If you want to know why someone does or doesn’t buy, you have to understand how the environment shapes behavior. Divorcing the quest for understanding from the context in which it takes place is a recipe for leading yourself astray. To maximize sales or the impact of communication, the environment has to be right.
Over the past 20 years numerous studies have revealed how our behavior is influenced by elements of the retail environment that should, logically, have no bearing on what we choose to do. While it may not be a revelation to learn that music and lighting can affect our mood and, as a result, our behavior, the extent to which both can cause people to spend more is surprising, as well as further evidence that we’re not equipped to be aware of what shapes our own behavior.
In general, consumer research is conducted in a place that is convenient to the researcher.1 Indeed, research tends to come labeled according to where the data is obtained: in-street interviews, online surveys, in-home, hall tests, viewed (viewing facility) groups, and so on. Implicitly, the message is clear: it doesn’t make any difference where you ask questions, you’ll get the same response. As I will discuss in the next chapter, there is much to be learned from watching what people buy and how they buy it, but first I should explain why the environment matters so much, how it can change what people do, and what a dramatic difference it can make to sales.
Charles Areni, who specializes in studying the environmental psychology of commercial space, set up a test in a shop that sold wine, playing either top 40 or classical music. He found that people spent more than three times as much on a bottle of wine when classical music was playing compared to when pop music was selected.2 Of course, all the people involved assumed that they were buying the wine they wanted to buy and would be able to provide apparently rational justifications for doing so, but they didn’t know that the quiet background music was the only variable being altered. Recently, a wine industry provocateur has said he believes the taste of wine is influenced by what music is playing; a bizarre theory perhaps, but one that makes sense when one considers the impact of unconscious mental associations and potential for misattribution that psychological studies repeatedly encounter.3
For instance, researchers have found that the type of music played can dramatically alter the amount of time people stay in a store and how quickly they move, and can change their perception of how long they’re kept waiting or how crowded the shopping area is.4 Not surprisingly, these influences on behavior and perception can lead to greater spending; a comparison of slow and fast music in a supermarket found that the former led to 39% higher sales.5 Again, no one would suggest that these people walked out thinking: “I definitely spent more because all the music was less than 60 beats per minute.”
In the US, psychologists experimented with changing the lighting in two retail displays, one featuring tools in a hardware store and another with belts in a western apparel and feed store. They installed additional 500-watt lighting in the ceiling, which they could control independently from the main lighting in the room. Through videoing the shops’ customers they recorded the amount of time people spent at the display, the number of items they touched, and the number they picked up. They found that consumers who engaged with the displays touched more of the items and spent significantly longer there when the additional lighting was turned on.6
It is well documented that light levels have an effect on brain chemistry: light regulates the body clock and is associated with the release of serotonin, which plays an important role in the regulation of mood, anger, and aggression. However, only people who’ve been diagnosed with a condition such as Seasonal Anxiety Disorder might be expected to recognize that they’d feel better if they got more light. It is entirely reasonable to suppose that if the light in the research environment is significantly different from consumer reality, people may feel and react differently.
Beyond the way in which changes in lighting and music cause people to behave differently, more subtle variables such as the proportions of the room have been shown to change how people think; the very nature of their thought processes appears to alter. Two marketing professors created four rooms that were identical except for the height of the ceiling, which they set at either 8 or 10 feet. By giving participants different tasks that required different types of mental processing and analyzing the results, they discovered that people in rooms with higher ceilings performed better at tasks requiring relational processing (to do with identifying and evaluating the connections between different sports), whereas when the ceiling was lower participants performed better at item-specific tasks. They also found statistically significant differences in how two products were evaluated.7 Of course, no one taking part was told that ceiling height was the focus of the study. Just like the research on subliminal smells and images revealed in the previous chapter, these environmental influences take place at an unconscious level and, through some twist of evolutionary fate, our conscious minds are oblivious to what’s really driving the thoughts, feelings, and behavior that result from them.
Another influential aspect of context that consumer research routinely fails to consider is who else is present during an influential phase of the consumer experience. Anyone who has ever been in a shop with a young child will know that the retail experience can be dominated to a far greater extent by the child’s actions than by any other environmental variable. A 2 year old will want to stop and touch a huge proportion of the products and displays. Depending on the circumstances this could have a number of effects on the adult concerned: noticing something that would otherwise have been screened out; reduced sensitivity to unconscious environmental influences; or the desire to abandon the consumer experience altogether!
In one study, the time spent shopping at a DIY store was compared for different combinations of consumers.8 Women shopping with a female companion were observed to spend on average more than 75% longer in the store than a woman and a man shopping together. It is relatively easy to analyze one’s own experiences and speculate on why this might be the case. Irrespective of the reasons, the fact remains that people’s thoughts and feelings as evidenced by their behavior must be markedly different in these contrasting circumstances.
As consciously capable, self-rationalizing beings, we prefer to believe that we are the sole masters of our choices and our destiny, however much an objective assessment of our lives may contradict this. How many of us recognize that we have been influenced by an advert or the actions of a salesperson? Even where we can grudgingly acknowledge their presence, most of us prefer to believe that a salesperson’s involvement was only one (very marginal) factor in our decision, rather than the critical point of influence that determined the outcome of our experience. Of course, any retailer that monitors sales on an individual staff basis will be able to point to one person who generates considerably more sales than another, over the same period of time in the same retail outlet. Their efficacy could conceivably be attributable to work rate, but when you spend a reasonable amount of time watching different salespeople in action you soon see that some have skills in sizing up customers, developing a rapport with them, and then tailoring their sales approach as required. Ultimately, the outcome of a consumption occasion can be entirely determined by the human interaction that takes place with the salesperson.
It is fascinating to observe how good salespeople get customers to tell them what they will be influenced by, and then leverage that dimension later. I’m not suggesting that this is duplicitous in nature; anyone whose job it is to make sales and who has the interpersonal flexibility to try different approaches will unconsciously assimilate an understanding of what works and what doesn’t. Sometimes it’s as simple as asking if the customer has a particular brand in mind. If the customer volunteers a brand name that the store stocks, the salesperson can eulogize about that brand and have a very good chance of making the sale. The alternative, suggesting that a brand the customer proposes is somehow inferior to another, carries a significant risk of undermining the customer’s confidence and causing them to go away “to think about it.”
Of course, a combination of poor training and an absence of interpersonal flexibility can create a very different effect. I vividly recall going to look at a car at a dealership several years ago and the salesman repeatedly asking me what was stopping me buying the car I’d taken for a test drive. My issue was a fundamental one – the car wasn’t quick enough – but rather than acknowledge this and offer an alternative car, he pursued a bad script and asked: “If we could overcome that would you buy it?” I could see no practical way of changing the car’s performance and he didn’t suggest one, so the conversation quickly became mystifyingly abstract. I didn’t buy the car, or a faster one, from that salesman. I don’t know if he is still selling cars, I’m not sure if he ever sold a car come to think of it, but I suspect that there is a salesperson somewhere who would have approached things differently and persuaded me at least to stay in touch with the dealership.
The impact of the presence of other people is demonstrably significant even when, logically, it should make no difference. As you will see throughout the book, and particularly in Chapter 5, there are numerous reasons to question the validity of responses obtained to questions. However, one aspect is the way people’s answers to questions change depending on the way they are asked. Online surveys are becoming hugely popular because of their relatively low cost and high speed, but people give different answers to certain questions when they are sitting in front of a computer screen alone from those they express when someone is there to ask the question. In one study, the answers to questions such as “How do you manage on your income?” varied between 29.9% and 47.7% saying they were “comfortable,” depending on whether the question was answered in the presence of someone else or not.9 Another comparison, conducted by a large opinion polling company, found that 73% of people interviewed by telephone who thought prison sentences should be served in full disapproved of muggers being spared a term in prison. The same poll conducted online at the same time found that only 52% fell into the same category. Despite the pollsters ensuring that the two samples were weighted to reflect the demographics of the UK population as a whole, they obtained statistically significant differences to 22 out of 28 questions!10
The market research industry’s reaction to such anomalies is to scrutinize the approaches used and ask: “Which method is more accurate?” The answer to this inherently biased question is “neither”: what we think is a by-product of where we are, what and who we’re surrounded by, and how we happen to be thinking at that moment in time. It may well be that, in the presence of someone whose job it is to stand in a street on a cold day and plead with people to come and answer a few questions, more people consider themselves well off. Conversely, when they are sitting in front of their computer answering tedious questions in the hope of winning £5,000 in a sweepstake for participants, part of them asks: “How has my life come to this?” Of course, if you’ve just been stopped in the street by someone collecting for starving children in Africa, or watched the people of Haiti fighting for their lives, your answer will be different again. In trying to quantify such nebulous concepts in any way, market research is really missing the point.
Another aspect of the consumer environment consists in the products and product communication that surround any given item. Research has shown that the result of advertising is influenced by the context in which it is placed. In tests where the context of the magazine or program was similar to the subject of the ad, it resulted in better understanding and greater likability of the advert. A good feeling about the context is misattributed to the ad itself.11
Where participants who knew a lot about cars were asked to evaluate an advert for a Honda, they rated it more favorably when it was surrounded by ads for prestigious brands like Armani and Rolex, than when it was in the context of less premium brands like Timex and Old Navy.12 When Simonson and Yoon compared how people evaluated the attractiveness of a series of products, including lawn mowers, food processors, and cars, they found that the strength of preference for a product was influenced by the context of choices presented at the time. For example, when a pen was selected from a set where it was significantly better than another, participants would pay more for it and think it wrote better than when the same pen was selected from a more balanced set of options.13 With the vast sums spent on advertising, a relatively small investment replicating Simonson and Yoon’s study for your own products and media options could lead to a dramatic difference in the way people feel about your brand.
With such diverse and intermingling influences shaping consumer behavior excluded from the market research process, it is perhaps unsurprising when its conclusions are wide of the mark. When McDonald’s developed the Arch Deluxe burger in the mid-1990s, the company was confident that it had a winning product that would appeal to adult consumers. In the context of its market research the product performed very well, but in the context of a McDonald’s restaurant, complete with “Happy Meals,” Ronald McDonald, and other child-associated cues, the reaction was very different. Ironically, the advertising concept, which featured Ronald McDonald taking part in more grown-up activities, probably reinforced the contradictory associations customers were battling with.14
The corporate desire for control and standardization is understandable. It could be argued that it is essential to success for certain functions like accounting, procurement, and branding. However, as McDonald’s discovered, centralized processes don’t always offer the answer. When it comes to market research, the desire for a steer on development away from the complications of the retail environment are extremely risky. Devoid of the context in which it will eventually be sold, consumers can’t respond authentically to a product. McDonald’s developed its “Burger with the Grown-up Taste” from its Oak Brook headquarters in a direct move to appeal more to adults. Away from the plastic seating, bright primary colors, and menus of familiar, child-friendly alternatives, respondents rated the product highly for taste, freshness, and satisfaction. Despite more than $200 million of expenditure, at least $100 million of which was spent promoting this product that research had shown was so appealing, it failed and was withdrawn. According to one source, most of McDonald’s successful product innovations, including the Big Mac, Fillet o’ Fish and Hot Apple Pie, were invented in operators’ kitchens, not remotely at head office. While these franchisees may not have had “robust” survey methods with which to research their inventions with customers, they could at least test their products in the environment in which they would be eventually be sold.15
All human behavior is heavily influenced by the environment. As Kevin Hogan, author of The Science of Influence, says:
Humans, like animals, interact with and respond to their environment far more than we are aware of at a conscious level. If you want to change your own or someone else’s behavior, the first thing you can do is change the environment. Changing the environment is uniquely powerful in changing behavior. There is no greater single influence.
Psychologist Stanley Milgram conducted a pioneering and now famous study on how context can change behavior, illustrating very powerfully that what people are prepared to do can change dramatically when the environment alters. The context can determine not just how the person behaves, but how differently they act from the way they might have expected to, and, in most cases, how they would like to tell themselves they would. In his experiment, 37 out of 40 participants administered potentially lethal 450-volt electric shocks to another participant in the test, simply because an apparently authoritative person in a well-regarded university’s science lab asked them to do so.16 (In fact, the person being shocked was in on the experiment and was only acting out the pain from the electricity the participants thought they were delivering.)
In the Stanford Prison Experiment, psychology professor Philip Zimbardo collected a group of students and assigned them at random to be either prison guards or prisoners in a mock jail that he’d created. Before the experiment was concluded (eight days early, after the intervention of the psychologist’s girlfriend who had been brought in to conduct some interviews with participants and objected to the conditions that had developed in the mock prison), the guards had attacked prisoners with fire extinguishers, refused them access to a toilet, made them sleep on concrete floors, and subjected some to sexual humiliation. The experiment only lasted six days.17 These dramatic examples are more extreme than anything market research is likely to encounter or create, but they illustrate the psychological principle involved.
When Mattel decided to introduce a new version of Barbie’s companion Ken in 1993, it asked its target audience, 5-year-old girls, how they wanted Ken to look. The result was almost certainly a reflection of what girls of that age regarded as iconic of attractive masculinity at that time: boy bands. The groups often had gay impresarios or stylists and were happy to target the gay movement that was so vibrant as it worked actively to establish itself in mainstream life.18 The resultant “earring” Ken doll wore a lavender mesh top with matching faux-leather sleeveless jacket, had two-tone frosted blond hair, and sported a small silver ring on a choker that was swiftly identified as a cock ring (a popular adornment of gay clubbers at the time).19 This wasn’t something most parents were happy to buy for their children; nor, once it had been pointed out, was it an image Mattel was happy to be associated with, and the doll was hastily withdrawn.
The artificial nature of the research environment can also be responsible for not flagging up something that, in the real purchase environment, is unconsciously reinforced and hugely significant in determining a product’s fortunes. When Heinz developed All Natural Cleaning Vinegar it was a logical concept: the company already knew that people used their “eating” vinegar for cleaning and media interest in more natural cleaning products was high. Away from the context of the supermarket, a specific Heinz cleaning product seemed like a good idea. However, in the environment of a supermarket and, in particular, the context of the company’s food products, it was hard for consumers to reconcile those unconscious culinary associations with those of cleaning products that they associated with dangerous chemicals and germs; the product failed and was withdrawn.
Often the influence of the environment on behavior is so subtle as to be imperceptible. Aron and Dutton set up an experiment involving two groups of male students.20 Ostensibly the purpose of the research was to examine the impact of scenery on creativity; however, an attractive female interviewer was asking the questions and the two groups were in differing locations. The individual interviews for the first group took place on a wobbly bridge over a deep ravine; the second group were located on a bench on the other side of the bridge. The researchers were in fact interested to see how many of the participants asked the researcher out on a date. The difference was dramatic: 60% of those interviewed on the bridge telephoned the interviewer; just 30% of those on the bench did so. The researchers concluded that the people on the bridge misattributed their psychological arousal from the unstable bridge to the girl. Put another way, they knew they felt something, and their conscious mind wrongly diagnosed the sensation as a feeling of attraction to the researcher standing in front of them, rather than the fear of falling to their deaths from an unstable bridge.
Something similar was probably a significant factor in the failure of Peugeot’s 1007 car.21 When it was initially revealed as the “Sesame” concept car at the 2002 Paris Motor Show, the positive public reaction led the company to build and launch it; according to Peugeot’s 2005 annual report, the “creative design delighted” executives. Whether the reaction was attributable to the novelty of the car’s sliding doors, something Peugeot or another manufacturer was doing to generate excitement around its stand at the show, or the general buzz of the motor show is impossible to identify at this distance. However, the poor sales since the model was launched suggest that the positive reaction people thought was attributable to the new small car was misattributed excitement from something else going on at the time.
Creating the appropriate mood around a product – be it by staging an exciting event, wrapping a “hot” celebrity around it, giving it to people when they’re having fun doing something else, or making them feel they’ve got a great bargain – can boost a brand’s appeal precisely because of the phenomenon of unconscious misattribution. Experiencing a powerful feeling at the same time as experiencing the product can be sufficient to make the less rational part of our mind perceive something in a way that it otherwise wouldn’t.
Online retailers have the luxury of being able to conduct live trials that enable them randomly to assign customers to one of two or more alternative page layouts or even entirely different website designs. This gives them both a large degree of control over what a customer experiences and an opportunity to explore how subtle alterations to the environment change behavior.
The challenge with so much flexibility is to know what to test. It is apparent from studying what has been successful in such tests that, unsurprisingly, the same unconscious preoccupations shape how people respond online as they do elsewhere. There are still people who won’t buy online because they feel it is too risky. However, most of us have evolved through a process of finding that an item we wanted was only available online or else was offered at such a discount there that we took the chance. The importance of feeling secure about buying online, particularly from a retailer that you haven’t used before or that doesn’t have a familiar high-street counterpart, has led to numerous protection schemes springing up. Their importance is evidenced by one website design test for luggagepoint.com, which that found sales increased by 5% and revenue per customer 11% when it moved the “Hacker Safe” crest a couple of inches to the right and removed a small banner next to it that promoted international shipping.22 I can still remember the fear I felt buying a guitar from the United States on eBay in the days before PayPal protection existed. My anxiety was heightened by the fact that the seller insisted in typing his emails in capital letters. This relatively unusual ignorance of the Caps Lock key heightened my worry that he was a fraudster. Eventually, frequency renders such experiences mundane, at which point what matters is that our expectancy of speed and simplicity is met. eBay tells me that I’ve now bought 70 times – I had no idea.
A lack of ease, or fluency, can be a cause of lost sales. Where customers can’t find what they want easily, and even when the first page of a site is slow to load, they will go elsewhere. One study has suggested that, unless there is something on screen telling people that information is being loaded, two seconds is as long as people are willing to tolerate before they will move off.23 Google found that by making the Google Maps website faster (it reduced the size of the page by around 25%) traffic increased by 10% in the first week and by 25% just three weeks later.24 I strongly suspect that the specter of the unconscious mind’s capacity to misattribute feelings is at the heart of this phenomenon: mild frustration at waiting for a page to appear can easily be felt as a dislike for what is on offer.
I now buy items online from my phone in the television ad breaks without the slightest concern. What was once a source of anxiety has been made fluent by Amazon’s one-click order process. I strongly suspect that many of the products I buy from the site would be available more cheaply from its online competitors, but it has made buying so easy I’ve never taken the time to check. While a sufficiently discounted price is a reason to take a risk on buying from somewhere new, it often doesn’t determine subsequent behavior, when factors such as fluency (habit) and social proof take over in importance. A 2001 study based on click-through analysis of North American web users found that only 8% were aggressive price hunters.25
Social proof, in the form of bestseller lists, testimonials, or customer reviews, is also a hugely influential factor. Retail clients tell me that those products that have the best reviews will sell best; it’s no surprise that, in the absence of seeing what other people are buying, such information is important to online customers.
Just as with Moll’s plant wrapper (see Chapter 1), the fact that so many studies show that people are unaware of how the environment and context affect their behavior and attitudes doesn’t deter them from offering an apparently credible justification for what they do and why they do it. However honorable conscious intentions might be, consumers can’t help but create and perpetuate myths about why they buy what they do when the researchers’ questions are asked.
In a local department store I watched as a large number of passing customers stopped to look at a display for a new iPod speaker system. The display included the iPod and was pumping out music, making it hard to ignore. I approached several of the people to ask whether they had been considering buying this particular system prior to seeing it in the store; almost three quarters of them said they had. However, since I knew that the system was new and recognized the “foraging” consumption style in the store, I doubted that so many knew it existed. I changed the subject and talked about the product’s features for a few minutes. When I asked these people later where they had first seen the speaker system, most of them said, as I suspected, that they had seen it in the store that day for the first time.
When I was researching the purchase of instant-win lottery scratchcards, respondents liked to believe that they bought them impulsively. However, I discovered a strange dimension to this impulse. When someone walked straight up to the newsagent’s counter to buy something they would often not buy a scratchcard, but when they were required to stand in a queue they would. The unit that displayed the scratchcards was positioned in such a way that customers only saw the side when they were actually at the counter. The Perspex holder displayed the side of the cardboard reel of the scratchcards, not the eye-catching designs with the distinctive silver panels and references to the cash prizes. When there was no queue customers were less likely to be visually prompted by the card display and consequently less likely to buy one; queues created the visual opportunity to notice the scratchcard unconsciously, and the sight of it triggered associations in some people that they interpreted as a desire to purchase.
Combined with the inability to acknowledge accurately what aspects of our surroundings have influenced our behavior is our capacity to be selective witnesses. It’s troubling to learn that a US study found that false eye-witness testimony contributed to three quarters of convictions that were overturned by DNA evidence.26 Psychologists at Iowa University faked a crime in front of students and asked them to identify the perpetrator from five suspects, none of whom was the actual thief; 84% of students were willing to point the finger at one of the innocent suspects. This increased to 90% when they were told that one of the five had confessed to the crime and their claimed confidence in the identification increased from 6 to 8.5 out of 10.27
People are just as unreliable as witnesses of their own experiences. In 1991 James Randi, a challenger of the paranormal, conducted an experiment with the British medium Maureen Flynn. One of the tricks mediums use is to throw out huge numbers of names (or a few names with a large audience) to find a connection. Some time after the experience, he asked a client of Ms Flynn, who considered her reading to have been “very good,” how many names she had mentioned during the 30 minutes the “reading” lasted. The client estimated that the medium had provided about six names. A transcript of a recording, however, revealed that she had mentioned 37 names, along with the initials N and L (which would provide a link to approximately 300 more).
The problems of context are further compounded by the time between the actual moment of interest to market research and when a question is asked about it. Wegner, Vallacher, and Kelly conducted an experiment to explore changes in how people define their actions over a period of time by interviewing people who were getting married. In advance of the occasion, they would often describe the event in romantic terms; closer to the occasion, they described it in terms of the details that were concerning them (such as getting flowers, wearing particular clothes, and so on); some time after the wedding, they tended to talk in terms of “getting in-laws” or “becoming a member of a family.”28
As I said at the beginning of this book, humans have an extraordinary capacity for seeing things in a distorted way. We’ll convince ourselves that a superstition or alternative therapy works, even though we’ve experienced numerous instances when it has failed to bring about the consequence we believed it had the power to influence. This trait, known as confirmation bias, also enables us to ignore our own apparently contradictory behavior. We can chastise a child for swearing and ignore the fact that, when we struck our thumb with a hammer earlier in the day, we used exactly the same language ourselves. Usually, no one is paying sufficient attention to point out our flagrant inconsistency, but once in a while an event occurs and it is exposed.
In 2007 an article in the Daily Telegraph magazine recounted the story of William Barrington-Coupe, who had passed off recordings of other classical musicians as those of his wife, Joyce Hatto – a pianist in her own right, but by that stage an infirm septuagenarian.29 One twist in this tale involved the review that one of Hatto’s “recordings” had received from the Gramophone critic Bryce Morrison in 2006. The article reports that he described her playing on a Rachmaninov Concerto as “stunning… truly great… among the finest on record… with a special sense of its Slavic melancholy.” However, 15 years earlier the same critic had described the same performance (albeit performed under a different name) by saying that “(the performer) sounds oddly unmoved by Rachmaninov’s intensely Slavonic idiom… devoid of glamour… lacks crispness and definition.” There was no suggestion that the critic was doing anything untoward. A myriad of factors may have influenced his perception of the track on that day: the excitement surrounding this prolific new pianist, an earlier argument with a colleague, a change in his hearing, the impact of other music he had played beforehand, the music system he was playing the recording on, the temperature of the room, the packaging of the CD, the recommended retail price, even the smell of it. One thing is certain: his well-considered, well-intentioned, professional critique was markedly different, but the music was identical. It’s tempting to suggest that his self-contradiction makes Morrison a bad critic, when in fact all we should really take from his unfortunate experience is that he is human and as susceptible as the rest of us to the subtle but significant influence of context.
Market research is gathered in whatever way seems most convenient; retailers don’t want clipboard-laden interviewers harassing their customers. But what chance is there of getting reliable market research if the environment changes how people think, feel, and act? And, given what we know about our inability to access the workings of our own unconscious mind, respondents won’t know they are being influenced and therefore can’t possibly report them or attribute their behavior accurately to them in research. The locations used to label the noncontextual collection point of market research (online, in-street, and so on) should be seen as health warnings of the inherent unreliability of their findings.
The place to understand consumers is when they are in their natural habitat, wherever their unconscious mind is being exposed to everything that might shape how they feel. And the good news is that we can learn a lot from watching what consumers do.