It’s a mistake to ask people what they think they want. As Henry Ford said, “If I had asked my customers what they wanted, they would have said a faster horse.” Although some may even have balked at that, saying they were very happy with the horse they already had (particularly if they had only just bought it). Similarly, it’s a mistake to ask people what they think about a new product, service, or marketing idea and let their collective opinion shape yours. When market research wanders into the realm of the future it is inherently reckless.
Most of the examples I have referenced in this book have involved an element of research futurology: the attempt to create a New Coke that consumers will love more than another product; the number of people who will decide they want to visit the Millennium Dome; whether enough people will decide they like Red Bull, Baileys, a Sony Ericsson mobile phone, or an advert for beer. While they highlight failings with the research process whatever its focus, they serve as useful examples because real life has had the opportunity to prove them wrong. The scientific understanding that has emerged in recent years from social psychology and neuroscience shows that most research is an exercise in wishful thinking. The only thing market research guarantees is that you will get some answers, not that they will be an accurate reflection of people’s motivations, needs, or desires.
Fortunately there is sufficient research agnosticism for people occasionally to trust their instincts, discount what research says, and go ahead anyway and, as a result, evidence of such research failings occasionally finds its way into the public domain. However, the vast scale of the research industry is testimony to the fact that hope of obtaining reassurance that a decision is good before it’s taken overpowers the experiences that show it can’t.
This is hardly a unique example of wishful thinking in the human condition: religion propagates the implausible but desirable notion of life after death, astrology that the positioning of celestial bodies can explain the present and predict the future, and homeopathy that the more diluted a substance is, the more potent its remedial properties will be. History shows that people are quite willing to embrace an idea if it is comforting and discount or discard experiences that undermine it.
There are companies who acknowledge that it is incumbent on them to make decisions about what to create, embracing their legitimate position as the ultimate arbiters of expertise and custodians of their own brand. The car manufacturer Porsche has never sought customer opinions of its car designs; it recognizes that it is for the company to determine if something is right. The benefit of its approach is evidenced by the Porsche Cayenne SUV. The car received very mixed reviews, with many experts deciding that it was ugly. However, through Porsche staying true to the visual cues that appealed to the 911 buyers who were buying competitors’ SUVs as family cars and giving people time to get used to seeing the design, the car became successful, and within a year was outselling every other model the company made.1
BMW’s radical redesign of its 5 series in 2003 led to mixed reviews from industry pundits. However, within half a decade many competitors’ new cars were echoing elements of the BMW design, and those that didn’t were starting to look distinctly old-fashioned. The German car manufacturer works hard to understand its customers, but it knows that it can’t rely on them to shape design. As BMW’s design director explained in an interview with the Wall Street Journal in 2008, the company would be asking customers to make judgments based on the world today, whereas the designs need to appeal to them eight years later.2
Perhaps the most ironic example of research futurology I’ve encountered occurred when the BBC explored the idea of commissioning a new version of the classic science fiction series Doctor Who, a programme about a “time lord” who makes frequent trips into the future. According to the show’s writer, “the research found that no one wanted to watch Doctor Who. Kids said it was a programme for their parents. The parents said it was a dead show. I expected it to die a death after a year.” The research had concluded that the program was a niche show for sci-fi geeks.3 Program controllers ignored the findings and produced the program anyway, and over the last five years it has been a huge BBC ratings success. The BBC might just as well have asked the fictitious doctor whether people would watch him in the future.
In most walks of life people are understandably skeptical about claims of an ability to predict the future accurately. While some fall for the vague generalities proffered by astrologers and psychics, such practices have never fared well when subjected to scientific scrutiny. They are perpetuated because of the victims’ (or as the practitioners of these dark arts might describe them, clients’) unhealthy levels of confirmation bias regarding whatever mystical package the forecast has been wrapped in: they will attribute importance only to the times the predictions are right and dismiss those that aren’t as irrelevancies.
Market research routinely makes the implicit claim that it works as a tool to see into the future. Will customers buy a particular product? Will they buy it again? Which packaging will encourage them to purchase more or more often? Will they come back to a shop, restaurant, or venue again? How will they vote in an election? All these depend on the notion that the response to some abstract stimulus, be it a question or a real-life example taken entirely out of context, can be relied on as an indicator of what will happen at some later point in time; all are contingent on the capacity of each respondent to predict the future.
The basis for this “faith” in consumer research must stem from one of two alternative hypotheses: either that if you ask people what they like now, they will tell you honestly and it will remain constant in the future; or that if you ask people what they will do in the future, they will tell you honestly and then go on to do what they’ve said they’ll do.
The desire to predict the future is an understandable one, and one into which much time, effort, and money is invested. In the late 1980s I was working for one of the UK’s largest financial services organizations in its research and planning department. An econometric model of the UK housing market had been commissioned at considerable expense from one of the world’s leading management consultancies. After statistically modeling all of the available data a statistically accurate model of the market was obtained, illustrated by a chart with two very closely matched lines; one showing the actual average house price and the other the one predicted by the model. While the technical basis for such a model is to explain the movement in the data historically, the appeal is its ability to forecast what will happen in the future, and this was very much the focus of attention when the model arrived. Unfortunately, within a few weeks the stock market crashed and house prices started to fall dramatically; within four years the average price of a property had halved (a 35% drop in real terms). The model hadn’t foretold the crash and couldn’t explain it.
So is the faith in consumer research’s ability to predict the future justified? Can consumers reliably tell us what they will do in the future?
In Chapter 1 I explained how our behavior is frequently unconsciously driven, and that we have no direct access to the mental processes that influence what we do so dramatically. It’s no surprise that if we have no access to such processes when the events are “live,” we can’t be expected to forecast how we will behave under their influence in the future.
When people are asked to imagine how they would feel if something fantastic or terrible happened to them, they invariably exaggerate the extent of their response. Many people would love to win the lottery and believe that they would take the sudden transformation to having great wealth in their stride. And yet there are a surprising number of well-documented cases of big winners who held such views when they bought their ticket finding life no happier and sometimes much more difficult once they’ve received the multimillion-pound check. I know someone who inherited a successful manufacturing business and a vast amount of money. While he had the wealth to take whatever holidays he pleased, could drive the car of his dreams, and buy an amazing house, he suffered huge stress in managing the business – something he was poorly equipped to do. Further anxiety ensued about how his wealth should best be invested and, when his investments performed poorly, the magnified distress of knowing he’d lost a significant sum of money as a result. He hadn’t anticipated any of these issues and it seemed that all the advantages his new-found wealth provided were counter-balanced by something he found difficult to deal with.
Social psychologists have conducted research that supports this perspective on our ability to anticipate our futures accurately. It seems that we have a tempering psychological mechanism for positive and negative events. While people seek out happiness, as soon as it is attained its value declines; when sadness occurs it tends not to stay for as long as people predict.4 In time, people return to their personal baseline level of contentment come what may. The same tempering mechanism that helps us overcome grief also removes the shine from happy events. When the unexpected happens people work to make sense of the events, and in doing so the impact that was novel becomes more familiar and more standard as a result. In this way the shock of attending a first car-crash fatality helps ameliorate the experience for paramedics, rather than letting the effect of subsequent experiences be compounded emotionally.
Such processes aren’t predictable or consciously applied. Most of us would, I’m sure, prefer to retain the euphoric feeling of a winning lottery ticket or the triumphal victory of our favorite sporting team for far longer. In one sense I should have been far more disappointed than I was that the Euro 2008 soccer tournament was taking place without the England team. However, rather than each match serving as a reminder of what I was missing, I was able to enjoy the absence of stress that each match would have involved had my own nation been represented. Before the start of the tournament I thought I would have no interest in it; the press campaigns asking me who I would support were a reminder that the team I supported wouldn’t be playing! However, when the Dutch team hammered the Italians in their first group match I found the combination of attractive soccer, connections to friends in the country, and our positive historic national relationship with the Netherlands a sound basis from which to cheer them on. I would never have anticipated it would happen, but I soon found myself making the same diary space to watch the games in which the Dutch were playing as I would have done had England been participating.
Again, studies suggest my experience is normal. People taking part in an experiment where they received unflattering feedback from a personality test predicted their negative reactions would last much longer than they actually did, a phenomenon known as durability bias.5 Research by Timothy Wilson also found that American college football fans thought that the result of a game would influence their happiness for two or three days, but by the following day they were back to their usual level of contentment.6
One could argue that people will be able to state the direction of their feeling accurately: they know they will feel good or bad, they just overestimate to what extent. The problem when using consumer research to gauge feelings about, say, using a new product when it is launched is that the strength of feeling will determine how people behave. The feeling for the product needs to be strong enough to bring about the change in behavior required to purchase it. But people aren’t conscious of this requirement and, as a result, the mechanisms they use to forecast how they will feel in the future aren’t accurate.
Imagine you are given a gun to shoot a golf ball that you know is approximately 200 yards away from you, but you don’t know where. You have a powerful rifle with an equally powerful telescopic sight, so once you’ve found the ball, shooting it will be straightforward. The problem with using a telescopic sight is that it is very good at giving a close-up image of the thing you are looking at, but very bad (hopeless in fact) at putting that image into its context. You may be able to see a small white ball magnified to a useful-sized target when you’ve found it, but locating the target either requires looking at everything in view with the naked eye (virtually impossible) or sweeping the magnified image back and forth looking at each square foot of ground one “picture” at a time. This must be the reason that, even in America, frustrated golfers don’t carry the ballistic means to vent their frustration at wayward shots.
Using market research to explore the future suffers from a similar problem, known as focalism. When considering the future people routinely fail to take into account everything that is going on in their life at any given time and focus too much on the issue being considered. They think too much about how something could happen, and too little about how and why it might not.7 It’s not just the “focus” of a focus group that creates this type of bias, almost every form of consumer research seeks to take customers through a process that has them analyze elements of purchase decisions, concentrating on different aspects of the consumer issue in question. In part, this is a by-product of a process that seeks to rationalize what is often largely or partially not rational, but it is also the result of the way in which the research process is structured. Getting respondents can be time-consuming and so it seems prudent to squeeze them for information, be it through the depth of interviewing, the duration of the discussion, or the number of questions they are asked. The decisions that companies have to make are usually complex and involve numerous elements spanning the responsibilities of several departments and external agencies: manufacturing want to know if their product is liked, marketing that their proposition for it will be well received, finance that the price will be sufficient, and so on. Each of these issues has a multitude of subsidiary questions: How is each element of the product received? How well does it function? Could it be made better? Each respondent will be required to consider every aspect that the company commissioning the research has to consider, however fleeting and superficial their real involvement with that product would be in reality. The more people focus on something (or are required to focus on something because of the research process they’re participating in), the more likely they are to ignore factors that will have a bearing on the issue when it actually happens.
Last week I took my children to a local zoo. In the café we stopped at for lunch, the children’s food arrived in an animal-themed cardboard box and included a cheap animal toy. After a small amount of inter-child toy envy and negotiation with the kind woman at the checkout, both ended up with a six-inch rubber snake. Despite having a monetary value of just a few pence, in the eyes of a 6-and a 3-year-old child these were the most desirable items on the planet. Twenty minutes later, when it transpired that one snake was missing in a play area the size of a football field, I knew that no amount of discussion about the financial worth, meagre quality, or inauthentic design of the item was going to cut it with my daughter, and I was either going to have to start a search of forensic proportions, or else buy another lunch that no one wanted in the hope that there was still a snake available that could be traded for whatever toy came in the box. The experience was all the more frustrating (until I found the snake by chance) because I knew that the novelty would soon wear off. Seven days later I can report that I haven’t seen the snakes; my children have no idea where they are and have no interest in looking for them.
As is so often the case, what one observes in children is an amplified version of a trait we all have to a greater or lesser degree. When we see something for the first time it has an appeal borne of its newness that will, over time, dissipate. We wouldn’t be the extraordinarily creative, progressive, and successful species that we are if we weren’t intrinsically attracted to new things.
One factor that can temper our propensity for novelty bias is our aversion to risk. However, when being asked to evaluate something for the purposes of research there is no risk. There is no purchase to be made, no money to be spent, and no previously favored alternative to reject as a result of selecting the new alternative.
People are creatures of habit. Perhaps because of the natural cycles of life on an orbiting planet, our lives are divided into neat sections of hours within days within seasons within years, compounded by the semi-fixed physiological requirements for food and sleep. Our days become routine filled, and most people can, for any point in the near future, predict with reasonable accuracy where they will be and what, broadly, they will be doing.
When we make new decisions – real decisions with consequences, not those requested by consumer research – we experience anxiety. Will the decision prove to be a good one? Our unconscious mind plays its role in making us aware of the potential risks of whatever we’re considering. Once that new decision is made, and repeated, and no bad event befalls us because of it, we develop a sense of faith about our choice. For example, when I lived in North London I had to park my car on the road overnight. When the obvious (convenient) choice of the road I lived on had no spaces I had a choice about which road I selected as an alternative. The first time I chose an unfamiliar road I was anxious about the risks. Would the car be OK? Would the slightly rougher-looking buildings house people whose idea of a good night out was to vandalize my car? When I found the car undamaged the following morning I felt better about my choice, and with each vandalism-free night my car enjoyed, so my faith in the place I’d chosen to park increased. Very quickly I didn’t have to think about where to park my car if my own road was full; the second road was fine. As it turned out it was on a day when I had been able to park on my own road that, as I walked back home from the train station, I watched someone stroll up to my car and smash the wing mirror off.
What we really choose (as opposed to what we think we’ll choose when we’ve been asked) has much to do with what’s easiest, and that, very often is just a case of doing what we’ve done before. This is one the ways in which advertising changes what we buy: through repeatedly seeing a product in an advert it becomes unconsciously familiar and we are correspondingly more favorable to it. And only the unconscious needs to notice the ad; research shows that conscious awareness of having been exposed to something isn’t a prerequisite for this fluency bias to occur.8 I suspect that the unconscious mind is duped into concluding that something it has encountered on several occasions without coming to harm is “safe.” This mechanism of unconsciously accepting familiar things would have made sense when our ancestors encountered new objects for the first time: “Is that woolly white thing going to attack me if I get closer to it?” After a few encounters it is deemed safe and there’s no need to waste all that energy on tiptoeing past the sheep; unconscious and conscious attention can be directed elsewhere.
In one study Song and Schwarz tested reactions to different fonts by giving participants instructions for either cooking or performing an exercise in an easy- or hard-to-read font. They found that when the font was harder to read people assumed that the same behavior would take longer, be harder, and necessitate greater skill. We’re so preoccupied with doing what’s easiest that even a typeface can shape our response.9
There’s also a fascinating leading principle with fluency. When something is primed by a related word or image that thing becomes more likeable. In experiments when people are shown the word “key” or “lock,” they like a subsequent picture of a lock more than people who’ve been primed by an unrelated word such as “snow.”10 It would seem that when seeing one thing the brain opens up access to other related items just in case, and when that object appears the ease with which it can be accessed is interpreted as us liking that thing more.
But it doesn’t just require a direct link between two items to trick our brains into liking what’s easy or fluent. Labroo, Dhar, and Schwarz found that a highly familiar yet unconnected image on a wine label, such as a frog, increased liking for the product compared with a label without a distinctive and familiar image.11
Several years ago I was asked by a company whose sales were in decline to explore perceptions of its brand and food products. During the research I gave respondents a product to taste but didn’t tell them anything about it. The response was overwhelmingly positive and the consensus was that the company would benefit enormously from launching the product; almost everyone I interviewed intended to buy it. However, the product had already been on the market for more than a year, in a clearly differentiated pack design, and the brand users I was interviewing had been taking their usual product off the shelf and ignoring the new one entirely.
The extent to which this preference for what’s familiar can stretch is quite extraordinary. An analysis of share performance found that both in an artificial laboratory simulation and in real-life stock markets, shares with names (ticker codes) that were easier to say were predicted to perform better and actually did so!12
Another dimension to this affection for the known emerges from the bizarre discovery that people are more likely to live in a town or have a job that has a link to their name, either through it sounding similar or by virtue of it beginning with the same letter.13 American researchers Pelham, Mirenberg, and Jones attribute this irrational outcome to our capacity for implicit egotism, borne of our preference for things that are connected to ourselves. Ultimately, in real consumer situations (as opposed to research situations), people are attracted toward what’s familiar and easiest to process mentally, rather than what’s new.
Attempts to use market research as a forecasting tool are notoriously unreliable, and yet the practice continues. Sometimes this happens because companies have geared up their business on the basis that it works. For example, the research on the Pontiac Aztek didn’t highlight the car’s lack of appeal. Conversely, Chrysler’s research on the PT Cruiser led to a dramatic underestimate of how many it would sell. On other occasions research continues to be used because accepting the alternative is inconvenient. Opinion polls give politicians and the media plenty of ammunition for debate, but nothing they would attach any importance to if they considered their hopeless inaccuracy when compared with the real data of election results (and that’s after the polls have influenced the outcome of the result they’re seeking to forecast).
When exploring something new in research, its appeal may be enhanced by the focalism of the research process and its risk-free novelty or, as Red Bull illustrated, devoid of the social context that will ultimately make it appealing. Futurology-style research also fails to take account of the fluency challenge facing whatever is being researched when it is encountered in the real world, where the familiarity of what we usually see and do takes over, resulting in the new product being either entirely overlooked or rejected. Just as Chapter 1 showed that we are poor witnesses to our own past behavior because of our inability to access the unconscious processes that drive so much of it, so it is when we attempt to predict our future. It’s worth noting too that the issues I’ve described can be compounded when several new things are tested. Such exercises become implicitly comparative and inevitably something will be deemed “best.” So what, if anything, can be done?
Understanding customers in the present and appreciating how they have changed in the past are valuable tools. Of course, given what I’ve discussed up to this point, I’m not advocating self-report and self-analysis-style market research informing about these issues. The challenge is to observe and interpret the behavior of your customers and place it in an appropriate historical context. In this sense, launching a new product is much like buying a present for someone. The better you know the person, the better you know how they’ve reacted in the past, the more likely you are to buy something they really appreciate, but there is always the possibility that you will miss the mark.
When looking to the future it is important to separate the stages of research that are actually concerned with future response from those that are gauging current behavior. It may be more convenient to discuss what consumers currently do and then introduce your “solution,” but such a process is a recipe for inaccuracy. Depending on the nature of the problem and the style of research this can cause one of two issues. It risks constructing the problem and then presenting a logical conscious solution, in which case why wouldn’t someone think it’s a good idea? Or else it reinforces what people currently do, encouraging them to construct conscious reasons, and then offers them an alternative that requires them to devalue the rationale they’ve only just constructed!
Where future research is a question of “Will customers like this alternative version of something that exists if we offer it to them tomorrow?”, the closer to the real-world context it can be placed the better. That may be through a large-scale trial, but at the very least a customer’s real purchase mindset should be recreated and as much contextual information as physically possible provided to mirror the competing alternatives and aspects that will be unconsciously processed in reality. Naturally, the more artificial the test, the less weight should be attached to it in the overall decision-making process.
When future research is a question of the genuinely innovative, we must accept that consumers can’t make those predictions any better than us, just because they might buy the product concerned. As Patrick Dixon, futurist and author of Futurewise, explained using his mother as an example in a speech to the leadership team of cellphone network MTN:
So we listen to our customers, we get close to them, listen to my mother please, but just don’t believe what she says. We build an image of what her life is like, then using your techno-vision we build a vision, a future, that my mother may live in and we try to imagine how she will behave in that place.14
When an initiative involves something genuinely new, the challenge is to forecast the future ourselves by estimating how the product itself and the marketing created to influence how people feel about it will affect the consumer landscape. The key issue then is not what research techniques can be devised, but how an organization can gear itself up so that it is equipped to test such innovations quickly and cost-effectively, roll them out with deference to the fact they may not succeed, and accept the financial pain and opportunity to learn when what are essentially entrepreneurial endeavors fail.