1. People don't think about how they feel

Emotions drive most of our decision-making, with the brain processing emotions almost instantaneously. Dr Paul Ekman is the world's leading scholar in emotions and was a consultant on the Pixar film Inside Out. He describes six primary emotions: fear, anger, disgust, sadness, surprise and joy. Emotions are processed quickly as part of our survival mechanism. If you begin to swallow sour milk, you'll feel disgusted quickly enough to spit it out. If you see an attractive person, you'll feel it before you think ‘that person is cute'. But we can't process and articulate these emotions as we feel them.

We generally feel the emotion and then fill in the cognitive gap. For example, if there's an ‘instant attraction' between you and someone you've just met, you'll look for evidence to confirm that emotion. You'll say, ‘Wow, we like the same movies, isn't that nice?', or ‘Wow, we like different movies, isn't that nice?'. Applying this to the world of brands, we feel emotions when viewing advertising and beautiful products, but it's hard to articulate these emotions. This is especially the case in focus groups. If they can't describe how they feel, participants fill in the gaps with rational explanations. As Jonathan Haidt has said:

The conscious, rational brain is not the Oval Office; it isn't there making executive decisions in our minds. It is more like the Press Office, issuing explanations for decisions we have already taken.

When filling out a survey or participating in a focus group, credence is given to the rational and explainable because that's what the consumer can articulate. Unfortunately, it's not how we make decisions.

2. People don't say what they think

Our self-concept encapsulates who we think we are — physically, emotionally, socially and spiritually. It drives how we present ourselves, how we act and what we say. Most of us think our self-concept (how we like to see ourselves) is how we are (actual self). If this was the case, then Actual Self (AS) = Ideal Self (IS). Unfortunately, there is often a gap between who we are and who we aspire to be. A focus group participant or someone completing a survey and questionnaire is more likely to answer according to their Ideal Self rather than their Actual Self.³

In psychological psychometric testing, questionnaires build in ‘lie scales' to ascertain how much a subject is lying. However, there isn't time or money (or motivation) to develop sensitive tools for focus groups, or lie scales in quantitative measurement tools. Has a market research company researched the accuracy of market research? Have participants been asked if they have answered questions randomly to finish faster? How many people select only one brand because it means the questionnaire will be shorter? This information would make for more meaningful market research. In focus groups, it's common and normal for participants to say they don't watch TV, they read The Guardian newspaper and actively follow world politics. Most never admit an emotional rather than rational ad persuaded them to make a purchase. People present their Ideal Self, not the Actual Self. They don't say what they think, even to themselves, and especially not to a researcher.

3. People don't do what they say

Market research attempts to find out what people are planning to do and asks questions along the lines of, ‘Is this more likely to make you buy X?' or ‘Will you do X as requested by the communications?' Research conducted in this way is often inaccurate because consumers are bad at predicting their behaviour. If people did what they said they were going to do, intention to purchase would define every future consumer market. But it's quite a poor predictor of actual behaviour.

Focus groups are still a cornerstone of market research. They are usually conducted offline, with people invited to a venue with a one-way mirror for a two-hour conversation about a brand they may never think about again in their lives. They watch and respond to ideas and ads. Typical questions include, ‘Who do you think this is targeting?', ‘Do you think this would make you buy the brand?', ‘Would you be more likely to consider this brand if you saw these ads?', ‘Would you use this new product?' and ‘Would you talk about this new product with your friends?' Up to a hundred or so questions are fired at these people as they project into the future and attempt to guess their anticipated behaviour.

There are many other social inconsistencies and biases to deal with, such as domineering respondents, or the type of respondent who'll drive an hour each way to attend a group on a cold, wet evening for $50, or the fact that the same people can end up attending hundreds of groups. In the industry, they are called groupies. The premise that focus group attendees can accurately anticipate the impact of messaging is the most significant inaccuracy of all.

A jaded senior qualitative market researcher once said to me, ‘Adam, the only thing focus groups are useful for is for designing focus group rooms. Everything else is a waste of time.' When people are asked about the here and now they are more reliable. But who knows how you'll behave in the future? If you ask people if they like the room in which a focus group is taking place, they can answer honestly. But if you ask them if they're more likely to buy a particular type of insurance in the future if they view this specific ad, they wouldn't have a clue.

The gap between what people say in market research and what they actually do was identified as early as the 1970s. If intentions failed to match actions back then, the difference is even greater now because of increased choices, new technology and overall complexity. The ‘purchase funnel' is far less a funnel and more a pinball machine, pinging us all over the place. It's not surprising that people's stated intentions fall by the wayside.

The paper ‘When do purchase intentions predict sales?' in the International Journal of Forecasting found that many elements need to be in place for this to happen. The results indicate that intentions are more correlated with purchases⁴:

• for existing products, not new ones
• if the product is durable
• if the products are to be consumed/bought soon (not in the future)
• if the intention is for something particular, not vague.

There were other issues with the link between intention and sales, and these were primarily around measurement. What constitutes an intention? How do you link the intent to purchase? What are you comparing this data against?

TV viewing reveals the gap between what people say they do and what they do. Over the years, there have been predictions about the death of linear TV, with viewers claiming to watch far more streaming than broadcast TV. Poor research design saw a series of outlandish claims from the digital disruptors that simply never materialised. This happened partly because skewed samples of context-sensitive questions combined with misguided analysis to generate misleading conclusions. But it's also because respondents are much more likely to remember new experiences (streaming), and are unable to remember the number of hours they spent watching traditional broadcast TV. They said they watched more hours of on-demand and online TV than broadcast TV, but it is a far cry from their actual behaviour.

Here are Rory's thoughts on consumer insights and market research:

My general advice here is that you should spend a great deal of time, money and effort hearing consumers. But you shouldn't spend too much time listening to them. A very good principle in problem-solving is that you should not spend a minute trying to solve a problem until you have spent a day trying to redefine it. And so hearing the consumer is a very good idea. It's a good idea because consumers are alert to problems to which businesses may be entirely deaf. That's important. The fact that people are upset about something really matters. But listening to customers can be highly misleading. They don't always know what they are upset about. And they certainly can't always explain what the solution may be. In post-rationalising their emotional state, consumers have already framed the problem in a rational context. But this may bear very little relation to the real reason that gave rise to their feelings, and so may be almost useless in helping you define the real problem to be solved.

To demonstrate the difference between listening and hearing, Rory references a scene from the terrible 1990s movie White Men Can't Jump. Wesley Snipes' character, Sidney Deane, and Woody Harrelson's character, Billy Hoyle, start arguing in the car when ‘Purple Haze' by Jimi Hendrix plays on the cassette player (yes, times have changed). Deane says white people can listen to Hendrix but can't hear it because Hendrix was a black artist making music for black people. He says, ‘Look, man, you can listen to Jimi, but you can't hear him. There's a difference. Just because you're listening to him doesn't mean you're hearing him.' The takeaway: don't listen to what people say, but hear what's behind what they say. Hear the real problem for which they want a solution. In the next two chapters, I'll explain that even if you hear the consumer, delivering on what they want may remove value from your brand and banish it to homogeneity. But more on that later. Rory's insights (and referencing of the movie White Men Can't Jump) helped me to develop the title of this book.

They're not the only ones who lie

It's not just consumers who can't be trusted. It's not only in the frivolous world of advertising that people behave in ways that appear entirely irrational. As I mentioned at the beginning of this chapter, I witnessed it with prisoners in the legal system. However, it also applies to those who wield the gavel. My father was a judge, as was my father-in-law. I know that judges, even more than advertising types, believe their hype. But in their case, the hype is that they're rational, logical, humble and conservative. They don't take risks, and they only act on the evidence. Judges are the very embodiment of reason and reasoning in our society. However, I've been around enough of them to know this isn't always the case.

A fascinating study in 2011 found judges handed out harsher sentences before lunch.⁶ The reason is that they're hungry and, after a busy morning, their mental resources are depleted. After they have a bite to eat and feel replenished, the judgements return to the norm. They are human, after all. The study shows that influences outside of our control cause our behaviour: ‘Our findings suggest that judicial rulings can be swayed by extraneous variables that should have no bearing on legal decisions.'

If you asked a judge if they were more lenient because they were in a good mood with a full tummy, they would reject the notion out of hand. Judges sees themselves as rational human beings above such trivialities. It's a lovely example of how inaccurate our ‘judgements' can be and how open to suggestion and influence we are. If we can't trust a judge to make an accurate decision, how on earth can we trust the consumer to predict the future?

Beating the lie

So market research is inaccurate. Everyone knows it is. It isn't like ‘science', which is objective. There is generally vested interest in the answers, and often the research company is trying to protect their commercial interests and seem relevant despite woefully out-of-date methodologies. Or it could be the commissioning business just wants to find supporting evidence to go ahead with an idea; this is more akin to gathering confirmatory evidence than research.

The lies people tell themselves get in the way of brand-building marketing. Therefore, it's a question of how to properly conduct research. I'm imploring you to put brand thinking ahead of consumer thinking, but I don't want you to stop using research completely. I asked one of the best marketing researchers I know, Wiemer Snijders, what he suggests for those who need to do research and better understand the customer.

Wiemer told me a story about buying a new car. The salesperson told him some of the car's features were created after consulting with car users. People suggested storage for a laptop and a hook for groceries so they wouldn't topple over. Wiemer was baffled that the salesperson seemed to think he had just offered him a critical insider tip, and said,

Be careful when it comes to listening to consumers. Not everything people say they want is what they value enough to buy when it comes to it. We are selling; we have to filter what the end-user says and the data we collect with commercial sense.

Wiemer says three things have helped him determine whether to pay attention to consumer research — or ignore it. The first is timely answers.

People are terrible at predicting what they'll do in the future, so we need to be wary of this type of information. If we do ask people about their future buying behaviour, we should refer to an appropriate time frame — which will vary by category.

If asking people how likely it is they'll switch to a different insurance provider, ask in the month before renewal rather than six months ahead. People will happily offer an inaccurate opinion if they have no skin in the game. It's just pretend and hypothetical to them. What does it matter if they lead you down the garden path? If possible, make it real for them. If you're researching the type of drink they want, then also conduct research when they're actually thirsty and about to buy a drink.

Secondly, randomness in samples can be a trap. A/B testing may be a useful way of screening potential improvements; it measures what people prefer. However, many tests don't last long enough to match the full purchase cycle or offer enough data points:

It means running the risk of misinterpreting short-term results because you've potentially failed to measure the regression to the norm. It's a bit like walking up a hill and stopping before reaching the top. You fail to see what more there is and can miss the real conclusion.

Thirdly, size matters. Many so-called insights from research ask people what they think or feel, without relating it to actual buying behaviour:

This means that if we don't control for brand size or user type in brand image tracking data, we will likely miss the primary learning. Brands may, for example, seem more differentiated than they indeed are, leading managers to do things that focus too narrowly and may hurt sales and profitability.

What people say and do can be two very different things. And did Wiemer buy the car with the add-ons? No, he bought the standard one. My take-home message: where possible, watch people rather than ask them about predicted behaviour.

These are the typical methods in market research:

• In-depth interviews. A person is interviewed one-on-one about a particular issue or brand.
• Focus groups. A small group of people discuss an issue or brand. The conversation follows a free discussion guide, moderated by a researcher.
• Surveys. Questions and prompts that ask a predetermined set of people how they think and feel about a particular brand or issue.
• Questionnaires. People respond to a series of questions, either online or over the phone about a brand or issue.

But as this chapter has illustrated, each method has pros and cons associated with it. Even when they receive money, people lie and don't answer questions honestly. The research is laden with inaccuracies, costs and quality issues, and, in many cases, the costs far exceed the benefits. Research is now shifting into:

• behavioural research
• participatory research
• artificial intelligence.

Behavioural research looks at how people interact with the product or service in their environment. Technology allows the researcher to better understand how the product, service or brand could be adapted to meet the needs of the consumer. It involves spending time with consumers and watching them interact and use the product in situ. Companies such as Watch Me Think ask people to film themselves using the product and discuss its pros and cons. They're not interviewed, they simply do a talking review. Lookback.com does a similar thing with online and digital brands. You can watch people using your product in real life situations.

With participatory research marketers are attempting to close the gap between the producer and the consumer through what's known as ‘end-user innovation'. Give consumers examples of your product, and they'll innovate, customise or improve it in some way until it matches their needs.

‘Artificial intelligence' will seem far less magical and far more practical if you replace the term ‘artificial intelligence' with ‘computer program'. Artificial intelligence helps to process the swarms and lakes of data to develop insights into how people behave, think and feel about various brands and issues.

It's essential that strong marketing has insight embedded throughout it. But the insight needs to serve the brand and not the other way around. If your insights rely on revealed truths rather than people anticipating their future behaviour, it'll reduce your risk of insights corrupting the entire marketing process.