Now are you satisfied?


Work commissioned by the Legatum Institute (and led by Frontier’s Chairman) has recently laid out the case for using “wellbeing”, or “life satisfaction” as the overall measure of a country’s prosperity, rather than Gross Domestic Product (GDP). The rapid development of behavioural economics is the crucial underpinning to this evolution. Frontier has been using the new techniques that these advances have made possible to inform the customer strategies of major retail businesses, and this bulletin explores the lessons that can be applied to public policy as well.

In developing their strategies for sustainable success, retail companies have become increasingly sophisticated in their collection of data on customer satisfaction. They have been using new analytical techniques to help them understand how customers behave – as opposed to how we assume they will behave – and how that behaviour is influenced by the context in which their decisions are made. And – crucially – many of the most innovative have moved on from transactions-based targets to strategies aimed at engendering loyalty and maximising the lifetime value of their customers.

Now the public sector is playing catch-up. Policy-makers are trying to use these techniques, too. Governments, as well as companies, are seeking to develop performance measures based on an understanding of the life satisfaction of their “customers”, or citizens. And they still have a lot to learn from the evolution of commercial strategies developed by retail and services businesses.

What’s wrong with output?

GDP (or the slightly wider measure, GNP) has long been the yardstick of a country’s economic progress. It is the monetary total of all the goods and services we produce. But the inadequacies of such a measure – its exclusion of intangibles, its indifference to how the income we earn is distributed – have fed the appetite for a wider measure of national performance for almost as long.

Nearly half a century ago, Robert F. Kennedy pointed out that:

“…[GNP] measures neither our wit nor our courage, neither our wisdom nor our learning, neither our compassion nor our devotion to our country; it measures everything, in short, except that which makes life worthwhile.”

Throughout the succeeding years, economists have struggled to establish a wider measure of prosperity.

That’s not easy. Happiness, a worthwhile life, even the favoured concept of “wellbeing”, are – unlike GDP – subjective. Such concepts are:

  • hard to define; and
  • harder still to measure accurately.

Behavioural economics has prolifically demonstrated how people’s behaviour, including their responses to opinion polls, is heavily influenced by context, so that it is dangerously easy to be misled as to what people think. Moreover, what they think today and what they think tomorrow may differ greatly, making it hard to rely on information about a citizen’s “wellbeing” that may be highly subject to the mood of the respondent.

But customer responses are also subjective, contextually influenced and affected by mood, and yet businesses have learnt how to overcome the difficulties in interpreting the data sufficiently well to extract useful information. Advances in behavioural economics have helped them to filter data for identifiable sources of bias. Retail companies have become increasingly sophisticated in framing questions to extract useful answers, and in testing for “framing” effects.

A number of governments have already made effects to follow suit, in the measurement of subjective indicators of satisfaction. Within a framework provided by the Organisation for Economic Co-operation and Development (OECD), more and more countries are collecting comparable data on wellbeing to inform policy-making.

The table opposite, summing up their efforts, is taken from the report by the Legatum Institute on Wellbeing and Public Policy, published in 2014. In this a group of economists, led by Frontier’s Chairman, Gus O’Donnell, have sought to overcome some of the scepticism. The authors start by defining “wellbeing”; they look carefully at how to measure it; and then they explore the factors that affect it, with the aim of identifying a number of different ways in which policies should change to enhance it.

The report considers three possible measures of wellbeing:

  • How do you feel – how happy are you (the so-called “affect” measure)?
  •  How satisfied are you (i.e., how do you evaluate your life)? and
  • Do you feel your life is worthwhile (the so-called “eudaimonia” measure)?

Its important conclusion is that, in general, the second question (relating to “satisfaction”, rather than “happiness”) produces the most robust results and best fits policy-making purposes.

In assessing their satisfaction, respondents are nudged into some kind of evaluation rather than encouraged to express feelings. And the way they make this evaluation seems to be quite like the way they behave when asked to take an actual decision (eg, to buy one thing rather than another). Unsurprisingly, therefore, this choice of measure closely echoes the focus of commercial companies on “customer satisfaction”.

As the table shows, in 2011 only France and Canada had high-quality official measures of life satisfaction. But by the end of 2014 the majority of OECD countries will have such measures, although some of them with updates to the data available only every six years.

On the other hand, some national statistical agencies, such as the Office for National Statistics in the United Kingdom, collect the data more frequently and have sample sizes in six figures. In addition, there are private sources, such as the Gallup Daily Poll in the United States, which collects information from 1,000 respondents per day.

Table 1 Availability of official statistics on subjective wellbeingCan we move the dial?

Nonetheless, despite the enthusiasm of some leading political figures (including Angela Merkel and David Cameron), much scepticism remains. Even if the two hurdles of definition and measurement can be overcome, we would – the sceptics have argued – still have difficulty in understanding what would move the wellbeing dial.

So is there any point trying? If we moved our focus from GDP to wellbeing, would anything government does actually change as a result?

Well, governments have in practice long been trying, in a number of different ways, to measure – and influence – something other than economic growth. The focus of public policy, in all democracies, is clearly broader than GDP.

Most politicians are highly-sensitised to the elusive feel-good factor amongst voters, and understand that it is not perfectly correlated with national output – hence the debate in Britain about the extent to which the current economic recovery is or is not improving living standards. (Election history suggests that voting intention is much more strongly correlated with real personal disposable income than with GDP). But the behavioural approach can take this kind of analysis so much further.

Take, for example, the work on “loss aversion”, described by some behavioural economists as their greatest breakthrough in understanding how we take economic decisions.   Put crudely, loss aversion means that the pain of loss is more acute than the pleasure of gain – that is, in general, people strongly prefer to avoid losses than to acquire gains, even when the odds and scale of gains and losses make it rational to lean the other way.

Detailed experiments have enabled economists not only to identify but to measure the scale of this bias in different situations. In the private sector, analysis of loss aversion has greatly influenced the design of, for example, retail financial products.

Target practice

In the public-policy context, loss aversion is also important. The evidence is that people (ie, voters) are more disturbed by a fall in GDP than they are reassured by a corresponding rise, so that volatility in output inflicts more damage on wellbeing, or feel-good, than slower but steadier growth would do. This has obvious implications for macroeconomic policy.

Politicians, by and large, understand that boom-and-bust (or even bust-and boom) don’t win votes. The challenge, however, is to get governments to understand that they can improve policy-making by taking a more disciplined approach to the measurement of things they may believe they understand instinctively, and by setting objectives clearly linked to wellbeing measures.

Primed numbers

Work by behavioural economists has amply demonstrated the extent to which people’s decisions, or answers to questions, can be “primed” by deliberately setting the context in certain ways. This isn’t always as obvious as making A look cheap by placing it next to (more expensive) B. Professor Daniel Kahneman (winner of the Nobel Prize for economics for his behavioural work) explains in Thinking, Fast and Slow how people asked to guess at a number (the price of a bottle of wine, the age of a celebrity) will be unconsciously influenced by goalposts offered in preceding questions (“is the number more or less than x”?). Such effects are well-known to commercial organisations, and national statisticians have to be aware of them too, even when collecting objective data. But when eliciting subjective responses they have to take even greater care. For example, it’s been shown that in the United States, the answers people gave to Gallup Poll questions as to how they evaluated their own lives were significantly affected by asking them first where they thought the country as a whole was headed. So poll designers frequently use “buffer” questions to try to purge bias. And in the field of wellbeing, the statisticians may try to make cross-checks with – for example – the views of family and friends, or examine behavioural proxies (e.g., does so-and-so smile a lot?). Even when such cross-checks can’t be carried out, changes between one survey and another may deliver useful directional information. However, the key to success is to trial surveys constructed in different ways to build up knowledge of how these differences affect the way people respond to the questions. In the same way, commercial organisations have learnt (often the hard way) of the need to conduct rigorous, well-designed randomised trials of marketing and pricing strategies rather than leave these to the instincts of “born shopkeepers”.

It’s hard enough, policy-makers argue, to design policies to support economic growth; if we try to set targets based on the softer elements of wellbeing, we are likely to sink into a fog of uncertainty and poor value for money.

Of course this is a risk. But the growing focus on wellbeing should help to prompt a review of the way in which the effectiveness of government policies is assessed.

The current system of policy appraisal is based on a monetary cost-benefit analysis (CBA). This serves a simple output (GDP-maximising) objective reasonably well, although even for this purpose it has its weaknesses.

In carrying out a CBA, it is assumed that market prices indicate the worth of goods and services, at the margin, and so that using market prices to add up costs and benefits works well. The trouble starts with goods and services that do not have market prices, or whose prices do not well reflect the broader value people ascribe to them. There are many examples in such policy areas as health, social care, law and order and income distribution, which is one reason why the state is so heavily involved in them.

These are areas in which – historically – it has not is always been possible to make a well-informed choice in a well-functioning market. In a number of countries, governments of all political colour have sought to overcome this difficulty by introducing competition and choice in public services.


Choose your question

Introducing choice into public services not only empowers citizens, it also opens them up to more of the satisfaction measurement techniques used by commercial organisations – in particular, advocacy measures such as Net Promoter Scores (NPS).

NPS is a customer loyalty metric formally launched in the Harvard Business Review in 2003, and based on the question: “How likely are you to recommend us to family/friends?”. It is normally measured on a scale of nought to ten, and calculated as the balance of those giving scores of nine or ten over those giving nought to six (with scores of seven and eight being ignored).

NPS copyright has been established by its developers, and its use has become widespread. It was enthusiastically adopted by some highly customer-focused businesses, such as Apple. It is also quite often used to measure employee satisfaction, in staff surveys (“How likely would you be to recommend working with us?”).

However, NPS has its problems, not least in measuring satisfaction with goods or services people don’t much like the notion of “recommending” (e.g., cigarettes). And frequency of use has itself become a problem, with some customers rebelling against constant demands for scoring, so that the development of new feedback mechanisms has become an important component of retail strategies.

In markets where customers make frequent transactions, the question can be asked more simply (“Would you use us again?”). But NPS tries to measure something broader than satisfaction with a particular product – the degree of loyalty to a particular brand or institution.

The public sector has a lot to learn in the search for feedback. It is, of course, less easy to calibrate responses where the “customer” has no choice (you cannot meaningfully ask citizens if they would recommend the services of the police force to their friends). And even where competition can be introduced, the fact that most services are free at the point of use distorts the customer view of value for money.

A particular problem arises, in both the public and private sectors, in areas where it is easy to make bad choices, out of lack of information or in response to contextual influences. Obvious examples are financial services or medical treatments; but there are other areas of difficulty, where competition has increased choice but also made it more complicated (telecoms, energy).

For the past 20 years, public policy has been focused on increasing or improving competition by informing customers more. Only recently have policy-makers come fully to appreciate that more and more information may not always be what customers really want, or what will help them to make better choices. So researchers are now trying to build policy models that reflect the way people actually behave, rather than the (supposedly rational) way in which the traditional models assume that they do.


Quality before quantity

In some areas of public policy, of course, the pursuit of wellbeing (or something like it) is already embedded in appraisal methods. The benefits of new medical treatment are measured in terms of quality-adjusted life years (QUALYs), incorporating different elements of quality of life such as mobility, freedom from pain, self-sufficiency and the ability lead a social life.

However, research suggests that gathering subjective data on these elements would improve their appraisal: not least, it would cause factors to be weighted differently in the calculations, with mental suffering carrying a greater weight than at present, in comparison with limitations to physical mobility.

Of course the addition of subjective factors does not do away with all the problems of policy assessment. Distribution remains an issue to be added to the appraisal. Traditional CBA does not take account of whether an increase in income for the poorest should be weighted equally with an increase enjoyed by the rich. But equally, an analysis of wellbeing still leaves you to decide whether increases should be weighted according to how much wellbeing was initially enjoyed by the different groups of people affected.

The Legatum report grouped the drivers of wellbeing in three separate categories: economic, social and personal.

  • Economic influences. Research highlights the effect of education and work, as well as income. The psychological effects of unemployment are a major source of unhappiness; and measures of deprivation that only relate to income miss a lot. Back-to-work policies, therefore, have a wellbeing value over and above the output and income increases that will be picked up by a traditional CBA.
  • Social influences. Family life, community life, values, and the environment all show up strongly as determinants of life satisfaction. On this score, the report concludes that governments need to concentrate on building trust between people, on preventing corruption and protecting the freedom of the individual.
  • Personal influences. Research makes it very clear that health is the main personal determinant, to the extent that differences in health are more significant influences on life satisfaction than differences in income. And within the range of health issues, mental health stands out as the crucial factor.

Having sought to identify the strongest determinants of wellbeing, the authors discussed a range of possible public policies to enhance it. The four areas the report highlighted were:

  •  mental health and character building;
  • community;
  • income and work; and
  • governance.

One of its most important conclusions was that we should treat mental ill health as professionally as physical ill-health. Another was that we should focus not only on education per se, but on wider issues of upbringing – supporting parents, and building character and resilience in schools. At the community level, the authors concluded that we should promote volunteering and giving, address loneliness and create a built environment that is sociable and green.

As well as promoting economic growth, the report argued that we should aim to reduce unemployment through active welfare policies and encourage businesses to promote wellbeing at work. And the authors strongly believed that to enhance wellbeing, we should focus policy on treating citizens with respect and empowering them more.

Although some policy-makers remain sceptical, arguing that this is no more than what they try to do already, the value of this kind of work is to embed such objectives properly, and test policies rigorously for their effect. Moreover, assessing policies for their contribution to human dignity and life satisfaction is of obvious relevance to services where the quality of care – for example, of older people – is critical to performance.

The report came to some important conclusions as to how governments should alter their priorities. But much more work is needed to support these conclusions and turn them into policies. In particular, more experimentation in policy design is needed to discover what works and what doesn’t.


Hard Nudge, Soft Nudge?

A notable characteristic of some wellbeing objectives is that their achievement is likely to demand the use of policy tools other than legislation. “Soft law” – incentives or codes of good practice – may be the most appropriate.

“Hard law” is suitable – even essential – for areas where absolute compliance with the rules of civil society is required and must be enforced (for example, the criminal code). But to promote wellbeing, government is often seeking to encourage (e.g., volunteering) or discourage (e.g., obesity) rather than to insist (e.g., on everybody paying tax) or prohibit (e.g., fraud). So this is territory for the use of a number of tools developed by behavioural economists, and in particular the “nudge” techniques made famous by Professor Thaler.

Again, these are techniques already well-developed in the private sector.   Companies can rarely insist or prohibit – they have to persuade or discourage, using the vast advertising and marketing industries to help them.

Supermarkets aren’t laid out by accident, pricing is a science informed by the latest breakthroughs in behavioural economics, salesmen have been perfecting their patter ever since they set up the first market stall. Regulators, too, have had to build their expertise in this area, to prevent product mis-selling, subliminal or misleading advertising – as well as the essential underpinning of market regulation to prevent anti-competitive behaviour.

The need for this policing is greatest in complex markets – such as financial services – where manipulation or misleading claims are least easy for the customer to spot. But these are also markets where government may itself want to manipulate or “nudge” citizens into prudent behaviour.

Arguably, the justification for nudging is much easier when sharp elbows are being used to push people towards their own best interests rather than towards your own bottom line. But the differences between public and private approaches, and their justifications, may be greater in principle than in practice.

For a start, not all corporate nudging is unhelpful or unwelcome – a grocery website that reminds you what you bought last time usefully prevents you forgetting things as well as encouraging you to buy more. And far-sighted businesses know that it’s a mistake to nudge too far: there’s a degree of manipulation that will destroy trust and so, ultimately, sales.

Businesses that use prompts to suggest cheaper options (e.g., for railway tickets, or grocery products) are consciously trying to guard against this risk of a loss of trust. Equally, governments should know that they too can destroy trust by manipulating citizens, even if supposedly in their own interests.

The first thing government can learn from the private sector is in understanding that every communication it sends to citizens is an opportunity – and so a choice. In general, the civil service has attempted to make its stream of communications “objective”. But since all communications have some impact, government needs to think harder about what it’s doing.


Should we, can we?

The scope for change is huge precisely because government rarely designs its communications with any consideration (still less proper testing) of their behavioural effect. But once the public sector gets into this game, it has to think hard about both the principle (is it ethical?) and the practice (does it work?).

Governments, like companies, have two main tools they can use for “nudging” – information – trying to be helpful and/or influential – and money – offering financial incentives or disincentives. However, unlike companies governments can also use their powers to nudge behaviour through a mixture of compulsion and persuasion.

A particularly popular regulatory tool is transparency. Regulators may seek to change behaviour not by imposing obligations to act in a particular, prescribed way, but simply by requiring public disclosure.

For example, Britain’s set of “rules” for the governing of listed companies – the Corporate Governance Code – isn’t hard law but issued on a “comply or explain” basis. Companies don’t have to do what it says; but if they don’t, they have to explain to their shareholders why they don’t. Its advocates would argue this has enabled Britain to move faster and further in setting good governance rules than countries which do this stuff by statute, while at the same time leaving space for justifiable exemptions to best practice.

A related example is that all large quoted companies are required to say how much they pay their top executives in their annual reports. Sunlight, it’s argued, is the best disinfectant; and it’s hard to argue that you’ve been forbidden to do something if all that is required is to come clean.

Other examples of how governments can use a mix of hard and soft law to influence behaviour include the choice given speeding drivers to avoid licence points by attending a course educating them on the dangers. But the more nudging governments do, the more they have to be aware of the associated risks.

  • The ethical risk is that they will interfere ever more in people’s lives, and do so in ways that – unlike laws – don’t have to be approved by elected representatives.
  • The practical risk is that by interfering, governments lose information as to people’s preferred outcomes, and damage their ability to learn by their mistakes. And meanwhile, nudging may prove ineffective – the tax system is littered with failed incentives.

Pensioned off

When the British Chancellor announced, in the 2014 Budget, that people would no longer be required to buy annuities with their pension pots, he overnight ceased to be attacked for paternalism and became accused of irresponsibility. Before, pensioners had been complaining about being forced into poor-value products (particularly when annuity rates were low). After, the Chancellor's critics complained that he was allowing pensioners to blow their savings on cruise holidays and sports cars, leaving the state to pick up the bill when they ran out of money in old age. Industry responses meanwhile ranged between those who feared the loss of a relatively stable income stream and those who saw it as a great opportunity to sell new (and perhaps more complex) products. So what’s the answer? We start from the general assumptions that (a) giving people control over their own choices leads to better outcomes, and (b) allowing them to learn by their mistakes is a better way of getting to the “right” place than dictating it from the beginning. Forcing pensioners into annuities offended against both these assumptions: its inflexibility obliged many people to buy unsuitable products, and did little to encourage competition or innovation among providers. Moreover, the prospect of having most of your savings locked up in an annuity acted as disincentive to save. However, the learning-by-doing approach runs into difficulty when applied to one-off, financially complex decisions that don't allow for mistakes – and choosing what to do with your pension pot is plum in that category. So in the new world of free choice, pensioners needed to be nudged in a prudent direction. The essential second stage in this policy, therefore, came in July 2014, when the Chancellor announced that pensioners considering their new options would be given a certain amount of free, independent advice, funded by a levy on the relevant part of the financial services industry.

A reasonable defence against the ethical challenge is if you can show you are using nudging instead of legislation, since that reduces the degree of interference in people’s lives. “Soft law” avoids the need for several hundred pages of complex legislation, employing armies of lawyers to define, enforce and defend against.

But this moves the focus on to the practical challenge, since nudgers have to be able to prove that their approach is effective. And the best defence against this challenge is to engage in effective testing of policies.

The British Government’s Behavioural Insight Team (or “nudge unit”) won plaudits for a trial that it ran at a job centre in Essex, changing the way jobseekers were treated. The amount of paperwork that normally swamped the first meeting was reduced, leaving time to talk about getting back to work from day one, and helping claimants to set targets for action for the next fortnight. Advisers were given new tools to help build self-confidence and well-being amongst claimants still out of work after eight weeks. Comparison with a control group suggested a significantly greater chance that jobseekers treated in this way would be back in work within thirteen weeks.

The nudge unit has also taken some credit for the higher taxpaying response to letters from Her Majesty’s Revenue and Customs letting laggards know that others in their area had already paid up. However, the unit itself has itself recently been challenged to report on experiments that haven’t worked as well as on those that have.

An important example of the policy choice between rules and nudges came in the UK with the 2014 Budget (see box on the right hand side).

Both businesses and governments have to learn how to “nudge” not only effectively but efficiently. The response rate to a “nudge” technique acts as a useful measure of both its appropriateness and cost-effectiveness.

If the response is low, you’ve clearly failed to understand the underlying behavioural influences. But if you’re getting 100 per cent, perhaps you’ve been too heavy-handed – which may mean, as either a company or a government, throwing too much money at your target.

Again, the essential discipline is to trial your technique effectively before rolling it out completely. Incentives are hard to roll back without damaging customer trust. Testing the proposition may, of course, be easier for a supermarket than – say – HMRC. The last great experiment in trialling a tax (the introduction of the Community Charge, or poll tax, in Scotland before rolling it out across the rest of the UK) had political consequences still being felt today.



Behavioural economics is helping both businesses and governments to move on from output targets to measures that help them increase the satisfaction of customers and citizens. For both, that helps build sustainable strategies. Much that has already been developed in forward-looking retail and service companies is applicable to public policy, while customer strategies themselves are evolving fast.

Frontier has worked for a number of major retail businesses on customer lifetime value, choice and behaviours; and also on the introduction of competition and nudge incentives into public policies. This bulletin has attempted to bring together some transferable learning from both the public and private sectors, and some common principles for the development and testing of effective techniques.

Comments are closed.