Q: ‘Is authenticity important to you in Italian food?’
A: ‘Yes, definitely. All Italian food should be authentic.’
Q: ‘So what’s your favourite Italian dish?’
A: ‘Hawaiian pizza’
I suppose if people were consistent, I’d be out of a job.
What people say and what they do are often different. People are inconsistent, frequently irrational and their underlying motivations are difficult to uncover.
There’s lots of reasons why attitudes differ from behaviour. When talking about past events might be because people want to be viewed favourably by others, but a lot of the time it’s because people misremember. More fundamentally we are all subject to biases and are prone to make decisions emotionally and then rationalise them after the event. When talking about future events people overestimate the control they have over their own behaviour. Academics in the world of psychology have been struggling with this for years: dominant theories (e.g. theory of planned behaviour) are limited, with questionable predictive validity.
This is why we tend to look at consumers from a number of different angles: understanding attitudes (what people say), behaviour (what people do & their emotions in situ) and the market context helps us to solve market research problems confidently. The model below helps summarise this: It’s a triangle and has A-B-C on it. There’s also some methods listed as a prompt.
By way of example, if we wanted to understand a retail category we might combine some way of getting at attitudes (e.g. survey, focus group), a way of getting at behaviour (e.g. EPOS data, accompanied shop) and a look at the context (e.g. a market audit of competitor stores).
I tend to start with C – context – C-A-B.
Context here I define as how the market presents itself to the consumer.
You can have a good go at answering contextual questions before you’ve gone near a respondent; indeed, the list is a good prompt at proposal stage which can then be used as hypotheses to test in the project.
Taking an issue like “How can we get shoppers to purchase more ready meals at supermarket x?” as an example, I’d argue you cannot give a full answer without accounting for:
1. What the media says (e.g. ‘ready meals are bad for you – full of salt, sugar and fat’)
2. What people believe (e.g. ‘I don’t have time to scratch cook every night’)
3. What is the market context (e.g. the retailer’s 3 for £5 ongoing promotion across our retailer is being overtaken by a BOGOF in a competitor)
4. What the buying process is (e.g. majority of spend comes via habitual weekly purchasing)
5. What ‘arousal’ state do people buy in (e.g. time pressured, uninspired)
The issues which behavioural economics and choice architecture focus on fit nicely here. The former might include decision making biases and habitual error and the latter incentives and defaults. These prompts can be practical tools for our toolkit, and they might guide how decide which methods we use to understand attitudes and behaviour.
Examining attitudes brings us onto the standard qual / quant debates which I won’t enter into, except to say direct questioning is sufficient in many circumstances where you’ve got a simple issue.
With more complex issues adding methods to understand behaviour helps us reach a more nuanced understanding, and falsify our hypotheses. This gets to the truth when in a low involvement category (e.g. EPOS data shows penetration of washing powder A is highest despite people saying they buy B – because they’re in a similar pack) in situations where people have lower control of their behaviour than they assume (e.g. aiming to keep your alcohol consumption down but not managing it) or brings hyped situations back down to earth (e.g. respondents being intoxicated by a new product in a group vs. seeing it on the shelf in the cold light of day).
In summary, most of the issues market research seeks to answer are complex: overlapping several methods that cover attitudes, behaviour and context give us the best chance of getting us to a solution.