FEATURE1 March 2011

How will behavioural economics change research?

Behavioural economics shines a light on the factors that influence our actions. Jules Berry considers what researchers can learn from it.


Marketers were instinctively deploying behavioural economics long before it had a name. Prices ending in 99p and extended payment plans came into being because they worked, not because of a theory. Now, however, the marketing and advertising sectors are beginning to embrace behavioural economics more formally. Even the government is giving it a central place in policy initiatives. But what is it, how does it fit in with current research thinking, and how can we use it to improve our approaches?

“Marketers were instinctively deploying behavioural economics long before it had a name”

Behavioural economics (for anyone who has managed to avoid the books and press coverage that have brought this academic area to a broader audience) is a reaction to traditional economic theory, which has tended to assume that decision-making is based on individuals acting rationally. Behavioural economics shoots holes in this by demonstrating how our behaviour is, to a large extent, unconscious, irrational and socially driven. It attempts to systematically outline how these behavioural factors work so that we can first understand them, and then use them to our advantage. To paraphrase the title of Dan Ariely’s popular book, we may be dealing with behaviour that’s irrational, but that doesn’t mean it isn’t predictable.

Among the key ideas of behavioural economics are that we use rules of thumb to cut through complexity, we’re hard-wired to consider options in relative rather than absolute terms, we are heavily influenced by the people around us, and we hate losses more than we love gains.

Behavioural economics 101

?We use rules of thumb to cut through complexity
It would be impossible to make considered, rational choices for more than a handful of items when doing a weekly shop at a supermarket. To make sense of the sheer amount of information, we unconsciously adopt coping strategies, or heuristics. ‘If in doubt, pick a leading brand, because it must be OK.’ ‘The expensive coffee must be nicer than the cheaper coffee.’

We are hard-wired to think relatively
We focus on the relative advantage of one thing over another, rather than its absolute value. That means our opinion of the worth of a particular item can vary wildly depending on the context. The same bottle of wine may look a bargain at £30 in a top restaurant with a stellar wine list, but terrible value at the same price in the local gastropub where it is the most expensive on offer. How something is presented (‘framing’) and how our expectations have been set (‘priming’) are expressions of the same tendency to think relatively.

We are heavily influenced by other people
We tend to observe and copy what others do and, generally, we like conforming. When sponsoring someone, the amount you give will be heavily influenced by what other people have already put down (so make sure your first sponsor is generous to set the tone). And even when we think we are being individual, our choices are still influenced by aspirational people, experts or people we want to impress. Standing out from the crowd isn’t the same as not being influenced by it.

We think short term and we are loss-averse
The ‘buy now, pay later’ bias can affect everything from our choice of sofas to our pension planning and long-term health. Strangely, we also value things we already have more highly than things we could have, so we feel losses more keenly than gains. For example, at the birth of their first child, high-earning parents might not even have realised they qualified for child benefit – but you can expect them to be extremely unhappy at the prospect of it being taken away.

Behavioural theory in practice
The creation of a behavioural insight team within the Cabinet Office is evidence that behavioural economics has broken into the mainstream. The government plans to use this approach to ‘nudge’ us towards better decisions.

For example, only 27% of the population is on the NHS Organ Donor Register, although surveys consistently show that 65% would be willing to donate an organ. The problem is, at the moment you need to actively opt in to the scheme. This could be replaced with a system where people are prompted to opt in or out when completing, say, a driving licence application. Based on experience in other countries, that could double the number of registered donors.

Paradoxically the government is pretty overt about which buttons it is pressing with each initiative. In a discussion paper on behavioural insight and health, the Cabinet Office describes a campaign to persuade students to drink less by sending the message that levels of drinking were low, thus making people believe that heavy drinking was out of step with social norms.

“The creation of a behavioural insight team within the Cabinet Office is evidence that behavioural economics has broken into the mainstream”

There is, of course, always a danger that such attempts at manipulation do not have the desired effect. In the 1980s, regulators in the US made corporations reveal the pay of top executives, in the hope of shaming them into reducing their demands, and narrowing the pay gap between executives and average workers. The unintended consequence was that the CEOs could now see what one another were earning and fierce competition ensued, with salary league tables fuelling even higher demands. The ratio of CEOs’ to average pay has since soared to over 350. To put it in behavioural economics terms, it turned out that salary is more about peer relativity than adhering to social norms.

Strangely familiar
The research industry has been applying some of the principles of behavioural economics for a long time, if not by name. To address the gap between what people say and what they do, predictive quantitative techniques make adjustments for inertia, risk aversion, overstating of purchase intent and so on. For skilled qualitative practitioners, interpreting this gap is second nature.

The principle that everything is relative is built into research designs: we ask about perceptions of the value of a test brand in the context of priced competitors, we prime separate groups of respondents with different information to gauge the varying impact, and we use trade-off and choice techniques.

As for social influence, we can use depth interviews with friendship pairs, interactions between members in online communities and deliberative techniques. We can also include social factors such as acceptability (happy to be seen with), endorsement (have recommended) and aspiration (used by people I admire) in brand assessments.

But these are all ways of taking the mainly rational, conscious responses we get when we ask questions and trying to accommodate the influence of the irrational, unconscious elements in behaviour – isolating the bits we can measure and worrying about the complexity of the real world later. We know the process is far from perfect, but we have worked out a few fixes to make it serviceable.

In the qualitative world there are two traditions that address the discrepancy between what people are aware of and what actually drives their behaviour, at least at the data collection stage. First there’s the ethnographic tradition which prioritises observed behaviour over reported behaviour, and secondly the psychoanalytic tradition tries to dig out the underlying, unconscious drivers of how we act.

The argument against these approaches is that behavioural economics has undermined the notion that our actions can be usefully predicted at all through direct questioning. Wouldn’t it be better to cut out the rational part and observe actual behaviour, sensory experiences and brain activity?

These areas of research will undoubtedly grow in importance, particularly as the technology develops to make them more accessible. However, my bet is that question-based survey research will still be required to address the question of ‘Why?’ and to provide a link between the unconscious and conscious mind.

Now, how will these new insights into psychology and behaviour help us do better research?

The emphasis on relativity and context suggests we should get away from thinking there is a single optimal way of unearthing consumer perceptions and intentions. This will increasingly mean that we combine approaches to look at problems from different angles. Not just qual and quant, but also survey techniques with ethnographic approaches, consumer input via forums, sensory techniques and brain science.

Even when working with just one approach we can play with framing and priming to look at topics in different ways. For example, a manufacturer of contraceptives had the challenge of communicating the reliability of their product, which clinical tests had shown to have a 4% failure rate. We know that people are generally poor at evaluating risk, so the research was designed to test different ways of presenting the information. When told that the contraceptive had a one in 25 chance of failure, the majority of women thought it ‘extremely risky’. When told that it was 96% reliable, the majority thought it ‘extremely safe’.

Neither of these results is ‘the truth’, but both provide useful perspectives. If you launch a contraceptive and market it as 96% reliable, it will appeal to women looking for a safe method - but don’t be shocked if the ones who do get pregnant are upset about it.

Improved measurement and frameworks
To be really useful, the application of behavioural economic theory needs to be systematic. While a lot of the literature gives compelling examples, many are small-scale or anecdotal. How can we build models that usefully incorporate these learnings, and how do the effects play out on different brands and categories?

A first step is to view the problem and possible research solutions through the prism of behavioural economics.

Take, for example, newspaper pricing. A publisher is reviewing its price strategy and wants to understand what would happen to its brand at a number of price points. A live test market is out of the question. The classic approach would be to conduct a conjoint exercise – a highly artificial process which, while taking into account our tendency to think relatively, is essentially like a game where we highlight price changes by exposing people to multiple exercises. But we also know that people buying newspapers don’t tend to be very aware of the price, the price is not usually very visible, and people’s purchases are habitual. So the raw research results will over-predict price elasticity by a large margin.

Through a careful assessment of the behavioural factors at play, a framework can be built that helps to explain how research responses relate to behaviour in different situations (see diagram above).

An FMCG category will tend to behave more like a pure ‘market’. Research respondents might register a protest vote about certain issues but not necessarily follow it through to changes in behaviour.

Key barriers to changing behaviour are the difficulty in switching between brands or products (e.g. bank accounts), differentiation in product offerings (e.g. if your regular paper ups its cover price, will you really change to a different one?) and markets that act like monopolies (e.g. fuel, where prices tend to move together).

Of course, developing a framework is only a start. How important is each different factor? Which ones work together? How do these work at the individual level? Much work is required to produce the powerful models needed.

A catalyst for change
Behavioural economics is shining a light on the subtle and complex mechanisms that influence our actions. It is one thing to identify these mechanisms, but quite another to understand and measure them well enough to provide effective insights and accurate predictions. If we are up to the challenge of doing this, behavioural economics could be a game changer for the research industry.


13 years ago

Great article! Jeffrey Henning, from Vovici wrote a similar one on the topic http://bit.ly/e4K0Zw, which inspired my own post: http://www.relevantinsights.com/the-irrational-consumer

Like Report

13 years ago

Thanks for helping to raise awareness of this area. That, in itself, is a tough job given the Semmelweiss reflex; another cognitive bias which is the tendency to reject new evidence when it contradicts established paradigms! I would, however, take issue with your comment 'that question-based survey research will still be required to address the question of ‘Why?’ and to provide a link between the unconscious and conscious mind'. Human beings are virtually incapable of accessing the unconscious (autopilot / implicit) brain in a conscious way, yet traditional question-based surveys address the conscious (pilot / explicit) brain. This is good for measuring some things eg purchase intent but inadequate for telling us why eg fit with a neuro-psychological goal. So my bet, based on what our clients are finding to be cogent, is that research will evolve to combine the explicit and implicit - and that implicit research will replace many traditional surveys. I'm presenting on this topic at the Annual Conference and would be delighted to discuss it further with you or any interested parties. www.decodemarketing.co.uk

Like Report