The Science of Shopping : LUSENET : Unk's Troll-free Private Saloon : One Thread


Comfortable shoes, the American commercial camouflage uniform -- khaki pants, olive polo shirt, no aftershave and good, thick, dun-colored socks.

Okay, stroll, stroll, stroll...stop.

Get out the clipboard and pen.

Shhh. Stay behind that potted palm. This is the first track of the day.

The subject of study is the fortyish woman in the tan trench coat and blue skirt. She's in the bath section. She's touching towels. Mark this down -- she's petted one, two, three, four of them so far. She just checked the price tag on one. Mark that down, too. Careful, her head's coming up -- blend into the aisle. She's picking up two towels from the tabletop display and is leaving the section with them. Get the time. Now, tail her into the aisle and on to her next stop.

It's another day of fieldwork; the laboratory, another troubled department store. The focus of our analysis is the domestic department as per the science of shopping. But let's start by addressing a fundamental question: Since when does such a scholarly discipline even exist?

Well, if, say, anthropology had devoted a branch to the study of modern shoppers in situ, a fancy Latin way of saying shoppers out shopping, interacting with retail environments (not only stores, but also banks and restaurants), including but not limited to every rack, shelf, counter and table display of merchandise, every sign, banner, brochure, directional aid and computerized interactive informational fixture, the entrances and exits, the windows and walls, the elevators and escalators and stairs and ramps, the cashier lines and teller lines and counter lines and restroom lines, and every inch of every aisle -- in short, every nook and cranny from the farthest reach of the parking lot to the deepest penetration of the store itself -- that would be the start of the science of shopping. And if anthropology had already been studying all that...and not simply studying the store, but what, exactly human beings do in it, where they go and don't go, and by what path they go there; what they see and fail to see, or read and decline to read; and how they deal with the objects they come upon, how they shop, you might say -- the precise anatomical mechanics and behavioral psychology of how they pull a sweater from a rack to examine it, or read a box of heartburn pills or a fast-food restaurant menu, or deploy a shopping basket, or react to the sight of a line at the ATMs...again, as I say, if anthropology had been paying attention, and not just paying attention but then collating, digesting, tabulating and cross-referencing every little bit of data, from the extremely broad (How many people enter this store on a typical Saturday morning broken down by age, sex and size of shopper group?) to the extremely narrow (Do more male supermarket shoppers under thirty-five who read the nutritional information on the side panel of a cereal box actually buy the cereal compared to those who just look at the picture on the front?), well, then we wouldn't have had to try to invent the science of shopping.

But anthropology didn't pay attention to those details, and so down the hall from my office is a room containing around fifty cameras, mostly video but with some still and digital cameras and a couple of old-fashioned Super 8 time-lapse film cameras thrown in. Next to them are piled cases of blank 8mm videotapes, two hours per tape, five hundred tapes to a case. We go through about fourteen cases, seven thousand tapes, a year. (In 1992, when we shot a lot of time-lapse Super 8 film -- about $60,000 worth -- Kodak told us we were the single largest consumer of Super 8 film in the world.) We also have maybe a dozen handheld computers on which we take down the answers from the thousands of shopper interviews we conduct, and there are some odd laptops in there, too, plus all manner of tripods, mounts, lenses and other camera accessories, including lots of duct tape. Oh, and hardshell cases for everything, because it all travels. A lot. We have enough gear in that room to equip a major university's school of social anthropology or experimental psychology, assuming the university has a deserved reputation for generating tons of original research gathered all over the globe.

Despite all that high-tech equipment, though, our most important research tool is a low-tech piece of paper we call the track sheet, in the hands of the individuals we call trackers. Trackers are the field researchers of the science of shopping, the scholars of shopping, or, more precisely, of shoppers. Essentially, trackers stealthily make their way through stores following shoppers and noting everything they do. Usually, a tracker begins by loitering inconspicuously near a store's entrance, waiting for a shopper to enter, at which point the "track" starts. The tracker will stick with the unsuspecting individual (or individuals) as long as the shopper is in the store (excluding trips to the dressing room or the restroom) and will record on the track sheet virtually everything he or she does. Sometimes, when the store is large, trackers work in teams in order to be less intrusive.

Befitting a science that has grown up in the real world, meaning far from the ivory towers of academia, our trackers are not an taken from the usual researcher mold. In the beginning we hired graduate environmental psychology students, but we found they were sometimes unsuited to the work and tended to come to the job burdened with textbook theories they wanted to apply. As a result, they often didn't possess the patience necessary to simply watch what shoppers do. The other problem we had with grad students involved stamina: While we don't work in the dusty heat of Mesopotamia, twelve hours on your feet under the fluorescent lights at Kmart is no picnic either. Fieldwork in any physical or social science is difficult. We found that, for our purposes, smart, creative people -- artists, actors, writers, a puppeteer -- often have what it takes. Beyond the fact that they have no theories to uphold or demolish, their professional skills are often rooted in their ability to observe. Also, it does not hurt that they have flexible schedules, so that when that Brazilian brewer or Australian tampon manufacturer or American fast-food operator happens to call, they have the open calendar and curiosity to be willing to go take a look.

When we find someone we think has the temperament and the intelligence for this work, we first put him or her through a training session. There's a lot to learn -- how do I watch and simultaneously take notes, for instance, or how can I tell whether someone is reading a sign or just staring at the mirror next to it? We have to teach the most important tracker skill of all: How do I stand close enough to study someone without being noticed? It's crucial to our work that shoppers don't realize they're being observed. There's no other way to be sure that we're seeing natural behavior. Fact is, we're all still surprised by how close you can stand to someone in a store and still remain invisible. We find that positioning yourself behind the shopper is a bad idea -- we all know the sensation that we're being watched. But if you stand to the side of a shopper, his or her peripheral vision "reads" you as just another customer -- harmless, in other words, and barely worth noticing. From that position you can get close enough to see exactly what a shopper is doing. You can be sure that he's touched, say, nine golf gloves, not eight or ten. Then we throw the tracker hopefuls out into the real world, in a store setting, to see them in action. Most of them wash out at this point -- you can teach technique, but not the intelligence or the slight case of fascination required to do this work well. Over half of our core group of thirty U.S. trackers have been with us for more than five years, some for a decade or more. It's hard work, but addictive, too. in teams of three to ten people, led by a member of our staff, they crisscross the United States and Canada, as well as Europe, South America and Australia, visiting every kind of retail business imaginable, from banks to fast-food restaurants to high-fashion boutiques to hangar-size discounters and everything in between. To make our international work easier and more efficient, for three years we have had research teams based out of Milan, Italy, and for two years out of Sydney, Australia.

In addition to measuring and counting every significant motion of a shopping trip, the trackers must also contribute incisive field notes describing the nuances of customer behavior, making intelligent inferences based on what they've observed. These notes add up to yet another, this time anecdotal, layer of information about a given environment and how people use it.

The forms our trackers use have evolved over the two decades we've been doing this research. They are the key to the entire enterprise, an achievement in the art of information storage and retrieval, nondigital division. Our earliest track sheets could record maybe ten different variables of shopper behavior. Today we're up to around forty. The form is reinvented for every research project we undertake, but typically it starts with a detailed map depicting the premises we're about to study, whether it's a store, a bank branch, a parking lot (for a drive-thru project) or just a single section or even just one aisle of a store. The map shows every doorway and aisle, every display, every shelf and rack and table and counter. Also on the form is space for information about the shopper (sex, race, estimates of age, description of attire) and what he or she does in the store. Using the system of shorthand notation, a combination of symbols, letters and hash marks, a tracker can record, for instance, that a bald, bearded man in a red sweater and blue jeans entered a department store on a Saturday at 11:07 a.m., walked directly to a first-floor display of wallets, picked up or otherwise touched a total of twelve of them, checked the price tag on four, then chose one, moved at 11:16 to a nearby tie rack, stroked seven ties, read the contents tags on all seven, read the price on two, then bought none and went directly to the cashier to pay. Oh, wait, he paused for a moment at a mannequin and examined the price tag on the jacket it wore. We'd mark that down, too, just as we'd note that he entered the cashier line at 11:23 as the third person in line, waited two minutes and fifty-one seconds to get to the register, paid with a credit card and exited the store at 11:30. Depending on the size of the store and the length of the typical shopper's stay, a tracker can study up to fifty shoppers a day. Usually we'll have several trackers at a site, and a single project may involve the simultaneous study of three or four locations in separate cities over a series of different weekends.

By the end of a job, an incredible amount of information has been crammed onto those sheets. They come back to the office where the job captain spends a day "cleaning" the forms -- making sure that each hash mark is visible and that every box that should be filled out has been. Then our data department spends another day or so entering all the information, every single notation on every track sheet, into a data base.

Over the years we've spent tens of thousands of dollars and countless frustrating hours with computer programmers, trying to come up with a data base system that could handle the kind of work we do. The big problem is that while we crunch the same numbers in the same ways from job to job, each project usually requires us to do something a little differently -- to collect different kinds of data, or to devise new comparisons of facts we've uncovered. We've hired fancy consultants who've spent six months at a crack with us, trying to build us a computer system. They ask us to list everything we want our program to do, but every week we add six new things to the list that negate all their work from the previous month. And, of course, our turnaround time must be swift, so there's no time to change the system completely for each job -- we may need to do one new comparison for a project today and then not have to perform that function again for seven months.

Until recently, most of our work was done in Microsoft Excel. Excel is not a data base program but a spreadsheet program, intended to help accountants do relatively simple flat calculations. Excel's beauty is its open architecture -- you can get in there under the hood and tinker, and soup it up. And that's exactly what we've done. It's as though Microsoft built a very nice bicycle ten years ago and we've turned it into a databusting all-terrain vehicle. Today we run much of our work in FileMaker and SPSS, but still vet it in Excel.

When the videotapes come back from the sites, it's someone else's job to screen every foot. Depending on the size of the store, we may have ten cameras running eight hours a day trained on specific areas -- a doorway, for example, or a particular shelf of products. We videotape around twenty thousand hours' worth of store time annually. The video produces even more hard data; if, for example, a study is meant to determine in part how a particular cash register design affects worker fatigue, we may use the video and a stopwatch to time how long it takes for a clerk to ring up a sale at 10 a.m. compared to 4 p.m.

The list of particulars we're capable of studying -- what we call the deliverables -- grows with every new project we take on. At last count, we've measured dose to nine hundred different aspects of shopper-store interaction. As a result of all that, we know quite a few facts about how human beings behave in stores. We can tell you how many males who take jeans into the fitting room will buy them compared to how many females will (65 percent to 25 percent). We can tell you how many people in a corporate cafeteria read the nutritional information on a bag of corn chips before buying (18 percent) compared to those lunching at a local sandwich shop (2 percent). Or how many browsers buy computers on a Saturday before noon (4 percent) as opposed to after 5 p.m. (21 percent). Or how many shoppers in a mall housewares store use shopping baskets (8 percent), and how many of those who take baskets actually buy something (75 percent) compared to those who buy without using baskets (34 percent). And then, of course, we draw on all we've learned in the past to suggest ways of increasing the number of shoppers who take baskets, for the science of shopping is, if it is anything, a highly practical discipline concerned with using research, comparison and analysis to make stores and products more amenable to shoppers.

Because this science has been invented as we have gone along, it's a living, breathing field of study. We never quite know what we'll find until we find it, and even then we sometimes have to stop to figure out what it is we've seen.

For example, we discovered a phenomenon known as the butt-brush effect almost accidentally. As part of an early study for Bloomingdale's in New York City, we trained a camera on one of the main ground-floor entrances, and the lens just happened to also take in a rack of neckties positioned near the entrance, on a main aisle. While reviewing the tape to study how shoppers negotiated the doorway during busy times, we began to notice something weird about the tie rack. Shoppers would approach it, stop and shop until they were bumped once or twice by people heading into or out of the store. After a few such jostles, most of the shoppers would move out of the way, abandoning their search for neckwear. We watched this over and over until it seemed clear that shoppers -- women especially, though it was also true of men to a lesser extent -- don't like being brushed or touched from behind. They'll even move away from merchandise they're interested in to avoid it. When we checked with our client, we learned that sales from that tie rack were lower than they expected from a fixture located on a main thoroughfare. The butt-brush factor, we surmised, was why that rack was an underperformer.

As I was delivering our findings to the store's president, he jumped up from the conference table, grabbed a phone, called down to the floor of the store and had someone move that tie rack immediately to a spot just off the main aisle. A few weeks later the head of store planning called me to say that sales from the rack had gone up quickly and substantially. Since that day we've found countless similar situations in which shoppers have been spooked by too-close quarters. In every case, a quick adjustment was all that was needed.

Another such "accident" of patient observation and analysis happened during a supermarket study we performed for a dog food manufacturer. When we staked out the pet aisle, we noticed that while adults bought the dog food, the dog treats -- liver-flavored biscuits and such -- were often being picked out by children or senior citizens. We realized that for the elderly, pets are like children, creatures to be spoiled. And while feeding Fido may not be any child's favorite chore, filling him up with doggie cookies can be loads of fun. Parents indulged their little ones' pleas for treats here just as they did over in the cookie aisle.

Because no one had ever noticed who exactly was buying (or lobbying for the purchase of) pet treats, they were typically stocked near the top of the supermarket shelves. As a result, our cameras caught children climbing the shelving to reach the treats. We witnessed one elderly woman using a box of aluminum foil to knock down her brand of dog biscuits. Move the treats to where kids and little old ladies can reach them, we advised the client. The client did so, and sales went up overnight.

Even the plainest truths can get lost in all the details of planning and stocking a store. A phrase I find myself using over and over with clients is this: The obvious isn't always apparent.

While studying the cosmetics section of a drugstore chain, we watched a woman in her sixties approach a wall rack, study it carefully and then kneel before it so she could find the one item she needed -- concealer cream, which, because of its lack of glamour, was kept at the very bottom of the display. Similarly, in a department store we watched an overweight man try to find his size of underwear at a large aisle display -- and saw him stoop dangerously low to reach it, down near the floor. In both cases, logic should have dictated that the displays be tailored to the shoppers who use them, not to the designers who made them. Move the concealer up, we advised, and put something aimed at teen shoppers down near the floor -- the teens will find their products wherever they're stocked.

In some studies, we synthesize every bit of information we can possibly collect into a comprehensive portrait of a store or a single department. A major jeans manufacturer wanted to know how its product was sold in department stores, so in one weekend we descended on four sites, two in New England and two in Southern California. Each department was similar -- the jeans section was a square area that held from eight to twelve tabletop displays and some wall shelving. We started by drawing a detailed map of each, showing the displays and the aisles leading into and out of the sections, but also where any signs or other promotional materials were posted. During that weekend we tracked a total of 815 shoppers and observed many more on camera, both video and time-lapse. We paid particular attention to the "doorways" -- our term for any path leading into or out of an area of a store. Until the client knew which paths were most popular, it was impossible to make informed decisions about where to stock what, or where to place the merchandising materials meant to lure shoppers.

By the time our study was completed, we could say which percentage of customers used which paths into each of the sections. Once we knew that, it was clear, for instance, that much of the signage was misplaced -- common sense dictated that it be positioned to face the main entrance of the store, but we found that most jeans shoppers came upon the section from a completely different direction. Even the client's big neon logo and a monitor showing rock videos were facing the wrong way if their job was to signal the greatest number of shoppers. We tracked shoppers from table to table, seeing where they stopped, what signs they read, whether they noticed the video monitors, and how they handled the merchandise, including whether they took anything to the dressing rooms. If they seemed to be showing jeans to a companion, we noted that, too. Some of the shoppers captured on video were also questioned by our interviewers, so that their demographic information and their attitudes and opinions could be correlated with their behaviors -- to see, for example, whether young shoppers with high school educations who say they depend on name brand when choosing jeans read price tags. After the research is done and the numbers are crunched and analyzed, we see what sense can be made of what we've learned. For example, if we were to find that a high percentage of male shoppers buy from the first rack of jeans they encounter, and that shoppers tend to enter the section through the aisle leading from men's accessories rather than from the women's side of the store or from the escalator, then we would advise our client to ask for the display table nearest men's accessories. Or maybe there's another determining factor -- maybe men who are accompanied by females and entering the section from the women's department buy more jeans than men who are alone. in that case, the best table would be nearest the women's merchandise. But no one knows for sure until we collect the data.

In other instances we're hired to study some small retail interaction in great detail. One such project was commissioned by a premium shampoo maker that wanted to know about the decision-making process of women shoppers who buy generic or store-brand beauty products. The client was interested in the "value equation" women bring to each shopping experience -- how does the shopper who buys from the generics section at the supermarket in the morning and then from Nordstrom in the afternoon decide which product she'll buy where? Does she judge that her skin deserves the premium brand but her hair can settle for the generic? Once upon a time only the budget-conscious bought store brands, but now you find them in everyone's shopping basket. Let's call her shopper number 24, a thirtysomething woman in yellow pants and white sweater, accompanied by a preschool girl, who enters the health and beauty aisle of a supermarket at 10:37 a.m. on a Wednesday morning. She has a handbasket, not a shopping cart, and has already selected store-brand vitamin C capsules, a large container of Johnson's Baby Powder and a packet of snapshots she picked up at the photo-processing booth. She is also holding a shopping list and the store circular. She goes directly to the shampoo shelves and picks up a bottle of Pantene brand, reads the front label, then picks up a bottle of the store brand and reads the front label, then reads the price tag on the Pantene, then reads the price on the store brand, and then puts the store brand in her basket and exits the section forty-nine seconds after she entered it. In that brief encounter, there was lots of data to collect -- what she touched, what she read and in what order, about twenty-five different data points in all. If, in one day, we track a hundred shoppers in that store's health and beauty aisle, it can amount to 2,500 separate data entries. As the woman exits the section, we interview her, asking twenty questions in all. So each of the twenty-five data points has to be cross-tabulated with each of her twenty answers -- a cross-tab challenge, take it from me.

No university, to my knowledge, has ever attempted behavioral research in the retail environment to the degree that we have. My old colleagues in the world of academia regard what we do with envy and horror -- envy because we get to do what we do and get paid for it, horror because we actually stick our necks out and are held accountable for the success or failure of our suggestions. After almost twenty years of work, our client list is as blue-chip as they come, and while we do get it wrong sometimes, three-quarters of our clients who buy us once come back for more.

-- Cherri (whatever@who.cares), May 06, 2002


If a shopper tracker were to track me in a mall, they would see me:

1) Walking through the mall directly to the store I think may contain my intended purchase, as quickly as possible, with a pained look on my face.

2) Walking as directly as possible to the section of the store I think may contain my intended purchase, as quickly as possible, with a pained look on my face.

3) Locating the merchandise as quickly as possible within the section. At this point the look on my face becomes less pained, because I sense that soon I will be able to leave.

4) Scanning the shelves and displays for the item I think may best match my needs, picking up some items to look at the price tag. This sometimes reintroduces a pained look to my face.

5) If an item appears to match my needs at a reasonable price, I evaluate it for quality, trying to determine if it has hidden defects.

6) If the item is satisfactory, I pick it up, proceed to the nearest cashier, wait my turn patiently, politely decline the offer to apply for a credit card, pay for the item, and leave the mall as quickly as possible.

7) I do not return to the mall again, except under dire provocation.

-- Little Nipper (, May 06, 2002.

Moderation questions? read the FAQ