Different light levels affected productivity.Just as we are unaware of how our physical environment influences our thoughts and behavior, we don’t appreciate the subtle but significant influence that the actions of other people can have on us. Cults, religions, and brands all rely on some aspect of group influence to spread their message, sometimes with astonishing speed.
History is littered with examples of times when groups have been influenced to behave in a way that can seem incomprehensible to others. When it came to power, the Nazi Party had two million members; by the time of its demise, it had more than eight million.1 While many of these people joined for career reasons, it has been estimated that there was an active membership of at least one million people, many of whom were in senior positions in the national government and to a greater or lesser degree integral to its nefarious objectives. In 1933, 44% of the German electorate voted for Hitler’s party. No doubt most of us would like to believe that we wouldn’t have been persuaded by the rhetoric of the day, but the reality is that it’s not just what is said that matters: the number of people around you who are nodding their heads can change what you think.
Shaping what the crowd thinks doesn’t necessarily require a large number of people. In the late nineteenth century, following some genuine medical breakthroughs, America was awash with miracle cures as opportunists sought to cash in with remedies of their own. They soon learned that a couple of people, used in the right way, could influence opinion and transform their fortunes. The show would arrive in town, providing hours of entertainment interspersed with short pitches for the nostrums. A couple of accomplices in the audience would buy the product, drink it, proclaim themselves cured, and, with animated desire, rush to buy another bottle. Soon people would be clamoring to buy whatever dubious mixture of alcohol, plant oil, herbs, and paraffin had been packaged up with an appropriately medicinal-sounding label, convinced that it would help them.2
Many political organizations and brand owners use market research focus groups, in the belief that they help them obtain a deeper understanding of what people think. They do so unaware that the susceptibility of people to what one or two others say and do is just as prevalent in modern-day focus groups. While it may not matter to most people if a brand of floor cleaner gets corrupted by this approach, everyone should be concerned about a research technique that shapes the national agenda of many countries when used by political parties.
The appeal of focus groups is driven by the belief that they can elicit in-depth information on a topic: by taking a group of similar people and facilitating a discussion over a protracted period of time, insights will emerge about what those people think. The theory is that, with skillful moderation, comments from one person will trigger additional thoughts from another and so on, until the group has explored its collective thoughts on the issues at hand. One advantage of groups is that a relatively large number of people (perhaps eight or more) can reach the given depth of subject exploration at the same time; another that a common view can be established giving relative efficiencies of time, ease, and cost when compared with speaking to people individually.
It is worth noting that the use of groups in psychotherapeutic work is precisely because of their capacity to affect change in people; and yet it is implicitly assumed that focus groups when used in market research don’t change people at all. While the role of the therapist clearly has a part to play, the fact remains that, were he or she to be the sole point of influence in group therapy, there would be little need to put patients through the additional pain of sharing their psychological problems with strangers.
So why are we so susceptible to what other people think, how does it influence us, and why is it that, however good the moderator, focus groups generate false findings?
I have described the phenomenon of psychological priming and the problems it can cause for research. The capacity for what we’ve just heard to influence what we “choose” to say is part of our inadvertent capacity for copying each other. Any human interaction is a potential source of such primes, and so they are an inevitable but uncontrollable by-product of the interpersonal dynamics of a focus group. Factor in the likelihood that the subject of the discussion is often of relatively minor consequence and there is a very high probability that people will go with the conversational flow. One person choosing to talk along similar lines to the previous one isn’t necessarily evidence of agreement, it’s simply the nature of the way we interact.
It’s not just what we hear that can cause us to follow a similar mental direction. We have a tendency to copy what others do without realizing. Evidence of this emerged in research published in the Journal of Consumer Research.3 People were asked to watch a video of someone talking about a series of advertisements; in fact, the people on screen had been told to eat one of two types of cracker placed in bowls in front of them as they talked. The study found that the people watching mimicked the selection of the person talking in the video, taking the same cracker from the two choices available to them.
One aspect that can contribute dramatically to the success of any initiative is the extent to which people will copy others they see using the product or hear talking about it. This can be achieved in numerous ways: devising a tag line for an advertisement that people like to repeat, like Budweiser did with the Wassup? campaign; making a product’s visibility integral to its appeal, as Apple did with the iPod’s white headphones; creating a buzz around a product by releasing a limited amount of intriguing information about it in advance (provided you can live up to your own hype), as Hollywood likes to do when it lets it be known that a particular new movie is in production, that a particular actor has been cast, or that an on-screen romance wasn’t all acting.
Fads and fashions evolve from this aspect of our desire to mirror others. For no specific reason that we can identify, a shirt that we decided we absolutely had to buy and used to love to wear becomes one that we choose not to put on and sometimes, eventually, one that we’re embarrassed to see ourselves wearing in old pictures.
Unfortunately, focus groups don’t simulate this propensity to be unknowingly influenced in a helpful way, because contextual influences and the wider canvas of day-to-day life are substituted for an entirely abstract and artificial focus on the subject of interest to the research. As a result, the copying we’re so predisposed to manifests itself in the way in which people respond to the topic of the group. The impact this can have was illustrated in a brand development project I conducted recently (the “politics” of the project demanded that focus groups be used; my client shares my perspective on the inherent problems with the methodology). I was asked to gauge the potential effectiveness of a new advertising campaign using several executions that had been developed into videos comprising still images with a scripted voiceover. One of the executions featured a 1970s sitcom-style joke involving a risqué double entendre. When this execution was shown first, the respondents were primed to make sexual associations and went on to find sexual connotations in the other ad executions that were certainly not intended, and that were not perceived by the groups who saw these executions before the one containing the double entendre. I was able to anticipate and discount such blatantly primed comments, but it is not always so obvious when a prime has had an effect.
In the same advertising research, I had reached a point where the artificially rational nature of the process of deconstructing an advertising concept had, unsurprisingly, driven respondents to the point of saying that they wanted totally rational advertising: my client should simply tell them that it existed as a retailer and could provide the products it sold. Customers, the respondents were convinced, would then decide whether or not they wanted these products and act accordingly. Anxious that someone viewing the group might take what these people were saying literally, I needed to expose the artificiality of what they were seeing, so I asked the respondents to name their favorite advert. They all named ads that were nothing like the one they’d requested from my client: emotionally evocative and devoid of rational or tangible claims. More revealingly, perhaps, after the first man to speak gave a car advert as his favorite, all seven subsequent respondents mentioned car ads too; it was as though no one could think of another product category.
Despite what we might like to tell ourselves about our pioneering and independent nature, most of our behavior comprises doing much the same as the people around us. The chances are that we’ll be one of the thousands of people buying the book about an explorer that we’ve seen on the bestseller list, not the intrepid soul who did the actual exploring in the Amazon rainforest. The evidence shows that we can’t help but care what other people think and will go to great lengths to conform.
In 1935 the pioneering social psychologist Mazafer Sherif invited people to take part in an experiment using the autokinetic effect. Participants looked at a point of light in a darkened room and were asked to report whether they thought the light was static or moving, a recreation of a natural phenomenon first observed by astronomers who thought that stars were moving. When participants were asked individually opinion was equally divided; however, when they were put into groups people tended to agree with the majority, even if this meant contradicting what they’d said originally. Later, when asked individually, they continued to subscribe to the group view. In other words, when placed in the context of a group, people will devalue their own opinion in the interest of developing an arbitrary position that is acceptable to the group.
It is relatively easy to demonstrate that an unconscious “group influence effect” exists. Get one person to stand and look at an abstract point somewhere and you will find that people take little or no notice. However, if you get three or four people doing it, virtually everybody stops to see what’s so interesting.
The neuroscience of how group influence affects behavior is still in its infancy. One recent study explored the mechanisms that cause people to tend to like what their friends like. Neurologists conducted fMRI scans of teenagers’ brains while they were listening to unfamiliar music spanning several genres. Each participant was played a number of songs and asked to rate how much they liked them. Then they were shown how popular the song was among a large reference group. To make sure that people weren’t contrary for the sake of it, participants knew that they would receive a CD containing their favorite tracks at the end of the study.
As they expected, the researchers found that people did adjust their ratings to conform to the “popular” opinion of the tracks. However, what they discovered from brain activity throughout this process was what was so fascinating. From the areas of the brain involved (the left and right anterior insula was active in those who changed their preference), it seems that people switched their preference because they were anxious that their opinion didn’t match up with those of other people. This neural activity is distinct from activity for reward and utility; in this case it seems that the music became more appealing not because it was liked or appreciated for its own sake, but because not liking it was worrying.4
One marketing case study that reflects these mental processes in action is the energy drink Red Bull. It was discovered by Austrian businessman Dietrich Mateschitz while he was traveling in Thailand. The drink, which was already called the Thai equivalent of Red Bull, was a cheap tonic sold by a pharmaceutical company and used by factory workers to help them stay awake at work. The results of taste tests were far from positive. The market researchers concluded that no other product had ever performed so poorly in consumer testing: the look, taste, and mouth-feel were regarded as “disgusting,” and the idea that it “stimulates mind and body” didn’t persuade anyone that the taste was worth tolerating.5
When it was initially launched in Austria in 1987, the product didn’t get widespread distribution. However, it became popular with clubbers and snowboarders, to whom the reviving properties appealed and who started to mix it with alcohol. Despite its being a working-man’s drink in Thailand, Mateschitz set the price of Red Bull well above other soft drinks; knowing how price can alter perception, this almost certainly contributed to its success.
The ingredient mix of Red Bull led to a lengthy delay in its German launch while regulatory testing took place. During this time people started to talk about the product they were encountering just over the border in Austria. These discussions about whether the drink was safe were intriguing and spawned excited debate, particularly among a young adult audience that is highly disposed to risk-taking behavior. In subsequent markets the brand replicated this model of exclusivity, carefully selecting edgy venues and activities to be associated with and spurning requests from establishments and retailers that wouldn’t help form the profile for the drink that the brand owner wanted to create. As recently as September 2009, two Swedish convenience store chains banned sales of Red Bull to children under 15, a move that is only likely to increase its status with young people and is unlikely to upset the company that markets it.6
Despite the disastrous research results Red Bull was hugely successful, having powerfully tapped into social curiosity and leveraged both priming and social proof. It wasn’t just that drinking it was cool, but that there was a risk that not doing so might make you look bad to your peers. By 2006 the company had sold more than three billion cans of its “disgusting” drink, achieving sales of over €2.6 billion.
It is crucial for marketers and politicians, and anyone else who hopes to assemble a mass following, to understand the nature of group influence. However, the same influences are unrepresentatively present when a small group of people is assembled to focus on a business or political issue. No amount of careful moderation can mitigate against the fact that what you hear will be a byproduct of the group dynamic rather than a reliable indication of what people in general think.
As the nineteenth-century American salesmen discovered, the quickest way to exert intentional influence over another person is to solicit the help of a group. If several people tell someone something they will be more inclined to believe it; they may even start to doubt their own prior judgment on the matter and accept the “group” view in place of their own.
In 1953 Solomon Asch published the results of a vision test. All but one person in each group of people were “plants,” who had been told by Asch to give an incorrect answer to which of three unequal lines matched another line. More than a third of the people taking part altered their answer to conform to the prevailing view; it required only that three other people confidently state an incorrect answer to generate this change.7 Critics of Asch’s experiment have questioned the motivation of the participants, suggesting that they modified their view because they didn’t feel strongly about the issue and didn’t wish to create conflict. While I accept this as a reasonable concern when considering the wider applicability of Asch’s findings, I would contend that the vast majority of market research involves topics about which it would be unreasonable to expect the participants to be strongly motivated (indeed, one might worry if they were).
More recently, Dr. Gregory Burns took Asch’s work a stage further using brain scans. As before, a group was constructed with a number of people in on the experiment who had been primed to provide uniformly correct or incorrect answers, and an unwitting participant who was genuinely attempting to match the rotated geometric shapes that Burns was using in the “test.” What the brain imaging showed was that when people gave answers after being influenced by the group they weren’t making a conscious decision to go along with what they had heard, they had actually come to believe that what the group had claimed was true.
Where organizations can convince someone that lots of people think something is worth doing or having, they tend to do well. Being the “most popular” brand, being used by an impressively large number of people, or publishing lots of positive reviews or testimonials all provide social proof that we should think something is good.
Focus groups frequently produce a unified view on a subject when observation of the diverse nature of personal taste would make it apparent that such cohesion is extremely unlikely. The commonality is a result of the format of the research, rather than a true meeting of minds. Sony Ericsson discovered this to its cost when, alongside one of the main networks (carriers), it used focus groups to assess the appeal of a new handset, the W600, among American consumers. The results would determine whether the carrier took the handset and would be used to help forecast likely demand. Consumers weren’t particularly impressed and the carrier very nearly didn’t take the phone at all. In the end, the carrier decided that it would take it and forecast sales of just 5,000 units for the first quarter. When the handset hit the market, 10,000 were sold in the first fortnight and ten times the original estimate were sold in the first quarter. As one Sony Ericsson worker described it:
The amount of back orders and supply chain havoc this caused was a nightmare. Our final forecast would indicate around 75,000 pcs would have been sold if we’d been able to meet demand.
It only needed one or two people to express a negative view of the new phone and others followed. Whether they were primed by someone drawing their attention to one less than compelling feature, a strong familiarity and preference for a competing product, or a bad experience with one of Sony Ericsson’s products in the past, it just so happened that a negative consensus prevailed.
When it comes to considering a new product’s potential success in a market, it’s worth reflecting on the basic math of the matter. How many people in a group did Sony Ericsson need to like its new phone? Apple’s iPhone is rightly regarded as a huge success and yet, one year after it was launched, its share of the smartphone market was around 10%.8 Sony Ericsson’s phone was never aiming for such a dominant share of the market, but even if it was, it only required that one person in each group be sufficiently enthusiastic about it to decide to buy one in the subsequent 12 months.
Some market researchers might point out that it is wrong to allow a qualitative methodology to gauge the potential market for a product. However, such techniques are routinely used to screen initial ideas prior to wider development, at which point a product that one person present likes will be swiftly abandoned. In any event, the same basic mathematics applies with a quantitative approach: few companies are going to launch a product that only 5% of respondents say they will buy, and yet this may be all that is required (or all that can be achieved) in the first quarter of a new product’s life.
As long ago as 1961, James Stoner found that people changed their attitudes after discussing a subject with a small group of people.9 David Myers and Helmut Lamm conducted an extensive review of research into group discussions and found substantial evidence that they have a polarizing effect on the people who take part in them across a wide variety of situations. When people lean in one way or another when considering something individually, discussion with a group tends to amplify that opinion: by the end of the discussion a relatively minor preference or dislike will become a much stronger one. The reason for this is interesting. Research suggests that people enter a group discussion with a misconception of the position of the other participants; they tend to assume that they will have a stronger view than the group, and to have an ideal position that is more extreme than the one they’re prepared to voice. When the arguments raised in the group discussion support the initial position, people feel a need to shift their declared position in that direction.
In other words, we like to perceive ourselves as more in the socially preferred direction than the people we compare ourselves with. It seems that we run a constant mental scorecard assessing what the social average is, to make sure that we position ourselves just above it.
Interestingly, reading or listening to arguments generally produces less effect than actual participation in the discussion. It has been suggested that it’s the mental process of actively rehearsing or reformulating an argument that brings about the shift in position; through the process of expressing it to others, we convince ourselves of our own argument.10 Building influence by instigating debate around a subject or brand is what makes viral marketing and political blogging so effective. When the topic is skillfully released or the fuse of debate lit in the right way, the resulting impact can be dramatic.
The challenge for focus groups is compounded by the frequently humdrum subject matter on which they focus. It is one thing to feel that you will stand your ground in a debate with strangers about the death penalty or a solution to the problems in the Middle East, but the packaging of a breakfast cereal or your reaction to a new biscuit is not something most people are likely to feel passionate about. Research analyzing discussion content has shown that the largest shifts in attitudes occur where the subject matter is mundane and the argument put forward novel. Many focus groups will create the attitudes they report, rather than reflect views that are representative of people who haven’t taken part in the discussions facilitated for the research process.11
Often a moderator is seeking a sense of agreement from the group about a topic: how a brand is perceived or positioned, the merits of a new product, the appeal of an advertising campaign. If several people offer a similar opinion on a topic, it seems reasonable to surmise that that opinion is widely held and to report it as such. However, social psychologists have experimented and found that just one person repeating a point of view several times is very nearly as influential as several people making the same point independently.12 There is a real risk that someone listening to a group will be swayed not by an actual consensus, but by one repetitive voice. This should come as no surprise to companies that advertise: part of what makes a message effective is the number of times people get to hear it.
The way in which a point is made by another group member will also contribute dramatically to the extent to which it influences the others present. It’s not only the tone of voice that makes a point more commanding and influential, the nature of the point itself has been shown to change the extent to which people are influenced by it. When a statement is presented as something that could or should have been evident beforehand, people revise their attitudes and intentions as though they really did know whatever it was in advance.13
Just how persuasive one voice can be was illustrated in a fascinating article by Robert Harley, editor of The Absolute Sound, a high-end audiophile equipment and music review magazine. He described a blind audio test conducted by Swedish Radio, which wanted to establish if one of a number of low-bit-rate codecs (systems to compress and play back music) was good enough to replace FM broadcasting in Europe. A careful “double-blind, triple stimulus, hidden reference” test was constructed, in which 60 “expert” listeners would make more than 20,000 evaluations each, involving first listening to the unprocessed signal and then hearing two other versions of the same music; they were asked to identify which had been processed by the codec. Eventually, Swedish Radio had narrowed down its search to only two codecs, both of which were believed to be good enough to replace analog FM broadcasts. The test seemed extremely thorough, totally unbiased, and entirely fair.
However, after reaching its conclusions Swedish Radio sent a tape that had been compressed using the new codec to an acknowledged expert in digital audio, Bart Locanthi, who listened to the track knowing that it had been subject to compression. He immediately identified that the compression had introduced a distortion. When he told Swedish Radio what he had found its staff had no trouble hearing the same issue for themselves. In a few minutes he’d identified what all those blind tests had failed to appreciate.14 Irrespective of what all the people concerned had thought when listening in one set of circumstances, one voice had caused them to reevaluate their opinion.
What can work so effectively as a marketing technique undermines focus groups as an objective tool for exploring what people think. Knowing how influenced people can be by one person, particularly someone perceived as expert or a celebrity with whom they feel an affinity, provides a powerful mechanism for shaping perceptions of your brand. For example, in the early 1990s Pizza Hut wanted to give its brand a lift. It successfully used a number of different celebrities in its advertisements, from supermodels to racing drivers. As a result, people who had previously not considered visiting the restaurants changed their opinion and sales started to climb.
Conversely, in a focus group, one well argued, novel, or authoritatively expressed point can sway the entire outcome of the debate, even though those taking part feel certain that their views are their own and not a consequence of what they’ve heard in the group itself. Rarely do all the people present have the same degree of experience and commitment to a topic. The focus group format – whereby a question or topic is put to the group and the group is invited to respond – encourages the person with the strongest involvement and view and/or the most confident person present to speak first. The process fosters the emergence of a leader and one person’s opinion frequently influences the responses of others.
When a group of people make decisions jointly or work together to reach a conclusion about something, there is a risk of group-think, a phenomenon first explained in detail by American psychologist Irving Janis back in the 1970s. He realized that groups making decisions had the capacity to reach those decisions with insufficient critical analysis and with too much deference to the prevailing point of view. This, he explained, had contributed to a number of political fiascoes, such as the failure to prepare adequately for the Pearl Harbour attack, the Cuban Missile Crisis, and the attempted cover-up of the Watergate scandal. Both Lord Butler’s 2004 Review of Intelligence of Mass Destruction in the UK and Senator Pat Roberts’ intelligence review in the US cited group-think as a factor in the failure of the intelligence behind the decision to invade Iraq.
When one considers the elements that contribute to these infamous decisions, it’s easy to draw parallels with consumer focus groups. Janis identified eight symptoms of groupthink:
1 The illusion of invulnerability creates excessive optimism that encourages risk taking. It’s hard to conceive of a more invulnerable group than a consumer focus group. If they like the product, ad, or whatever is being tested, they’re under no compunction even to part with the few pounds the product would cost; in fact, under the UK Market Research Society’s code of conduct, it’s understandably important that the line between selling and research is kept very clear.
2 Collective rationalization – discounting warnings that might challenge the group’s assumptions, rather than reconsidering them. The goal of most group facilitators is to arrive at a consensus view (even if only unconsciously because it will make writing the report easier). Combine this with problems caused by priming and people’s tendency to want to stick with what they’ve said rather than risk being perceived as inconsistent (cognitive consistency theory), and the propensity for collective rationalization is certainly present.
3 Unquestioning belief in the morality of the group causes members to ignore the consequences of their actions. In my experience few respondents are self-critically judgmental in focus groups, being far too preoccupied with how they’re being perceived by the strangers that comprise the rest of the group. However, the problems are compounded by the fact that there are no consequences (to the respondents) of their focus group comments. While making respondents legally responsible for subsequent marketing failures is an interesting notion, it’s possible that research response rates would drop!
4 Stereotyping those outside who are opposed to the group in a derogatory way. Resistance to ideas from the company behind whatever is being investigated may hinder concepts that may be persuasive under other circumstances.
5 Direct pressure on dissenters – members are put under pressure not to express arguments that go against the group’s views. Again, not many people are willing to make a stand on principle for a consumer issue that they consider either insignificant or academic in the face of several people opining a contrary argument.
6 Self-censorship of ideas that deviate from the group consensus. I’ve already discussed the polarizing nature of groups. The desire for social cohesion works in a number of directions, and one would have to question how strongly a respondent would need to feel about a topic in a consumer focus group to introduce disharmony.
7 The illusion of unanimity among group members – people take silence as agreement. Rarely, if ever, does a research moderator actively canvas the opinions of everyone present, not least because it would break the flow of discussion and create a very unnatural interaction that would work against the primary goal of the focus group (i.e., to get people talking about the issue in question).
8 Self-appointed mind guards who shield the group from dissenting information.
Leaving aside point 8, which I would say only applies with the worst of focus group moderators, and point 2, which I concede is marginal, there is an argument to say that six of Janis’s eight symptoms of groupthink are present in focus groups. The propensity for a focus group to reach a “bad” conclusion is therefore significant.
It should go without saying after the issues raised in Chapter 3 that the context, whereby a number of consumers are placed together in a room to talk about something, bears little relationship to the environment in which a consumer’s response would normally occur. The artificial focus of discussing a consumer issue for a long period is a recipe for distortion and it’s all too easy for that focus to miss the point entirely, either because the consumer response isn’t determined at this level of mental processing, or simply because the abstract nature of the discussion means that something that seems irrelevant is glossed over. However, if all of this weren’t enough, the market research world has conspired to create a way of making focus groups even more artificial: the viewing facility.
If the arguments and supporting science provided up to this point have convinced you (as I hope they have) that the environment influences how people think and behave, then what follows will come as no surprise. Nevertheless, the widespread use of viewing facilities and the flagrant way in which they ignore human psychology mean that they merit inclusion in this book.
I accept that some of the emerging evidence on the extent to which the unconscious drives our behavior, and our inability to post-rationalize it accurately, isn’t intuitively apparent. Indeed, it can be quite uncomfortable discovering how extensive is the illusion of conscious will (to use Daniel Wegner’s phrase). However, I don’t believe that most of the inherent issues with viewing facilities should be so unapparent, and in many ways they serve to illustrate the extremes of artificiality that are widely accepted in market research. It’s as if someone sat down and thought: “OK, how can I find a way of finding out what people think that is as unrealistic as possible?”
For the uninitiated, viewing facilities are specially constructed to provide a room in which research can be conducted, usually with 10 or 12 comfortable chairs, a coffee table, and a television (for showing stimulus material such as advertisements). Almost one whole wall of the room is replaced by a two-way mirror (sometimes confusingly called a one-way mirror), on the other side of which is a second room in which a similar number of observers can watch the proceedings without being visible to the respondents. Sound is captured by microphones in the respondent room, and in almost all cases there are one or two video cameras to record the conversation.
So far so good, you may think. However, in order for the observers to remain invisible to the respondents, it is necessary to keep the respondent room very brightly lit, while the observers sit in semi-darkness. In addition, and just in case it wasn’t obvious from the cameras and microphones, respondents are told (normally verbally and through signs in the room) that they are being videoed and recorded. Which of us, hand on heart, honestly believes that we would be ourselves in such an environment? There are more than 150 such facilities in the UK and over 600 in the US, charging several hundred dollars per group. Some larger manufacturing organizations use them so frequently that they have invested in their own viewing facilities.
As is so often the case with market research, the convenience of hearing consumers say something is accepted irrespective of the probable reliability of what’s heard. Let’s explore why the results are likely to be unreliable.
Viewing facilities manage to create problems even before the first topic is raised for discussion. Long before respondents reach the artificial reality of the comfortable chair in the bright, mirrored room, they have to deal with arriving at the “facility.” If you spend any time observing people moving from one space to another, you will see that they change pace as they make a transition from one area to the next. They make unconscious adjustments from whatever was in their mind as they traveled – getting to the right destination, being on time, anticipating what will happen there – to sizing up their new surroundings. This is an element that good designers are adept at influencing to suit their goals; in retail environments it can help encourage customers to engage with more products more quickly and spend longer in the store. Most viewing facilities inadvertently create the feeling of entering a secret government compound! Because they are secured behind door-entry systems – it would be impractical to have a staffed entrance – respondents arrive and announce themselves over an intercom to an unseen receptionist. There then follows a series of stairs and/or corridors to reach the first holding area, or in the case of one manufacturer-owned facility, a seven-floor journey in an elevator.
To the best of my knowledge, no one has studied how such an entry influences what people say, but given the ways in which subtle environmental factors shape the way people respond, it is hard to conceive that they wouldn’t. Even if such an arrival is soon forgotten, the use of a mirror, having people watching, recording equipment, and light levels all demonstrably do change how people think.
The mirror, essential to the viewing experience, is a problem because it, too, changes how people think and behave. Most people are surprised (for me it’s unpleasantly) when they see themselves on video, because when we consciously look in a mirror we don’t usually see ourselves as we are. If we are psychologically healthy we will filter out the negatives and fix on the parts of ourselves we like; if not, we will either fixate on the parts we don’t like, or else like ourselves so much we become unbearable to others. This mechanism enables people to buy clothes that don’t suit them, and that they would view disparagingly on others, simply because they only look at one aspect when they try them on. The same mechanism, albeit to an extreme degree, enables an anorexic to see themselves and still think they need to be thinner.
On such a scale, and when seated in front of it, a viewing facility mirror works more like the unwelcome and unflattering video of oneself, providing occasional reflections to sensitize respondents to their unconscious expressions and mannerisms.
“Does this self-awareness necessarily change anything?” I hear you ask. The evidence suggests that it does. Having spent hours and hours observing consumers in retail places, Paco Underhill realized that when people pass shiny surfaces they slow down, but when there are too many reflective surfaces it becomes disorientating.15 When Arthur Beaman and his colleagues set up an experiment on Halloween to see if the presence of a mirror altered the amount of candy children took while believing that they were unobserved, they found that children who had a reflection of themselves at the time were far less likely to take more than they’d been instructed to.16 Another study found that the performance of people copying foreign text was improved by the presence of a mirror.17 Diener and Walborn reduced the proportion of students who cheated on a test by taking extra time at the end from 70% to just 7%; the only difference was the presence of a mirror.18
Given that people behave differently when they’re observing themselves, it’s no surprise that they do so when they know they’re being watched by others. Aside from the expected increase in self-awareness and self-consciousness, I have also found that it can increase defensiveness and aggression, particularly in men. Research by anesthesiologists found that patients who were told they were being observed changed their behavior during and after surgery, including the amount of pain they felt after surgery (people being observed felt less). The study concluded that the action of observing patients in clinical trials could invalidate the results of an experiment.19
The bright lighting in the respondent room that is necessary to allow the two-way mirror to function properly exerts its own influence on respondents. Countless research papers have studied the impact that the level and type of light have on people, including the Hawthorne research discussed overleaf.20
Essentially, our bodies are regulated by light levels and we have a circadian rhythm repeating approximately every 24 hours for cycles of sleep/wakefulness, body temperature, hormone production, and alertness. With such a fundamental link between our physiology and light, it isn’t surprising that studies have found differences in behavior and mood when light levels are varied. At the most simplistic level, more light is associated with greater engagement and, in work environments, higher output levels. For example, one study found that sensitivity to an unpleasant noise increased as the level of illumination was increased.21
The sum impact of the viewing facility environment is to induce a sense of hyper-consciousness, dramatically at odds with how most consumer behavior occurs. The transition through an entrance system that often involves consciously expressing the reason the respondent is there; the journey along unfamiliar corridors to a room in which respondents are told that they are being watched, filmed, and audio recorded; being observed by unseen strangers in a room that is brightly lit and where a large mirror reflects how others will be seeing them – there are not, from my perspective, a recipe for discovering consumer truths. Short of having a silent observer standing right next to each respondent, I struggle to conceive of a way of creating a research environment more at odds with consumer reality.
Typically, when people refer to the Hawthorne effect it is to support the view that a variation occurs in how people behave when they know they are being observed; in the experiments, researchers looked at how the productivity of workers changed under different environmental conditions (initially changes in levels of lighting). Since the original projects were conducted, several other studies have examined the research and suggested that the changes in productivity could be attributable to other variables. However, it is interesting to consider all the variables that could have accounted for the changes in production observed at the Hawthorne factory:22
Different light levels affected productivity.
Impromptu team work among the participants increased their effectiveness.
Being studied influenced the participants’ degree of motivation.
Feedback from the measurement of their work enhanced the skills of the participants.
In a viewing facility the first three of these influences will almost certainly be present, and there is a reasonable argument to say that the presence (and impact) of the moderator is analogous to the fourth.
When Sony Ericsson realized how much money it had lost as a result of the focus groups it ran on the W600 handset, it was forced to reconsider the use of such groups. Evidently up until that point the tacit belief had been that asking groups of consumers what they thought was a reliable gauge of something. While a few companies have moved away from them, virtually every research agency that describes itself as expert in qualitative methods offers focus groups as a legitimate market research tool. Consequently, numerous products, services, and pieces of marketing communication make it to market or are rejected because of feedback from groups of consumers interviewed collectively. There remains a notion that prompting a discussion with other consumers present will reveal more thoughts and feelings than might otherwise be identified, and of course such groups are quick and convenient to conduct.
Our own illusions of conscious control no doubt contribute to the perspective that we know we could be asked to discuss our opinion on something with a group of strangers and would ruthlessly stick to our beliefs; after all, we know what we think.
Much of the criticism of focus groups – and there has been a lot – involves issues with respondent recruitment. Are the people who are willing to participate necessarily representative of the market as a whole? Are they “professional respondents” who take part in such discussions with great frequency for the payment provided to participants? Others have pointed out that for a supposedly in-depth research tool the amount of discussion time per participant can be insubstantial: 12 respondents taking part in a standard 90-minute group discussion would have an average of 7.5 minutes of airtime each, less when any introductions or warm-up discussions are included.
However, these are moot points if any group inherently distorts what respondents think and say. When people think differently in groups, need to feel more strongly about a subject to stand up for it in a group situation, unconsciously allow their attitudes and what they focus on to be altered, become more vehement and are primed by what they hear other people say, it doesn’t matter how well the respondents are recruited or how expertly the group is moderated.
There are many reasons to believe that the information emanating from focus groups is a by-product of the group dynamics through which it has been collected, far more than that it is an accurate reflection of the consumer response of those same people in the real world. Given the relatively mundane nature of consumer research topics, any reasonable comment or reaction raised by a respondent in a group discussion is likely to face little opposition from the other people present. In psychological terms, the very fact that it has been voiced first will give it additional impact.
All of these problems are compounded by the issues I’ve discussed in previous chapters regarding the perils of artificial introspection and the potential impact of a moderator’s behavioral style unconsciously encouraging a particular type of response. The need to encourage respondents to feel comfortable and to open up is a recipe for the “please you” behavior driver, which encourages an upbeat response.
If accurate consumer insights are the objective, then by far the simplest “solution” is to avoid focus groups altogether. The only theoretical place for them would be if researchers believed that they could simulate the complex social influence that occurs in human groups. In such circumstances, hearing the group interaction take place may be illuminating. However, this isn’t a justification for recruiting a group of people who don’t know one another and moderating a discussion between them. Instead, the aim should be to take an existing social group, put them in the most accurate context possible, subtly release the initiative in their presence (and in the presence of as much of its competition as possible), and stand back and see what happens. Even then, the problems of the tacit leader’s views dominating the discussion or inadvertent priming taking it in a particular direction mean that it can’t be relied on.
When the Post Office decided to rebrand itself to create an identity that would reflect all of the things it did as a business and would work internationally, it reportedly spent more than £2 million on the process. It stated that the new name, Consignia, was “extremely well received in customer research.” Nevertheless, it was lambasted by the press; it seemed to be a word, but no one was sure what it meant. Worse still, UK consumers had a long and broadly affectionate relationship with the Post Office and felt like they were losing something fundamental. As you know by now, loss aversion is a powerful motivating force. Understanding the true nature of consumer behavior would have been far more valuable than whatever was wasted on market research.23
By all means, if the subject of interest is what people talk about when placed in a brightly lit room while being watched by a hidden group of strangers, use a viewing facility. Otherwise, using one is unlikely to be beneficial.
For the most part, the best way to consider groups is in terms of the role that social influence has on consumer behavior. As I’ve discussed in previous chapters, people are hugely susceptible to priming and to social proof. The appeal of a product or new brand name can be hugely influenced by who says it’s good or who is seen to be using it, irrespective of its apparent merits when considered consciously.
The truth is that consumer behavior, just like all human behavior, is very much a by-product of the wider social group. Understanding how the right combination of context and group influence can lead to commercial success (or the absence of it) is perhaps as close to defining a magic formula for marketing new products as it is possible to get. However, such studying of group interactions has to take place unobtrusively in the native habitat of the consumer, or else be considered in the broader context of the way we behave in relation to one another. It can’t be recreated artificially in a couple of hours with a moderator in a strange room.
Going back to the example of New Coke that I discussed previously, the group influence effect was another significant factor in Coca-Cola’s undoing. Inevitably, Coke’s customers didn’t carry out an independent, balanced assessment of the new recipe’s qualities as the research respondents had done: they heard the media, friends, and colleagues talking about it. The sentiment started to spread that the removal of the old recipe was somehow undermining the essence of America; one newspaper columnist compared changing the drink’s formulation to the removal of President Roosevelt’s face from Mount Rushmore.24 Demonstrating admirable, if misguided, faith, Coke’s executives continued to take comfort from the surveys they were conducting that told them people liked the new flavor. Instead, people were leaning toward loss aversion and copying the sentiment of those around them who were saying that something significant was being taken away. Ultimately, the group decided that it didn’t like the idea of what Coca-Cola was doing, and that mattered far more than what they as individuals might have thought of the actual taste.
Understanding the reasons that focus groups can’t help shed light on what people really think or do in itself provides a checklist against which to consider an initiative that you might, misguidedly, have asked such a group to evaluate:
Focus groups don’t work because… |
About your initiative… |
People copy each other. |
Will it be visible? |
People change their mind to fit in. |
Will people feel they must have it? |
People agree with the majority. |
Can you win over enough people early on? |
Discussion changes attitudes. |
Can you get people talking about it? |
One voice can sway the group. |
Can you get experts or ambassadors on side? |
One of the most popular applications for focus groups is exploring what people want and whether they like a product, policy, or piece of communication. This raises the question: Is there a reliable way of asking people what they would like in the future?