Why We Buy

by Paco Underhill

  Whyte, who started his career as an editor at Fortune magazine, was, essentially, a scientist of the street—the first one, which is amazing when you think of how long streets existed before he came along. His work has been used to make public spaces better and more useful to citizens, which in turn made cities better and more useful, too. Whyte’s methods were a kind of lens through which a physical environment could be studied and improved, and my work on behalf of shopping owes a great deal to his methods.

  Back in 1977, I was a part-time instructor at City University of New York, teaching courses in fieldwork techniques for the environmental psychology department. I was also working in an establishment of which I was part owner, the Ear Inn, a bar in downtown Manhattan. There, I had a customer who had been hired to design a system of signage at Lincoln Center, the performing arts complex that’s home to the Metropolitan Opera House, Avery Fisher Hall—about a dozen theaters in all. He told me they needed someone to look into the usage and circulation patterns of the underground concourse that connected the buildings to parking garages and the subway. There was a small, makeshift gift shop down there at the time, but Lincoln Center wanted to see if a larger store might be viable there. First, though, they needed to make sure that a store wouldn’t create congestion in the pedestrian walkways. With my customer’s help, I got the job.

  So I recruited a few of my students to help and we took some cameras, staked out our observation spots and went to work counting and mapping. The crowding question was easy enough to answer—we roped off an area exactly the size of the store they wanted to build, then watched and filmed pedestrians streaming through during the busiest times. We suggested then that with the room available, they should add some benches down there, to make it something of a destination rather than just a corridor. Our client declined to take our advice then, but today the benches are in place. I also strongly recommended that they double the size of the ladies’ room, and they declined to take that advice, too. Today, thirty years later, the line at the ladies’ room still goes out the door during busy times. Shameful.

  As I was compiling the data to write the report and looking at the many hours of film I had shot, I realized that from one of the camera positions I could see inside the gift shop, all the way to the cash register. There, as I watched, two customers lined up to pay. One looked to be a wealthy woman, probably an opera-goer, who had piled a small tower of boxes on the counter. Next to her was a teenage girl whose purchase required just one small brown paper bag. I couldn’t see enough to tell exactly what was going on, but I was intrigued.

  I visited the shop the next day and talked to the clerk, who told me that the woman was the wife of a Mexican diplomat who had decided to buy some fancy music boxes as gifts to take home with her. The boxes were expensive, and she was buying about a dozen of them, for a total sale of close to $9,000. She needed to pay quickly, before intermission ended, and she had to arrange to have the boxes delivered to her. There was also the matter of having sales tax waived owing to her diplomatic status. A complicated transaction, to say the least.

  But this had to wait while the clerk handled the transaction with the teenage girl, who had arrived at the register first bearing her selection—a ballerina pen.

  It was clear even to an academic like me that the cash register procedure could stand a little reorganization and clarification. These two transactions should not be competing for the same clerk’s attention. And then the lightbulb clicked on. Why not take the tools of urban anthropology and use them to study how people interact with the retail environment?

  Years earlier I had witnessed an argument between the esteemed sociologist and author Erving Goffman and Jack Fruin, the chief engineer of the Port Authority of New York and New Jersey, who was at that moment in the midst of a gigantic undertaking, the planning and construction of Newark International Airport. Jack was expressing his emphatic frustration with the world of academia; he had attempted to hire some scholar-experts to guide his engineers and architects in their work, but instead of the clear-cut advice he had hoped to receive, he was getting buried under the academics’ typical inability to assert any fact, no matter how small, that hadn’t been completely proven by research. Goffman held the intellectual high ground in their argument, but at one point I clearly remember thinking, I’d have a lot more fun working for Jack than for Erving. Erving’s hiding in his ivory tower. Jack is out there doing stuff.

  Not long after the Lincoln Center assignment, I was sitting with some friends at a nightclub in Greenwich Village. One of the guys at our table was a young executive with Epic Records, a division of CBS, and I described to him my bright idea of measuring what happens in stores—the thought that there might be something worth learning by turning scientific tools on shopping. And over the course of a few beers my idea must have sounded interesting, because the guy said, “Why don’t you send me a proposal?”

  Full of ambition the next morning, I rose early, dragged out my manual typewriter and drafted a plan. I sent it over quickly, then waited. For, oh, about a year. Of course I tried writing again and telephoning during that time, but no one ever returned my calls. These were the dark ages of the science of shopping, remember.

  And then, out of the blue, I heard from a woman who was in charge of market research for CBS Records. She said that they had found my proposal in a dusty file somewhere and were all quite fascinated by it, and was I still interested in studying a record store?

  Sure, I said, inwardly rejoicing that a major American corporation was actually going to underwrite—to the tune, I think, of about $5,000—my research into the habits of the modern shopper. I immediately called a few of my students, assembled some notebooks and time-lapse cameras, and made my way to a record store in a northern New Jersey mall.

  Now, nearly decades and close to two million hours of videotape and much personal observation later, that study seems almost charmingly rudimentary. But at the time, it felt as though the discoveries came flying fast and furious.

  For instance: In the late ’70s, when the study was being done, traditional singles—45 rpm records—were still big sellers. The store, wisely, displayed the Billboard magazine chart of bestselling singles near the racks of records, as a stimulus to sales. But our film showed that most buyers of 45s were adolescents—and the chart was hung so high on the wall that the kids had to stand on their toes and crane their necks to see what exactly was at the top of the chart. We suggested to the manager that the chart be lowered, and a week later he called to say that sales of 45s had gone up by 20 percent. Just like that! Lower the chart! It worked!

  We spent a lot of time that weekend watching people in line to pay at what the retail industry calls cash/wraps. Regardless of what store designers and merchandise managers think, in many ways the cash/ wrap area is the most important part of any store. If the transactions aren’t crisp, if the organization isn’t clear at a glance, shoppers get frustrated or turned off. Many times they won’t even enter a store if the line to pay looks long or chaotic.

  At this store, there were several big displays of new releases as soon as you walked in, just a few feet from the cashier. This was fine as long as the store was empty, but if customers were in line, their bodies completely hid the displays. Put up a stanchion and a velvet rope to keep the line off to one side, we suggested, and again, our advice had an instant effect—sales of records from the displays went up immediately.

  Doesn’t all this sound just the least bit obvious? It does to us, too, especially after we’ve spent so much time watching and filming and timing and interviewing and so on. Until then, however, these were the kinds of problems that had remained hidden in plain view.

  While watching the record store customers, we noticed an odd pattern: The LP section (this was pre-CD, remember) was always more crowded than cassettes, but sales were split evenly between the two formats. Following customers, the reason became clear: Because the LP covers were bigger, it was easier to read the song lists and s
ee the photos, so cassette shoppers would browse in LPs, make up their minds, and then go to the tapes section to find their choices. Our suggestion was to make the aisles wider in LPs so that shoppers wouldn’t feel crushed and rushed, a definite sales killer. Also, we thought the store should invest in more durable carpet for the sections that got significantly more traffic.

  My final memory from that study comes from a video clip I still show to audiences: a young man shoplifting classical music tapes. Only after watching him take the tapes over and over on the film did I notice that the bag he slipped them into was from a chain that had no location at that mall. I passed this tidbit on to the client’s security executive and told him that they should be watchful whenever such “wrong” bags were spotted in their stores (remember, this was before security tagging). I got back a note saying that they had prevented several thousands of dollars in theft using that method of detection.

  And thusly, a science was born.

  Before the science of shopping existed, there were at least two other ways to measure what took place in a store. The most common way of viewing a store is to simply examine “the tape”—the information that comes from the cash registers, which tells what was bought, when and how much of it. This is how virtually every retail undertaking, from the largest, most sophisticated multinational chain to the corner newsstand, does it. It’s a fine way to see how the store as a whole has performed this quarter, or this year, or on any given day, or even time of day, and is, in the end, the measure of a store’s overall health and growth (or decline) that counts. But as a diagnostic tool, or as a way of figuring out what happens in the store and how, it is not very useful. Sales research records your victories; what it does not do is look at where you are losing. What hurts is when you get the shopper in the door, down the aisle and in front of the product, and for whatever reason, they don’t buy. When businesspeople attempt to infer too much from sales data, it can be downright misleading. Here’s a good example, from a chain drugstore in a Massachusetts mall. This was the first mall store owned by this particular company, and so management was eager to see the results. Based solely on total sales, our client was pleased overall, and in particular with how the aspirin section of the store was performing.

  But based on all our many previous studies both of drugstores and of the aspirin category, one crucial figure was on the low side. The product conversion rate—the percentage of shoppers who bought—was below what we expected. In other words, plenty of customers stopped at the aspirin section and picked up and read the packages, but too few of them actually bought aspirin. And the conversion rate for aspirin is usually high—it’s not the kind of product you idly browse; you tend to go to that aisle only when you’re in need. So we spent some time specifically watching the aspirin shelves, and we trained a video camera on them, too.

  Over the course of three days, a pattern emerged. The aspirin was displayed on a main aisle of the store, on the path to some refrigerated cases of soft drinks, which tended to draw a great many customers to that part of the store. That might lead one to expect that the aspirin would sell well, but just the opposite happened. The main customers for cold drinks were teenagers, and our observation showed many of them entering and making a beeline for the coolers. In fact, this was a favorite place for the mall’s young employees to grab a quick cold soda during breaks.

  These young shoppers were supremely uninterested in aspirin. The shoppers, often seniors, who did want aspirin stood a little nervously at the shelves, searching for their usual brand or figuring out which was the better deal while also trying to stay clear of the teenagers tearing down the aisle. In fact, a substantial number of aspirin shoppers became so irritated or thrown off balance by the teenagers that they would prematurely break off their browsing and walk away empty-handed. It was a modified version of the butt-brush effect—the shoppers weren’t being jostled exactly, just a little rattled. You could see it plainly on the videotape—some customers were practically cringing and hugging the shelves, not the ideal shopping position. And when we timed shoppers, we found that they were spending less time at the shelves than our experience led us to expect.

  This is something that comes up in our work all the time: A store has more than one constituency, and it must therefore perform several functions, all from the same premises. Sometimes those functions coexist in perfect harmony, but other times—especially in stores selling diverse goods, like cold drinks and medicines—those functions clash. We also saw this in a Harley-Davidson dealership, where a roughly three-thousand-square-foot showroom has to make room for well-off male menopause victims looking to recover their virility by buying bikes, blue-collar gearheads who are there for spare parts, and teenage dreamers interested in the Harley-logo fashions. All three groups want nothing to do with one another. When a premises’ functions clash, a way must be found to accommodate as many uses as possible. In this drugstore, we advised our client about what we had learned and suggested a counterintuitive move—that the aspirin be relocated to someplace off the main drag. Fewer total customers would come upon it, we knew, but more aspirin would be sold. When they moved the shelves, sales rose by 20 percent.

  We performed research for a large bookstore that had recently put a big table of discounted books just inside the entrance, where every customer would see it first thing. And it performed admirably—almost everyone stopped for at least a cursory browse, and the percentage that bought at least one book was high. Which meant that, according to the cash register tape, the table was a resounding success.

  Except that as we tracked shoppers, we found that the number who would go to the table and then travel through the rest of the store was lower than it should have been. In a case like this, every hour on the hour a tracker would hurry through the entire store and note how many shoppers were in each section, including the register area, the coffee shop and so on. This is the density check that we perform as part of every store study, and it tells us a great deal: It gives an instant snapshot of the store’s population and where people are drawn or not; it suggests when something about the architecture or the layout may be inhibiting shoppers from visiting certain areas; and it shows how shoppers move (or fail to) through the premises. And in fact, taken section by section, the number of shoppers who were penetrating the rest of the store was uniformly down. Also, our track sheet maps of customer travels began showing a telltale shallow loop—shoppers would enter, hit the bargain table, then maybe visit one or two more displays, but they never strayed far from the front of the store before heading to the cashier. This was no coincidence, needless to say—customers were choosing from the discount table, then going directly to the register, paying for their bargains and leaving without even browsing the bestsellers or any of the other books selling at the normal profit margins. Our shopper interviews turned up an unfortunate side effect, too: Thanks to the prominence of the bargain table, the store was gaining a reputation as a discounter rather than as the place to go for the hot new book. The success of the table was causing the failure of the rest of the store.

  So much for what can be learned from the register tape.

  The second means of learning what goes on in a store, employed by the most famous names in market research (whether political, commercial or any other) is simply to ask people questions about what they just saw, or did, or considered doing. That can happen in person, online, on the phone or in a focus group—it’s all about asking people what they think.

  Let’s take a telephone poll conducted by the Democrats and the Republicans, for instance, or just the shopper interviews that take place as you exit a store or a shopping center. After a long list of questions, some basic demographic information is taken (age, education, income, sex, race and so on). From those two, a big fat binder full of suppositions is assembled: Forty-year-old Caucasian college-educated married mothers of two living in Northeastern suburbs and driving station wagons would prefer Jif even more if it were low-fat, for example. Or men who buy Co
ke at convenience stores say they would notice their brand less often if it were any color but red. Or one quarter of all college graduates eats pasta once a week. The possibilities for cross-referencing are endless, and there is undoubtedly some marketing wisdom to be gotten from such studies. But they don’t really reveal much about what happens in a store, when shoppers and goods finally come together under the same roof. There are surveys that do ask customers for information about what they saw and did inside a store, but the answers are often suspect. Sometimes people just don’t remember every little thing they saw or did in a store—they weren’t shopping with the thought that they’d have to recall it all later. In a fragrance study we performed, some shoppers interviewed said they had given serious consideration to buying brands that the store didn’t carry. In a study of tobacco merchandising in a convenience store, shoppers remembered seeing signs for Marlboro even though no such signs were in that store.

  If we went into stores only when we needed to buy something, and if once there we bought only what we needed, the economy would collapse—boom.

  Fortunately, the economic party that started the second half of the twentieth century has fostered more shopping than anyone would have predicted, more shopping than has ever taken place anywhere at any time. You almost have to make an effort to avoid shopping today. Stay out of stores and museums and theme restaurants and you still are face-to-face with Internet shopping twenty-four hours a day, seven days a week, along with its low-rent cousin, home shopping on TV. You have to steer clear of your own mailbox, too, if you’re going to duck all those catalogs.