles fit you perfectly, and it would tell you what sizes worked for you in other brands. That idea lives on in a firm called True Fit, whose technology Macy’s began using on its site last year. After you tell the site what’s in your closet and give it some basic information about your body type, it compares your size to its data on thousands of other clothes. Then, when you browse the Macy’s site, you’re given a score suggesting how well each item will fit based on your other clothes. The system keeps track of what you buy—and what you return—over time, so it constantly improves its predictions. The Wall Street Journal reported that in a test with 400,000 users on one premium-denim sales site, True Fit reduced returns from 50 percent to 20 percent.

Then there’s the most basic solution to the sizing problem: Get out a tape measure and take your own measurements. Retailers have shied away from this method because customers find it annoying (and there’s some worry that people will lie about their own sizes), but I’ve found it works really well. In the last year I purchased several shirts and a few pants from several online stores that make custom-fit clothing for men. Sure, it was a hassle to precisely survey my body, but you’ve only got to do it once, and it was worth it. Getting clothes that fit perfectly is surprisingly thrilling; it makes you feel so good that you can’t stop yourself from buying more.

That’s why Amazon would be wise to implement one of these ideas. Body-scanning is pretty wacky, but technology like True Fit’s—or even the trusty old tape measure—seems like it could work wonders for the site. If Amazon cataloged its clothing by size, and if it used your measurements to predict apparel that would fit you, its sprawling clothing store would be pretty hard to resist—the promised land of perfect pants that we’ve all been searching for.