Consumer research and Behavioural Economics are not always the easiest bed fellows. Through Behavioural Economics we now understand why people are often unaware of what influences their behaviour. So, by definition, as they are not aware of it, they are often incapable of accurately reporting these influences. However a core assumption of market research is that people do have complete insight into the working content of their minds, and hence predicting the future based on the responses to direct questions will be accurate.
In qualitative research (focus groups, interviews etc) good behavioural researchers can read respondents and get a feel of what their behaviour might be. However it takes a brave researcher to confidently predict respondents will do something when they have not said they would. But often the opinion of good qualitative researchers is not enough for clients to make big decisions, and they look to quantitative research to get ‘hard evidence’ that predicts the way people will behave. In other words, they feel the need to get ticks in boxes, or these days clicks on web based surveys, that categorically show the majority of a large number of people say they will do one thing or another. However, if Behavioural Economics is right, this approach may not always lead to an accurate prediction.
Put simply, people may believe they are reporting accurately on surveys, but are they unreliable witnesses to their own behaviour?
Certainly, recent opinion poll failings on huge issues such as Brexit, the 2015 election and Scottish independence, all point to this. And if people fail to predict - when asked - how they are going to act on major issues, how reliable are they going to be in reporting on the things that matter less, such as whether a new variant of cereal will be eaten by their children or whether a change in pack design will make it more appealing? The truth is no doubt somewhere in between, where people are able to accurately report some things about their opinions and behaviour but are, in fact, rather poor at reporting others.
But this still puts research in a tricky position, as it is commonly asked to produce hard evidence that allow marketers to make the important business decisions, while acknowledging that Behavioural Economics suggests the responses it gathers may not always be reliable as they wished. Clearly all ways in which to maximise the accuracy of responses in research need to be reconsidered.
So what is the problem?
Well it may sound like a daft question, but it is a fundamental one. What exactly is a brain for? This is something I often ask people and the answer that usually comes back is to think, or reason, or sometimes to feel. Modern thinking from psychological and neuroscience disciplines is somewhat different.
One undeniable truth about the brain is that it is a product of evolution. As such its core function, by definition, was to give us an evolutionary advantage. Make us better at being what made us the top mammal, or more accurately, what made us the best hunter gatherer (as in evolutionary terms not enough time passed for us to extensively evolve beyond that). Taking aside some of the brain’s core biological / autonomic functions (such as breathing, hunger etc), the received wisdom is that the brain’s primary job is to predict what happens next.
What this means is that in any given scenario or situation, it absorbs a huge amount of information in any situation we are in, then based on that, it automatically produces information in terms of feelings and emotions that give us information about what might occur in the next few seconds. Say, for example, one of our hominid ancestors in the plains of Africa, heard a rustle in the bushes. From previous experience and acquired world knowledge, it sounded like it was large, and could be a potential threat.
To allow them to survive, a quick automatic response, one that does not involve thinking or consideration is by far the most likely to allow our ancestor to survive. As such an instant feeling of ‘fear’ and the appropriate response of ‘avoid’ would lead to a much greater chance of survival. If our ancestors were actually designed to ‘think’ and ‘consider’ by bringing previous information to our mind and consciously considering it, any response would have been slow and ponderous, and their ability to end up as lunch would have been infinitely more likely than ending up as a parent.
The other factor of course is that evolution only allows for efficiency, that is the maximum output for minimum (calorific) input. Although the brain on average weighs 3 pounds (about 2% of the total body mass) it uses 20% of the energy we consume. If we think or concentrate for a long time we feel tired, with good reason; thinking uses a surprising amount of energy.
This is why the general belief amongst psychologist and neuroscientists is that the thing that gave us the biggest evolutionary advantage, was for our brains to evolve the ability to quickly, and accurately (enough), predict immediate responses to our environment. The implication of this ‘what happens next’ function means that the brain is functioning best (i.e. doing what it is primarily designed to) when it is immersed in a situation and is automatically generating feelings about it.
The implication for research, and its tradition of direct questioning of course is that the approach is not getting the brain to do its core function. Direct questions prompt the brain to have to consciously mentally reconstruct something, either a past event or a future situation. i.e. remember something that happened or imagine a new scenario. Although the brain can do this and it can be accurate, the quality of this ‘emulated’ experience is dramatically reduced.
Put simply, even though it is the most complex thing we will ever experience, it would need to be many times bigger, and vastly more complicated to actually recreate the totality of any experience. As such, the conscious mental reconstructions we end up with are summaries of real experiences, and because of this they can by definition not be completely accurate. So in research terms asking people questions, getting them to consciously reconstruct future or past events will only get a response that is derived from a reduced level of information, hence this is one of the reasons that what people say is not always what they do, and why predicting behaviour from direct questioning cannot always be entirely accurate.
So how can we maximise accuracy of responses?
Of course the only reliable way to understand behaviour is to observe it. Through ethnography we can see what people really do, how they behave and understand this without the danger of self-report clouding the answer. But such approaches are expensive and time intensive, and there are always issues over whether observing is actually changing the behaviour. There is no doubt that as life logging technology improves, it will give a fuller picture of behaviour and it is likely that this technology will become a staple approach in the research industry. However, we can only observe what currently exists. Often research is asked to provide commentary on scenarios that don’t yet exist; new adverts, point of sale, pack designs and product ideas and extensions, that cannot be observed in the real world. We believe answers to these kinds of research challenges can be found in the virtual world.
Catching the thing you didn’t drop
We had a client in our supermarket. He picked up a box of cereal and it slipped. His whole body instantly reacted as his left had shot out to grab the cereal, to stop it falling, accompanied by a small ‘Agh!!’. Nothing unusual here, except he had not actually dropped anything…
The supermarket in question was virtual and he was using a headset and controllers to interact with it. He’d picked up a box of Coco Pops, his finger slipped on the controller and the virtual representation of the world he was in represented it falling from the shelf to the floor.
Even though none of it existed, his instant autonomic response was to react as if it had really happened. What was more interesting was that he muttered a little and then picked it up with the controller and put it back on the shelf. He did not even realise that his left hand had tried to catch something that wasn’t there. It was only afterwards when we showed him the video of this that he saw the humour, as well as how his brain had been tricked into believing something had happened that never did.
Putting someone in a virtual world makes the most of the brain when it is performing its primary function. Being immersed in a situation and then responding automatically in a way that will predict what happens next. In the case above, if a thing drops you try and catch it. It’s what the brain unthinkingly does. Brains just happily respond to the virtual environment the same as if it were real.
Just as an aside, I will admit myself that I have been demonstrating virtual supermarkets for quite a while and I still find myself walking round the basket that we included on the floor for people to pick up and put things in. Even though I know it’s not there, my brain automatically still tells me to avoid it in case I kick it or trip over. It is all of these ‘brain stem’ based reactions, getting the brain to do what it was designed to do, that is the key to increasing the accuracy of responses in research.
The new virtual world that awaits
The debate Behavioural Economics haves started is the extent to which we are aware of the things that influence our behaviour. This has prompted other observations, specifically in research, that if we are not aware of those things, we are by definition not going to able to report them in a survey or to a researcher. Not because we don’t want to but simply because we don’t know, and because it involves our brains doing something that is not its primary function, namely consciously emulating future and past events. Virtual reality provides a solution to this.
The most common question we get asked is “isn’t it unnatural?” or “don’t people feel odd in it?" My response is that current convention of asking direct questions or getting people to talk about something they would not normally talk about in a group of people they have never met, is far less natural. Virtual Reality actually allows us to do research where the brain is functioning as evolution designed it to, experiencing an environment, and responding to it as if it is real. It is only through convention that we are used to surveys and focus groups as providing answers.
Direct questioning either in qual or quant has been an optimal tool in the researchers tool box up to now, but new VR and AR technology allows us to explore a new world of research that is likely to yield far more predictive research to these current approaches. The potential for VR based research is vast. From the creation of virtual environments through to clever combinations of 360 video and CGI allow us to put new shop fixtures, adverts and packs in stores that have been filmed. Although this is predominantly still a qualitative tool, we have collaborated with a research panel provider to produce the UK’s first VR panel, opening up the prospect of VR on a quant scale.
With companies like Google, Microsoft, Apple, Facebook, Samsung and HTC betting on this being the next paradigm shift, the future of commercial behavioural research could well see some astonishing changes in the next three years. As penetration of the technology expands, virtual quant research can be conducted to a depth that up to now can only be imagined. Respondents can, in their own home, take part in car clinics run on a huge quant scale with people being able to sit in, experience and change aspects of the design so brands can get real time feedback. Respondents in their living room will be able to walk around new store layouts and shop designs.
Already respondents can see new products on shelves or new pack designs, in the context they will see them in store, to see if they are appealing. All of this will allow a whole level of behavioural research that at the moment can only be imagined.
Exciting times are ahead.