In Thinking Fast and Slow, Daniel Kahneman provides a comprehensive and highly readable overview of how we all make decisions. In doing so he completely buries the already tarnished view of human behaviour as ‘homo economicus’ (rational decision maker) and demonstrates clearly and precisely that we are far from paragons of reason.
This is a book that every market researcher should read. The industry is still wedded to the traditional view that we are all rational, seek to maximize utility and minimize costs, make decisions in social isolation and have stable preferences. Many market research practices are based on the assumption that consumers have accurate memories, understand their own choices and behaviour and can predict their future choices. Unfortunately, reported behaviour, attitudes, beliefs and intentions are rarely predictive of future choices, consumers are influenced by those around them, their preferences change depending on context, and they are ‘cognitive misers’ and seek to make decisions as simply and quickly as we can.
While many market research practitioners have been moving in the direction of using observation, behavioural data and non-verbal questioning to address these issues, they are still in the fringe of market research, and the majority of clients continue to rely on verbal question and answer, attitudes and opinions and long verbal questionnaires to provide insight into their customers’ choices.
Here are three key ideas from behavioural economics that should be incorporated into all research studies.
- We are terrible witnesses to our own behaviour, and standard question and answer approaches encourage us to create fictions to explain our choices.
- Implicit knowledge is what makes humans really clever, and this can only be accessed through indirect (non-verbal) approaches.
- Behaviour depends far more on context than it does on stable preferences, personal beliefs and cost-benefit analysis.
Firstly, question and answer approaches assume that the respondent knows the answer and all the evidence shows that the vast majority of day-to-day decision-making relies on implicit knowledge (heuristics or mental shortcuts) that are outside our conscious mind. When asked to explain behaviour we are compelled to find answers (post-rationalization), but those answers are based on what our conscious mind thinks is a plausible explanation (and likely socially acceptable and maintaining a positive self image).
Many studies (e.g., with split brain patients or looking at memories over time) have shown that we invent answers that fit with our perceived reality and will ignore the truth even when presented with it. One of the main personal factors studied in behavioural science is how we like to maintain a positive self-image (e.g., cognitive dissonance).
More worryingly for research there is strong empirical evidence that asking questions interferes with memory of experiences and can alter the facts. For example, in criminal psychology it is well understood that asking witnesses to describe a criminal has a detrimental effect on any later identification and that the process of question and answer influences mental processing and changes the way the event is remembered. This is more than a question order effect, as any directed or detailed questioning forces us to reframe memories in specific ways before they are ‘re-remembered’.
Put simply, asking questions changes the way we think about any experience, product or service. So we should measure behaviour and not opinions, through behavioural data sets, facial recognition (i.e., emotional behaviour), click throughs, and by using more experimental approaches to designing research studies.
Secondly, most of our decision-making happens ‘under the hood’ (or in System 1 as Kahneman refers to it). Our implicit knowledge far outweighs our explicit knowledge, just as the richness of our experience through the senses cannot be adequately described through words (although some great poets have made good attempts).
Humans can mentally process a complicated image in fractions of the time that we can process a few words on a page, comparing it to our data bank of decades and making instantaneous predictions of what is happening in the world around us and what will happen next. By comparison, our verbal reasoning (conscious processing of System 2) is awkward, slow, verbal and effortful (although much more logical) and is only used sparingly (it’s much easier to rely on intuition than to have to figure out every problem anew).
As our implicit knowledge is mostly non-verbal (mostly based on the senses) it makes sense to access this knowledge via non-verbal approaches (projective techniques, imagery, less direct questioning). These approaches rely on our mind’s ability to make connections and associations that are buried deep in our memories, and cannot be accessed in logical verbal-based reasoning.
Thirdly, context (social and environment) has a huge impact on behaviour. Of special importance for researchers are copying and imitation in buying behaviour, the impact of context on memory and the importance of the frame of reference.
When I go drinking with friends, whatever I might say I prefer, my choice depends more on what other people buy. Most research relies on the individual and the assumption that their behaviour is their own choice with no social influence. This is not true as Mark Earls and others have demonstrated.
Many studies have shown that our memory relies heavily on context to provide cues as to which past experiences are relevant. In crime witness studies, taking someone back to the scene of the crime is done because recall of the event is much more accurate than in any other environment. When did you last take your respondents to a bar to ask them what drinks brands they can recall?
Framing has always been a consideration for researchers, which is right. Our perception and mental processing relies on relative differences rather than absolute information (which is why a certain colour still looks the same even when the sun has set). More mundanely, choices such as price depend entirely on the order of presentation and the highest and lowest cases (which is why price sensitivity responses are always in the middle of the range provided to respondents).
How can research take on the challenges from behavioural economics and adapt to the reality of implicit knowledge, unconscious decision-making and contextualized behaviour?
Here are three key lessons:
- Avoid asking questions and measure actual behaviour either directly or via proxies (such as facial expression).
- Move away from question and answer approaches to non-verbal research tools.
- Focus on the context of behaviour through observation, in-situ testing and framing questions with the right choice set.
And please read Daniel Kahneman too!