This Thing For Which We Have No Name

A Conversation with
Rory Sutherland [5.12.14]
Topic:

"No one ever got fired for buying IBM" is a wonderful example of understanding loss aversion or "defensive decision making". The advertising and marketing industry kind of acted as if it knew this stuff—but where we were disgracefully bad is that no one really attempted to sit down and codify it. When I discovered Nudge by Richard Thaler and Cass Sunstein, and the whole other corpus on Behavioral Economics…. when I started discovering there was a whole field of literature about "this thing for which we have no name" …. these powerful forces which no one properly understood—that was incredibly exciting. And the effect of these changes can be an order of magnitude. This is the important thing. Really small interventions can have huge effects. ...

...Markets actually work because they're adaptive. Bad things get killed, good new things sometimes get promoted. But most of the time what you'll find in business is no one has the faintest idea of why the things that work actually work. What's very useful here is that finally a group of academics with money, time, and immensely high intelligence were finally sitting down to codify and make sense of things, which we'd been aware of for years but which, to our shame, we'd never attempted to actually try and systematize.


(49:17 minutes)

RORY SUTHERLAND is Executive Creative Director and Vice-Chairman, OgilvyOne London; Vice-Chairman, Ogilvy & Mather UK; Columnist, The Spectator.

Rory Sutherland's Edge Bio page.


THIS THING FOR WHICH WE HAVE NO NAME

The strange thing about academics, which always fascinates me, is that they believe they're completely immune to status considerations and consider themselves to be more or less monks. In reality, of course, academics are the most status-conscious people in the world. Take away a parking space from an academic and see how long he stays. I always find this very strange when you occasionally get in the realm of happiness research, you get fairly considerable assaults on consumerism as if it's just mindless status seeking. Now, the point of the matter is, is that academics are just as guilty of the original crime, they just pursue status in a different way. ...

~  ~  ~

I have probably stolen, without realizing it, your own job title of "impresario" rather unfairly. The reason I did this was that occasionally people started writing about me online as a "behavioral economist" and I realized that, among academic behavioral economists, this would drive them practically apoplectic to have someone with no qualifications in the field so described. (I'm a classicist by background in any case). So my being described as a behavioral economist would make them practically deranged.

It's not my job to be a behavioral economist; my job is to actually popularize behavioral economics. The interesting thing to me—as someone who has a brother who's an academic—is that academics, however brilliant, are the worst marketers in the world. And there's a reason for this, which is they, of course, possibly rightly, despise the idea of marketing. Their idea is that there is a pure and objective truth and people should appreciate it. As they see it, the way in which you present that truth should be irrelevant to its acceptance. So indifferent are academics to presentation, it seems, that the PowerPoint slide which announced the existence of the Higgs Boson was partly written in the font Comic Sans.

My contention would be that there are ideas which, depending on how you present them—it's rather like the experiment with Wason Cards—may either be easily accepted and understood or are incomprehensible and/or repellant. The reaction to those ideas will entirely depend on how they're framed and presented, and have nothing to do with the objective truth underlying the findings. My personal view, as much as it may seem repellant to them, is that academics actually need impresarios, need marketers, and need popularizers.


The reason I'd happily describe myself as a behavioral science impresario is that many recent insights from six or seven interrelated fields of social science are extraordinarily important in terms of business activity, but more important still, public policy. If those things aren't widely known, appreciated, and understood, and if people aren't allowed to grasp them in the right way, then they will be crudely overlooked. I said elsewhere on Edge that in the social sciences, the good ideas aren't always influential and the influential ideas aren't always good. To have some very, very important ideas from Kahneman, for instance, or Jonathan Haidt from evolutionary psychology…. to have those overlooked or effectively rejected by practical people in positions of power or influence would be really tragic.


I was always taken by the Richard Feynman thing where he said, and I'm paraphrasing, "When you come up with some true thing, you can often find the same truth that's expressible in three or four different ways." And what those four different expressions say is identical in objective reality; however, the kinds of ideas and implications and applications that arise, depending on how you express the truth, are psychologically completely different. Expressing the same thing in four different ways is, in one sense, completely worthless as an activity. On the other hand, in terms of generating further ideas based on those insights, the kinds of things that will be generated will be dependent on how those things are presented. "Therefore psychologically we must keep all the theories in our heads, and every theoretical physicist who is any good knows six or seven different theoretical representations for exactly the same physics." That's a really important thing.

The reason I'd happily describe myself as a behavioral science impresario is that many recent insights from six or seven interrelated fields of social science are extraordinarily important in terms of business activity, but more important still, public policy. If those things aren't widely known, appreciated, and understood, and if people aren't allowed to grasp them in the right way, then they will be crudely overlooked. I said elsewhere on Edge that in the social sciences, the good ideas aren't always influential and the influential ideas aren't always good. To have some very, very important ideas from Kahneman, for instance, or Jonathan Haidt from evolutionary psychology…. to have those overlooked or effectively rejected by practical people in positions of power or influence would be really tragic.

The problem—in terms of understanding human behavior, and I go even further, in understanding economics—is, why is neoclassical economics so disproportionately influential among the social sciences in influencing policy and bank behavior and, indeed, business? I mean, the finance director—the chief financial officer—is now typically the second most powerful person in any organization. He will, unconsciously or not, base quite a lot of his decisions on the unproven assumptions of neoclassical economic theory. So why has neoclassical economics, which is a pretty appalling predictive guide to individual human behavior, achieved this influence? One of the reasons is it's mathematically neat so people just fall in love with the elegance of the thing and are happy to ignore the fact that, empirically, it ain't all that.

Another one, I suspect, is a peculiarity of academics. My brother is an astrophysicist. I once said to him, "Look, you do extraordinarily advanced math about galaxy clustering and such. Why don't you just bugger off to a bank for three years, make a stack of money, and then return with the freedom to do what you like?" And he said, "The problem is, for example, in astrophysics, disappearing off to work in a bank for three years is effectively career suicide." Now, there's one group of academics where working for Citibank for two years on a highly-paid gig is possibly not career suicide, and that would be among economists.

By contrast, I imagine that if you're an anthropologist and you spend three years working for Unilever, it's actually a career setback. My guess would be that if you were a sociologist in 1975 and you went and worked for an advertising agency, you might as well retire afterwards because no one else in academic sociology was going to give you a job. But economics has this disproportionate influence and my mission is to rebalance things a little bit.

The other problem in overcoming the disproportionate influence of economics is that to understand why conventional economic approaches are wrong—and to understand what is needed to replace them—you possibly don't have to understand one thing, you have to understand about five or six different things. You probably need a bit of game theory, a bit of evolutionary psychology, a bit of behavioral economics, a bit of complexity theory. Now, the problem then is if you have a case where in order to reject the consensus you need people who know a bit about six different things, then simply by statistical averages, the number of people who appreciate all of those five or six different things is going to be a hell of a lot smaller. And that's genuinely the case.

There are some good social scientists—Jon Elster in Understanding Social Behaviour, for one—who make this point: that unless you really understand game theory, you can't begin to actually understand human behavior. That's pretty fair. There's also an interesting question, which I'd like to debate, which is the disproportionate obsession with mathematics in the social sciences. It's not to say that mathematics is not valuable, it's simply to say that if you create disciplines where you aren't considered a science unless you use math, unless you can actually mathematicize what you're attempting to describe. That strikes me as extraordinary limiting.

First of all, the relative progress in mathematics is not without its own biases. If you look at it, the Greeks were extraordinarily good at geometry. Why? I suppose because they had two obsessive preoccupations: Astronomy in order to predict weather or seasons so you can plant crops—and perhaps in calculating boundaries between parcels of land. If you look at the history of math, certain things like an understanding of basic statistics and probability was extraordinarily late to emerge in mathematics. Someone told me that in the 17th Century people would assume that if you threw a die six times, each of the six numbers should come up once each. They were that shit and that was how bad it was. Game theory only really kicked off after the Second World War. There are many things that were impossible to model until the advent of modern computing power. So a theory expressed in words may be the best we can do for now.


Math may be an obstacle to good thinking because it's actually constraining. We have a perfectly good thing called words and grammar, which can describe very complex concepts. Regret, for example—maybe you can't mathematically express it very easily but it patently affects human behavior. The fear of regret is clearly fairly influential in affecting how people behave. I would argue that from my advertising perspective, one of the reasons people pay a premium for brands isn't because they're objectively better, it's because they're less likely to be terrible.


The BBC has recently run a series of podcasts, one of them being about Hyman Minsky—a kind of dissident economist—who did a pretty good job of explaining economic bubbles and crashes. They asked the question on the podcast: why was it that Minsky was so lacking in influence as an economist? And they said, "He didn't really do much math." The Austrian-school of economics also has a natural aversion to math because they claim you can't mathematically model human behavior—as a result, apparently, Austrians can't get jobs in economics.

Math may be an obstacle to good thinking because it's actually constraining. We have a perfectly good thing called words and grammar, which can describe very complex concepts. Regret, for example—maybe you can't mathematically express it very easily but it patently affects human behavior. The fear of regret is clearly fairly influential in affecting how people behave. I would argue that from my advertising perspective, one of the reasons people pay a premium for brands isn't because they're objectively better, it's because they're less likely to be terrible.

Whatever you think about McDonald's—it's really, really good at not being bad. If you understand satisficing—which would be another concept hugely important to the understanding of human behavior—we think we maximize and we describe our behavior as if we're maximizing but most of the time we go "I want something that's pretty good and definitely isn't awful." Why do we go to McDonald's? Is it the best food in town? Probably not. The search cost of finding the best place to eat in town, given that we've only got one shot at having a meal in a strange town, would be pretty high. But also when you go into McDonald's you know you're not going to be ripped off, you're almost certainly not going to be ill. By contrast I've become ill after eating at Michelin-Starred restaurants quite frequently. Once you understand the perfectly sensible evolutionary instinct to satisfice, then the preference for brands is not irrational at all: I will pay a premium as a form of insurance for the reduced likelihood that this product is appalling. Is that called a minimax approach? Someone help me out here. ...

Certainly there's a problem with numbers in that there are sophisticated things in life that we all understand perfectly well when verbally described. Should psychology be constrained by math? I mean, who has the better understanding of human behavior—Shakespeare or Eugene Fama? If you make mathematical expression a barrier to entry, to any kind of theory, you are undoubtedly limiting yourselves.

~  ~  ~

It always struck me that when I first discovered conventional economics—which was a weird period where I was ill for about five days and ended up reading endless books on economics—I had a strange sort of mixed feeling which is that, first of all, this is incredibly beguiling and seductive and elegant; the very elegance of the theory can't help but be captivating. To an extent I was a convert at first. It also has a rather good vocabulary, which I still use to the annoyance of people around me: "We could do that again but it has diminishing marginal utility." It's quite a useful thing. "Sunk cost bias" is a very useful concept. Understanding it is very useful because you can correct it in yourself.


It is true of quite a lot of progress in human life that businesses, in their blundering way, sometimes discover things before academics do. This is true of the steam engine. People developed steam engines before anybody knew how they worked. It's true of the jet engine, true of aspirin, and so forth. People discover through trial and error—what Nassim Taleb calls "stochastic tinkering." People make progress off their own bat without really understanding how it works. At that point, academics come along, explain how what works works and to some extent take the credit for it. "Teaching birds to fly" is, of course, the phrase that he uses.


One of the single moments where I realized this stuff is really useful is when I first tore up a pair of air tickets. My wife and I had some nonrefundable air tickets to go to Paris for the weekend and the day before we were due to travel, both of us went down with flu and were feeling appalling. We were there packing, thinking, we've bought these tickets, they're nonrefundable, we have to go to Paris. And I suddenly said, "Hold on, we're now going to spend another 300 or 400 pounds on hotels, art galleries, and everything else in order to feel crap in a hotel room rather than feeling mildly ill at home." The moment when I tore up those travel tickets that was almost a little Damascus Road experience where I realized some of this thinking is practically useful in everyday life.

It is true of quite a lot of progress in human life that businesses, in their blundering way, sometimes discover things before academics do. This is true of the steam engine. People developed steam engines before anybody knew how they worked. It's true of the jet engine, true of aspirin, and so forth. People discover through trial and error—what Nassim Taleb calls "stochastic tinkering." People make progress on their own without really understanding how it works. At that point, academics come along, explain how what works works and to some extent take the credit for it. "Teaching birds to fly" is the phrase that Taleb uses.As I say, I was seduced by economic thinking and the elegance of it, but at the same time having worked in advertising for 15 years, I was also fairly conscious of the fact that this isn't really how people behave. We'd always known, in those fields of marketing, like direct marketing, where you actually got results—you sent out letters to 50,000 people and saw how many people replied—there was something going on that we didn't understand. In other words, occasionally you might do incredibly elaborate, complex, and expensive work and have more or less no effect on the uptake of some product. Then someone would redesign the application form and slightly change the order of the questions on the application form, and the number of people replying would double. We knew there was this mysterious kind of dark force at work in human behavior.


"No one ever got fired for buying IBM" is a wonderful example of understanding loss aversion or "defensive decision making". The advertising and marketing industry kind of acted as if it knew this stuff—but where we were disgracefully bad is that no one really attempted to sit down and codify it. When I discovered Nudge by Richard Thaler and Cass Sunstein, and the whole other corpus on Behavioral Economics…. when I started discovering there was a whole field of literature about "this thing for which we have no name" …. these powerful forces which no one properly understood—that was incredibly exciting. And the effect of these changes can be an order of magnitude. This is the important thing. Really small interventions can have huge effects.


The extraordinary thing about the marketing industry is that, by accident, it was pretty good at stumbling on some of these biases which behavioral economics later codified. There's a wonderful/evil advertisement I mentioned in a recent piece, "How else could a month's salary last a lifetime?", which is a De Beers advertisement in about 1953 for engagement rings. Now, that's a brilliant case of framing or price anchoring. How much should you spend on an engagement ring? We'll suggest that whatever your month's salary is, that's what you should spend.

"No one ever got fired for buying IBM" is a wonderful example of understanding loss aversion or "defensive decision making". The advertising and marketing industry kind of acted as if it knew this stuff—but where we were disgracefully bad is that no one really attempted to sit down and codify it. When I discovered Nudge by Richard Thaler and Cass Sunstein, and the whole other corpus on Behavioral Economics…. when I started discovering there was a whole field of literature about "this thing for which we have no name" …. these powerful forces which no one properly understood—that was incredibly exciting. And the effect of these changes can be an order of magnitude. This is the important thing. Really small interventions can have huge effects.

Everybody assumes outside the advertising industry—outside business, outside marketing—that marketing is full of incredibly cunning, evil people. That you'll walk through an advertising agency and stumble through a door and find people experimenting with electrodes attached to the brains of rats. The embarrassing truth is those rooms don't exist. Most of the progress that's made in business is made through a kind of trial and error where you accidentally stumble on something that's successful. Of course, the way business works quite well is that things that are unsuccessful get killed off fairly quickly and things which are accidentally successful get invested in; a very crude feedback system but it kind of works, broadly speaking.


Markets work because they're adaptive. Bad things get killed, good new things sometimes get promoted. But most of the time what you'll find in business is no one has the faintest idea of why the things that work actually work. What's very useful here is that finally a group of academics with money, time, and immensely high intelligence were finally sitting down to codify and make sense of things, which we'd been aware of for years but which, to our shame, we'd never attempted to try and systematize.


The idea that capitalism doesn't work because it's efficient—that's a complete myth. Anybody who's worked in any kind of capitalist enterprise for any length of time will know that it's not even remotely efficient. There are spectacular inefficiencies all over the place.

Markets work because they're adaptive. Bad things get killed, good new things sometimes get promoted. But most of the time what you'll find in business is no one has the faintest idea of why the things that work actually work. What's very useful here is that finally a group of academics with money, time, and immensely high intelligence were finally sitting down to codify and make sense of things, which we'd been aware of for years but which, to our shame, we'd never attempted to try and systematize.

There are going to be people watching this video or reading the text who are going to have ethical dilemmas. The truth of the matter is there are of course ethical dilemmas because you can patently lead people to do things they subsequently regret. On the other hand, at its very simplest, nudging can be nothing more than the act of painting white lines in the middle of the road, or painting a pattern in a car park so that people all park in a way which allows more cars to fit into a given space. You know, as an absolutely purist libertarian, I could get really angry and say, "I hate car parking lines because they interfere with my right to park at the diagonal." But you'd have to be a fairly deranged libertarian purist to take that view, to protest against the lines in car parks. I suppose you could argue they're not obligatory in any case. You can, if you want to, park at the diagonal it's just that at that point, social and reputational mechanisms come into play and people just think you're an creep if you park across two spaces. So the lines in car parks are a reputational nudge in a way. That's cute.

I am immensely grateful for academics in the social sciences for finally taking these things seriously. They are a bit prone to highly purist feuds, i.e. The People's Front of Judea vs. the Judean Popular Front, or whatever. My point of view from a business perspective is, we find these feuds bizarre. And the reason is, is that, as I said, in business you don't have to be right, you've just got to be less stupid than the other guy. We'll happily use bad ideas if they're less bad than our competitors' bad ideas.

I know that academics go practically deranged at the mention of Malcolm Gladwell. They'll say some of the things that he says are unrepresentative, or slightly inaccurate, or he misrepresents science, or takes credit for this or that. My view is much more benign, which is, look, if 100,000 business people read his books and wake up the next day being slightly more open-minded or slightly happier to work with the counterintuitive than they were the day before, then that book's a net victory and I'm not that bothered. But that's because business and academia are fundamentally different. In academia you have to be right, in business you've just got to be less stupid, less wrong than your competitors. And so the approach we have to progress is fundamentally different. We just look for ways to be less wrong. And of course we don't need to know why something works, just that it does.

~  ~  ~

If you would accept me as the other impresario, I'd be immensely grateful. If this meant that really good, worthwhile, and useful ideas, which in many cases could be enormously beneficial, can start to percolate through business decision-making and replace the dumber ideas. What I talk about, by the way, when I talk about dumb ideas is two particularly dumb things that affect both business and government decision-making. One of them is the assumptions of neoclassical economics, which basically creates an imaginary species in order to be able to mathematically model it.


In the commercial world it's not in the interests of consumers to be rational. If you imagine a world of rational consumers where they all bought cars based on some formula, which was fuel economy multiplied by acceleration divided by depreciation or whatever it may be, what they would end up with would be really, really terrible cars. Car manufacturers would immediately game their predictable and easily understanding preferences and produce cars, which met all the metrics laid down for a desirable car, but the cars would be ugly, uncomfortable, and generally ghastly, and no pleasure to own. It is our interest, in a sense, as consumers to have a degree of unpredictability to our behavior and to be difficult to model. But homo economicus is really a contrivance in order to pretend that humans can be treated as though they were atoms or planets or something like that, which is a fundamental category error. It's what Hayek would call ".....a mechanical and uncritical application of habits of thought to fields different from those in which they have been formed."


Homo economicus, not only has it never existed, my contention is that if homo economicus evolved for some peculiar reason, it would die out very, very quickly or we would club it to death with sticks. It's not in our evolutionary interest to be rational optimizers, by the way. Any species that's rational also becomes highly predictable, and anybody who's highly predictable is rapidly dead because you can anticipate his behavior and game it. Sometimes being unpredictable is the best approach.

In the commercial world it's not in the interests of consumers to be rational. If you imagine a world of rational consumers where they all bought cars based on some formula, which was fuel economy multiplied by acceleration divided by depreciation or whatever it may be, what they would end up with would be really, really terrible cars. Car manufacturers would immediately game their predictable and easily understanding preferences and produce cars, which met all the metrics laid down for a desirable car, but the cars would be ugly, uncomfortable, and generally ghastly, and no pleasure to own. It is our interest, in a sense, as consumers to have a degree of unpredictability to our behavior and to be difficult to model. But homo economicus is really a contrivance in order to pretend that humans can be treated as though they were atoms or planets or something like that, which is a fundamental category error. It's what Hayek would call ".....a mechanical and uncritical application of habits of thought to fields different from those in which they have been formed."

Understanding human behavior is always, in my view, going to be slightly messy, imperfect, and so forth. Different people, by the way, will always use different heuristics in order to make decisions. No group of people will ever behave in an identical or predictable way, nor will they always maintain those behaviors or proxies for a long time.

Digital cameras always interest me. There was a brief period where, broadly speaking, people bought a digital camera and they asked the question, "How many megapixels has it got?" That was kind of a lazy proxy for how good a camera is it? Don't get me wrong, this wasn't a totally bonkers rule at first. A 3 or 4 megapixel camera is significantly better than a 1.2 megapixel camera, and an 8 megapixel camera is better than that. When you get to about 12, frankly, unless you're working for the North Korean government where your only requirement is to take photographs that are going to be projected onto buildings of enormous scale, 12 megapixels is good enough. I've asked professional photographers, "Do you use Raw?" They don't. It takes up too much space, basically. It's just too damned unwieldy and big. And the actual gains are, frankly, trivial.

Consumers, when they buy digital cameras, don't obsess about the number of megapixels any more because at some point that metric becomes very bad. You create overheating problems in the CMOS, and goodness knows what else. And they'll move on to design or robustness.

The most successful camera brand, arguably, is GoPro, which is not about picture quality at all. It's about being able to take pictures while you're doing extreme activities. The great thing about consumer markets is the very fact that they're volatile. The consumer markets, by being volatile and unpredictable and strange, ensure a fairly good supply—a diverse supply—of interesting new products. If people's behavior were kind of rational and uniform, the products we'd get would be awful—Soviet-style. Something quite interesting is going on here. This is always going to be messy. But the fact is there are very, very simple things where both businesses and governments are making appalling misjudgments. They look for a right answer. When in fact what you need is lots of non-wrong answers.

~  ~  ~

When people say, "Nudging is unethical because it's effectively exploiting innate, unconscious biases to get people to do things that consciously they wouldn't want to do." The simplest example I always give is  that we live on the second floor, the bag of trash gets too heavy for my wife to carry downstairs. Now, if she wants to carry the rubbish downstairs because the bag is now too heavy to manage, there are basically two approaches she can adopt. There's the first approach which is wait until I'm sitting half-dressed with no shoes on, watching a television program, and say, "Rory, can you take the rubbish downstairs?" This is mistimed, slightly annoying, irritating, and at a moment when I will now have to get up from the chair, put shoes on, and perform this slightly tedious activity. The alternative is that, without saying a word, she simply leave the rubbish by the back door. The next time I'm headed downstairs I'll carry it down without even thinking about it at all. ...

~  ~  ~


I'm very interested in the work of Gerd Gigerenzer because, of course, one of the points he makes, which is naturally music to the ears of a marketer, is that how you present true things has a huge effect on how we perceive and understand and act on those true things. The same thing could be true in four or five different forms, depending on how you display statistics or how you display probabilities but, nonetheless, the presentation of it in one form is in line with our evolved psychology and we're comfortable with it. The presentation in a different form is dissonant with our evolved psychology.


I'm very interested in the work of Gerd Gigerenzer because, of course, one of the points he makes, which is naturally music to the ears of a marketer, is that how you present true things has a huge effect on how we perceive and understand and act on those true things. The same thing could be true in four or five different forms, depending on how you display statistics or how you display probabilities but, nonetheless, the presentation of it in one form is in line with our evolved psychology and we're comfortable with it. The presentation in a different form is dissonant with our evolved psychology. His new book has a super section on the Monty Hall problem. I've often wondered that you could try reframing that problem, with Monty replaced with a Buddhist Monk, or someone assumed to be benign and helpful, to see if more people got the right answer. ...

If you look at the world of physical design—I drove you here today and I steered the car with my hands; every single car I know, including Formula 1 cars has a steering wheel. Now, our hands didn't evolve to steer cars. What we do very sensibly is we design cars in such a way as some evolved equipment that we have is quite good at steering, which is why nobody's attempting to devise an interface where you steer your car with your nose.

In the physical world we're very good at designing things that work well with our evolved physique. If you take the psychological world, we're much, much worse. We are, in many cases, doing the equivalent of designing a car you have to try and steer with your nose. Because our understanding and our appreciation of evolved psychology is much worse. First of all, we think there is such a thing called perfect reason. That we should be able to deploy reason and logic in every decision we make, when most of the time we simply don't have enough information to do so or the cost of acquiring that information will be absurd, or the information is not available in a comparable form. You know, it's very difficult—weighting. If you're looking at three or four variables in parallel when buying a car, what relative weighting do you attach to things? Are they linear? Are they even sort of monotonic in terms of how we perceive them? Lexicographical choice is much easier.


What fascinates me is that when we design physical things, no one's so dumb as to design a car that you steer with your little toe. When we design programs and government policies and things, we commit that error all the time. And the reason we commit the error is because, first of all, nearly all the decisions where we attempt to predict or understand human behavior are based on, as I said, this broken pair of binoculars. One lens is neoclassical economic theory and the idea of perfect rationality, perfect information, perfect trust. That's obviously wrong. The conditions of a rational actor possessed of perfect information and perfect trust engaging a series of independent, standalone, utility-maximizing behaviors is clearly nonsense.


I mentioned a few fields I thought were important for anybody understanding human social behavior. I mentioned, obviously, evolutionary psychology, I mentioned game theory, I mentioned behavioral economics, and so forth. I'd also include psychophysics in that. I mean, we said we don't perceive the world in objective ways. It's not in our evolutionary interests to perceive objective reality. It would be computationally very wasteful and in survival terms, it would be extraordinarily ineffective.

What fascinates me is that when we design physical things, no one's so dumb as to design a car that you steer with your little toe. When we design programs and government policies and things, we commit that error all the time. And the reason we commit the error is because, first of all, nearly all the decisions where we attempt to predict or understand human behavior are based on, as I said, this broken pair of binoculars. One lens is neoclassical economic theory and the idea of perfect rationality, perfect information, perfect trust. That's obviously wrong. The conditions of a rational actor possessed of perfect information and perfect trust engaging a series of independent, standalone, utility-maximizing behaviors is clearly nonsense.

For reasons of self-interest, I have a particular horror of this assumption: after all, in such a world of perfect information and perfect trust, marketing and advertising wouldn't need to exist. One of the reasons that people in economics have traditionally had an absolute horror of marketing is they believe there was an objective truth, a known valuation for everything, and that anything that interfered with the presentation of that truth was basically there to deceive people and to distort and effectively hack their perception or preferences.

I'm very, very interested in understanding how the brain processes information. I'm also hugely a fan of the understanding that what we have evolved a kind of heuristic toolkit. It's not, by the way, a constant toolkit either. We learn and replace and change the way we use of things. But in those things which we do reasonably regularly, we get pretty good at using those proxies. In other words, information sources that are readily available, computationally tractable and pretty good at helping us satisfice.

If you want me to explain, as an advertising person, brand preference—I'm not the first person to say this—Phillip Nelson and Toshio Yamagishi both got this—one perfectly sensible explanation is that someone who has a very valuable reputation that's been built up at great expense over many years is simply more reluctant to sell a bad product than someone who has no reputation to lose. If you want to put it very crudely, we like doing business with people we can hurt a bit. But if someone has a valuable but fragile reputation, we feel more comfortable buying from them than we feel buying from someone who's effectively anonymous and untouchable.

Our preference for brands, if you look at the reputational game theory, if such a field exists, there's nothing particularly irrational or foolish, and it's a basic proxy. I want to buy something, I'm happy to pay a little bit more for a feeling of reassurance and the unlikeliness of the product being a disaster or a rip-off, so I will buy from someone who is well-known—and has a reputation to lose.

The fact that we do this instinctively is not irrational—it's fallible, yes, but it's remarkably clever.

~  ~  ~

I'm pretty uncomfortable with the word "bias", because I'm not even sure you can say in most real world situations that there is a single best "rational" solution.

If you take a typical high school math question, it will be something like "two buses leave a town at an identical time, traveling at a constant 60 kilometers an hour, traveling in opposite directions. The reason you have this nonsensical imaginary world where buses travel at a constant 60 kilometers an hour is to give kids math problems to solve where there's a right answer. They're just buses, for God's sake. When does a bus ever travel at a constant speed? Okay, it doesn't happen. When do roads go in perfect straight lines? In the real world the information we have is kind of mucky; it's imperfect. What we have to do is latch onto a few of those reliable signals that might  help us make a reasonably quick decision without agonizing it over for ages, and use them. And probably then if they prove unreliable, then we might reject them. I would argue most people pay money for brands because most of the time when you buy something from a well-known brand, it doesn't disappoint. And that's, therefore, something that people continue to do in categories where that approach works.

~  ~  ~

One interesting thing that's regularly occurred to me in business, which no one in economics seems to have considered, is that the consumer surplus doesn't seem to bring any happiness, and that seems interesting. Would it be possible for you to enable people to understand the consumer surplus? Would that make them happy? For example, most people who are in the top quartile of disposable earnings, when they buy a flat screen television—first of all, it's worth remembering about a large plasma screen television that Louis XIV would've given you half of Burgundy in exchange for a modern television. So, if your frame of reference is: what would Louis XIV like? you would fill your house with electronic equipment and rejoice in every single last element.

The second thing is when you buy that television, you might be prepared, if such things cost that much, to pay $3,000 or $4,000 for a large plasma or LCD television. In reality the thing is priced at a level which effectively appeals to, let's say, 75 percent-80 percent of the market and you only have to pay about $800. But nobody leaves that shop going "That's extraordinary, miraculous. I am now extraordinarily happy because something I would have paid $3,000 for I can buy for 700 or 800." They've walked out of there with a net gain of $2,300 dollars, and yet it seems to generate no happiness at all. We simply think of things being worth what we pay for them. Now, if you solve that problem, you can synthesize an extraordinary amount of human joy—admittedly, mostly in the wealthier 50 percent of the populace. Thus, there is an interesting question, which is, people instinctively seem to think that framing is a dishonest activity—but can it help?

There was a very interesting finding fairly recently that Germans who go from being unemployed to being retired enjoy a very significant gain in happiness, even though their material state doesn't change at all. Simply by rebranding themselves retired rather than unemployed, they enjoy a gain in happiness that is equivalent to the gain and happiness enjoyed by newlyweds. Now, that has to be important, doesn't it? If you believe, as economists seem to do, that happiness can only depend on objective and financial conditions, you may be woefully wrong.

Jonathan Haidt is working on a book on this subject, which is the peculiar relationship we have with capitalism and business. Because we're overly preoccupied with intent rather than outcome, we don't really like businesses very much because any good work that they do is drowned out by the fact that their work is motivated by obvious self-interest, as much as they may try to disguise it with various activities and Corporate Social Responsibility activities. Nonetheless, we see our retailers and so forth as being self-interested and therefore we find it very difficult to admire or like them, whereas equally ostensibly philanthropic or altruistic entities, which may achieve little or nothing in the way of human progress or happiness, we have a disproportionate affection for. There was even a recent book called Pathological Altruism. If someone can find a way of solving that problem so that people had slightly more appreciation of activities which, although self-interested, seemed to result in fairly positive benefits for large numbers of people, that would be also a fairly interesting thing.

~  ~  ~

I went to speak to the Cambridge University Behavioral Economics Society and to be honest I was expecting a roomful of about 12 people sitting in a U shape around a table. And much to my surprise, there were about 170 people there. I asked 170 people "How many of you are studying economics?" and only three hands went up. Economists, i.e. purist economists, still don't seem to be very interested in the fact that arguably one should base one's theory on rational observed behavior rather than some elegant theory. However, there were about 25-30 people there from the business school. Business, for obvious reasons, is more interested in this than the conventional economic establishment. And you can quite reasonably understand why. Or, as Charlie Munger said, "if economics isn't behavioral, I don't know what the hell is."


I was a classicist at Cambridge. I joined as a graduate trainee, what was then Ogilvy & Mather Direct. Because it was the direct marketing division of Ogilvy, where we did advertising that had either coupons or reply cards, what you did was measurable. There was a feedback loop. Pretty quickly and early on, I realized that there was something going on there that nobody understood; that thing for which we have no name. Very small changes in the design of things would suddenly have immense effects in the number of people who replied, or the nature of the response, or people's readiness to pay—almost kind of butterfly effects. I was always fascinated by this, but there was no name for it and nobody in business was interested in quantifying it, or cataloguing it, or really investigating it in any kind of systematic way. I basically went quiet on it.


I was a classicist at Cambridge. I joined as a graduate trainee, what was then Ogilvy & Mather Direct. Because it was the direct marketing division of Ogilvy, where we did advertising that had either coupons or reply cards, what you did was measurable. There was a feedback loop. Pretty quickly and early on, I realized that there was something going on there that nobody understood; that thing for which we have no name. Very small changes in the design of things would suddenly have immense effects in the number of people who replied, or the nature of the response, or people's readiness to pay—almost kind of butterfly effects. I was always fascinated by this, but there was no name for it and nobody in business was interested in quantifying it, or cataloguing it, or really investigating it in any kind of systematic way. I basically went quiet on it.  I then spent quite a few years being very interested in Internet marketing because my brother—being an academic—had first used the Internet in 1987. I had quite a lot of advanced learning about how digital marketing worked.

Everything proceeded pretty much normally until about 2007, when I discovered Nudge and discovered a whole corpus of books on behavioral science. In 2009, I was appointed president of the IPA (Institute of Practitioners in Advertising). If you're an American, this is the equivalent to the Four-A's. It's the advertising agency trade body. Daniel Kahneman was, in fact, our second guest speaker. Just shortly after the publication of Thinking Fast and Slow, he came over and spoke to the IPA. I made it my mission to try and establish proper links between academic psychology and social science in the advertising industry.

Those links existed, by the way, in the 1960s. In the earlier episodes of Mad Men, there's a strange person floating around with a bowtie and a probably fake Viennese accent, who is going to be an in-house psychologist. There were some remarkable people. Louis B. Cheskin, who founded a kind of color institute. Bernays himself, obviously. Also, you knew someone called Howard Luck Gossage, who very, very early on founded something called Generalists, Inc., which was almost a precursor of behavioral economics. It was an advertising agency designed to look at human decision-making.

I thought that it's time we learn to understand this far, far better than the way things happened for the previous 20 years, by simply stumbling on things by accident. Broadly speaking, everybody assumes, as I said, there's this evil science thing in the marketing world. You may be guilty of base rate neglect there. The advertising world is much more left wing than you would expect. Everybody assumes it's full of Don Draper sociopathic individualists. In fact, politically I'd say the advertising industry is about representative of the population at large. Probably representative in no other way but it is politically representative.

The symbiosis here between what we discover in the real world and the academic work in properly codifying it and developing some sort of rigorous science around it, rather than simply pursuing lucky accidents in a kind of slapdash and opportunistic fashion, that struck me as having the potential for real progress.


There's another issue which is very, very important, which is that industry and government are now making extremely heavy use of market research. Anybody I talk to who's serious in academic psychology is deeply skeptical of self-reported anything. First of all, there are large parts of our brain and our decision-making apparatus that are opaque to introspection. We also have an extraordinarily dangerous tendency to post rationalize, when we don't know why we've done something, even deceiving ourselves. Timothy D. Wilson's book, Strangers to Ourselves, is pretty much un-improvable as a book title on the subject. We think we're much more rational than we are. I can't remember whether it was myself, Jonathan Haidt, or both of us who said that the rider on the elephant—the system two conscious brain—thinks it's the Oval Office when in reality it's mostly the press office.


But most important of all, even if you are an extreme leftist and you regard me as a horrendous materialistic bastard, we are united in one thing, which is that the influence that neoclassical economic thinking has on decision-making in business and, in particular, in policymaking is woefully disproportionate. Given that it's not really empirically based, the number of assumptions and axioms of neoclassical economics that have been allowed to pervade the business world in particular. By the way, I'm not criticizing really good economists here. What I'm criticizing is the kind of people who have done Econ 101 as part of some business school course have placed a mental tick box next to that mental model and then effectively rely on it without even consciously being aware of the influence it's had over their own thinking.

There's another issue which is very, very important, which is that industry and government are now making extremely heavy use of market research. Anybody I talk to who's serious in academic psychology is deeply skeptical of self-reported anything. First of all, there are large parts of our brain and our decision-making apparatus that are opaque to introspection. We also have an extraordinarily dangerous tendency to post rationalize, when we don't know why we've done something, even deceiving ourselves. Timothy D. Wilson's book, Strangers to Ourselves, is pretty much un-improvable as a book title on the subject. We think we're much more rational than we are. I can't remember whether it was myself, Jonathan Haidt, or both of us who said that the rider on the elephant—the system two conscious brain—thinks it's the Oval Office when in reality it's mostly the press office.

This is one reason why I've created Ogilvy Change, which is part of Ogilvy, which simply looks at behavioral science and what it can tell us about people that people themselves can't tell us, and which economics is wrong about. Then we can improve things quite a lot, or at least experiment.

First and foremost, let me give you a business example. The single best thing the London Underground did in terms of improving passenger satisfaction per pound spent wasn't faster, more frequent, later running trains, it was putting dot matrix display boards on the platform to tell you how long you were going to have to wait for your next train. There's something about the human brain, for whatever reason, which hates uncertainty. That's an interesting case because if you research how can you improve the Underground, most people would have said, "I want faster trains. I want more frequent trains." They would not have said, "I want less uncertainty."

Uber, the taxi-booking app, and Hailo, the British equivalent, are not that revolutionary in absolutely objective terms. You've always been able to book a cab by phone. You could pick up the phone and book a taxi. Big deal, okay. What makes Uber different is that when you phone for a taxi, in between that phone call and the taxi arriving, you enter the Twilight Zone of uncertainty. "Where is he? Why isn't he here yet? They said five minutes. I can't see him. Maybe he's outside. Should we go outside and have a look? What if he's left?" With Uber you watch the cab approach in real time on your map. And you go "Oh, look, he's stuck at those traffic lights. I'll make myself a cup of tea while I'm waiting." And you're both happier, you make better use of the time but you're also vastly less stressed in that period. Now, simply knowing that is really, really important. We don't like uncertainty.

Now, nobody in market research goes on about this much. Economists wouldn't understand this at all. They'd be entirely clueless about the effects of uncertainty, just as it's clueless about the effects of regret or fear of regret. But let's take this a step further, okay? If we know this and we also know that men are disproportionately reluctant to undergo certain kinds of medical testing, I can guarantee that quite sensibly and under the influence of good behavioral science, quite a lot of people have said, "Well, one way to get men to get tested for cancer more regularly is to make it easier and to use lots of nudges where you book appointments, you confirm appointments by text messaging and all that stuff." Absolutely true. And I commend that work, by the way.

There's another thing you can do which is interesting, and no one has tried this, which is to say, "If you have this test, we will give you the results in 24 hours." Because nobody's thought that's relevant. Nobody thought the delay between having the test and getting the result is relevant to the human propensity to have that test. Credit card companies discovered this by accident. The credit card people have discovered, they often said in the credit card ad "Apply now and you can get your approval within 12 hours." They clearly found, through testing or accident or experimentation or whatever, that this made a difference to people's propensity to respond. If you were simply using market research or simply using neoclassical economic assumptions, neither of those two things would tell you that time spent in a state of uncertainty might be an important factor. This is why I'm saying become a capitalist anthropologist, look at the behavior of consumers in small trivial money-grubbing ways. It's kind of like the Galapagos Islands for studying behavioral quirks.

Why do we prefer stripy toothpaste? When you think about it, once you put the toothpaste in your mouth, you mix it all up. Why does it need to be stripy? The strangest thing on the web is, there are hundreds of articles saying how do they make the stripes in toothpaste but there's no articles saying why. All those materials, the red, and the blue, and the white get mixed up in your mouth. It's completely pointless. Why do you do it? Something about the human brain just thinks if there are three different colors, it's easier to believe that that toothpaste is doing three different things: bamishes plaque, freshens breath, eliminates cavities. Because there are three colors, I find it easier to believe that this thing is doing three totally different things. As a greedy anthropologist, use these funny behaviors that are clearly part of our brain mechanism to solve bigger problems.

And then you can ask really interesting questions. What happens if the true reason for this problem isn't what people tell us it is? When you flew over from JFK to London, you probably found it irritating going through airport security. People do, okay? People just get really annoyed, taking their laptop out of their bag, removing their shoes, etc. I asked the question one day, "Why do people get irritated? First of all, it's there to keep them safe." But, secondly, I said, "It's not Guantanamo Bay. They're not being waterboarded. Why don't you just placidly take your shoes off, remove your laptop?"

The reasons people give for saying they're irritated are post-rationalizations of what's making them uncomfortable. If you take the rider on the elephant, which is Jonathan Haidt's model—think about the elephant as it gets disgruntled by things but it can't really talk—in many cases the rider is doing all the bullshitting about why I'm upset by this process and the elephant, which is the thing that's upset, can't really talk and can just kind of trumpet a bit or stamp its feet.

Let's do some experiments to find out if you can make airport security much less "system one irritating" by changing certain things. Is it the fact that this takes place in public? In other words, if you went into a little private cubicle, took all your stuff out of your bag, then in your own time returned to public view, placed your tray on the conveyer belt and went through, would it be less stressful? Is it subconsciously a feeling of humiliation in a public place that upsets us? Is it the fact that we're rushed, because there's a queue of people behind us and we feel if I cock up here and make a mistake and the thing goes bing, I'm going to have ten people behind me who go, "That guy's an idiot?" So is it fear of social stigma?


At this point, I'm excited because I genuinely believe you can synthesize happiness. Better choice architecture, better information, better structure of choices, fewer choices that are completely past-dependent and irreversible. You can improve things.


What's certainly true is if ever you go through a security lane by chance when there's nobody behind you, it's about 70 percent less stressful than if you have to do it at the head of a queue. My contention is that what we could do is do a series of experiments. You would design an airport security queue that was just as efficient or even faster than current queues but which didn't create the degree of anxiety and unhappiness this procedure seems to be creating at the moment.

At this point, I'm excited because I genuinely believe you can synthesize happiness. Better choice architecture, better information, better structure of choices, fewer choices that are completely past-dependent and irreversible. You can improve things. I'll give you one more example just to prove that this isn't just flash-in-the-pan nonsense.

Don't give people 24 white pills. If you want them to finish their antibiotics, don't give them 24 white pills. Give them 18 white pills and six blue ones and say, "When you finish the white pills, take the blue ones." The pills can, frankly, be identical, okay? People will simply be more likely to finish a course of treatment if it comes in two sequential colors. Fact. That's the mental equivalent of cars having steering wheels, not nose steering. It's just the way our brains work. Whether it's notionally rational or not is irrelevant.