2019 Conference

Rory Sutherland:    I'm going to start by asking something very cheeky of the financial services industry, which is what if they're wrong about everything? So generally if you want to win an argument in any financial, if you use economics, you're basically fine, aren't you? If you think about it in consumer decision making, what we're really trying to do is avoid regret. When we make a decision, if you want to simplify it very, very much, what we do when we make a decision is we're trying as far as we possibly can to reduce regret. When we make a decision in a corporate setting, we're trying to reduce blame, the risk of blame. Standard economics provides you with a very, very convenient blame reduction machine. It's rather like no one ever got fired for buying IBM. In the same way no one ever gets fired for making a suggestion which is economically rational.

I think there's a problem with that. I think the problem is this, if you think about it, in order to achieve it's mathematical neatness, economics makes some pretty enormous assumptions. It assumes that everybody makes a decision in conditions of perfect information. In other words, they know exactly what they wanted in advance. They know how much utility they're going to derive from it, and they know how much they're prepared to pay for it. And they completely trust the person who's providing it. Now for one thing, could we possibly have evolved that way? If our brains had evolved to be completely trusting and to assume that we knew to a huge degree of certainty what anything was worth, how much enjoyment we'd derive from every decision, we'd be a very, very strange person. Maybe there were people like that. I don't think they actually reproduced or survived very long.

The problem I think that comes when everybody assumes that economics is true, one of them, a smaller problem is that everybody in finance hates marketing. They really hate marketing, don't they? And that's because if you believe in your model of perfect efficiency in a world of perfect information and perfect trust, marketing wouldn't need to exist. So the typical finance guy sees marketing not as a source of value creation but as, at the very best, a necessary evil. It's essentially a cost to be minimized. It's not something which actually generates significant value. In that world, the only way you can actually improve things is effectively to improve the utility of a product or to reduce the price. There's nothing else you can do. And it turns business into a two dimensional game.

The first rule, if you think about it, when you try to make anything a science, the first rule is no magic. That's what science is really all about. Now in the physical sciences, if you're designing an aircraft, that works very well because you know you can define perfectly success in advance. You can define success in numerical objective terms, and therefore you can use science to determine whether or not you'll actually meet that objective. Don't get me wrong, when I fly back to London tomorrow, I don't want to think that the people checking the wheel nuts on the plane are wildly creative people. I don't want wildly creative or experimental people in air traffic control. Let's try reversing the runway just for the laws. No, I don't want that.

But there's a very, very large area of business decision making which involves human psychology where the same scientific approach isn't value, it's actually destructive. It creates a huge opportunity cost because as far as you see, there's no value to be created in perception at all. There's only value to be created in reality. And I think this leads to an enormous loss of potential value.

Just to give you an example, I met this guy for breakfast about 10 years ago. He was the Chief Executive at the time of the largest telecoms company in Britain called BT. And he just come from a meeting with city and financial analysts, and the meeting was all about one thing. It was about the fact that his broadband product to the home was the highest margin, highest price product, and also had the highest market share. Now call me old fashioned, you'd think in business that was a pretty good thing, wouldn't you? So you're selling a high margin product to a very large number of people. No. They're were giving him a hard time about it because although empirically you can sell high margin products to a large number of people. After all, what is Apple if it isn't doing that. They were convinced that home broadband was a commodity, and therefore essentially he'd either have to drop the price or he was going to lose market share.  Instead of taking him out for a slap up meal and saying, "Well done," they were effectively saying, "You have to do something about this. This extremely profitable state of affairs is unsustainable." Not because it doesn't exist in reality, but because it doesn't conform to economic theory. 

I checked six years after the meeting, and it was still the most expensive product in the market place and it had an even higher market share than it did before. So you have a strange thing in economics where people would almost rather be right in theory than right in practice. By the way, why is it possible that a large number of people might pay premium for a product? Well, if you understand that our brains have evolved with a very particular calibration towards risk. Now economists tend to think of utility as something that's additive. You get a bit of extra utility, and you add it onto your pile of utility. In evolutionary terms, it's not. It's multiplicative because apart from everything else, if you die at stage three, any utility at stage four is pretty much worthless. 

I've got a friend at the London Mathematical Laboratory who's a physicist. He also works at the Santa Fe Institute who says essentially the whole of economics is predicated on an idea of risk that's completely wrong. That no living organism would sensibly adopt, which explains why in evolutionary terms, we often make decisions which economists think are irrational.

Let me explain. These are four little targets, and they show high variance, low variance, high bias, low bias. In evolutionary terms, think about this, you have to make a decision. It could be who to marry. It could be what television to buy. It could be whether to climb a tree to rescue some cherries and eat them. Now in economic terms, it's very simple. You just work out the probable gain from climbing the tree, and reduce by 1% the value of the cherries by the 1% chance you'd die. Now if you think about it, that's a very sensible way to make a one off decision. But if you have that calibration of risk and you deployed it every day of your life, you'd be dead in about a month.

If you think about it, risk over time and amalgamated ensemble risk are the same thing. I'm not going to do this either you get permission. If I went to effectively, I went to 100 people and said, "I'll offer you $10 million to play Russian Roulette once." Quite a lot of them, a number of them would say, "Yes." If I went to the same people and said, "I'll offer you $1 billion to play Russian Roulette 100 times in a row," nobody would say yes. There's a difference between what you might call ensemble probability and time series probability. Economics doesn't understand this distinction.

Once you understand time series probability, we really, really want to reduce the chance. When you make a decision under uncertainty, you got to think of two things. What's the average outcome? In other words, how good is this going to be on average if I do this? But you've also got to add a second question, which we tend to do unconsciously, not consciously. We've also got to ask the question what's the worst that could happen?

And so as humans, we're happy to accept a degree of bias to minimize the risk of downside variance. So we'll pay a bit more for a branded television not because it's a better television but because the brand means there's a smaller chance of it being awful. Does that make sense? That every decision in conditions of uncertainty, you have to factor in yes, the average outcome matters and the higher the average outcome, the better. But you can't make those decisions without considering downside, particularly downside catastrophic probability.

And if you look at it mathematically, has anybody here come across the concept of ergodicity or heard the word ergodic? Don't worry because I hadn't until about a year ago. Don't look it up on Wikipedia either because blood will start to come out of your ears. It's a concept much beloved of sort of theoretical physicists and people like this. But I'm trying to explain it to you very simply. If you're adding things up, variance doesn't matter. 2+2+2+2+2+2 is the same as 1+3+1+3+1+3, whatever. If you're adding, doesn't matter. If you're multiplying, you want to reduce variance.

If you think about it in evolutionary terms, the things that survive and flourish are the things that optimize their time average growth rate. Now if you think about this, that explains mathematically human cooperation. If you've got two organisms and they say, "Right. Depending on what sort of day we have hunting, we'll either score three or two or one." And they say, "Right. If we both get a three, we both keep three. If we both get a one, we'll both have to put up with one. If we both get a two, we'll share two. But we're going to change the rules in one way. If I get a three and you get a one, we'll split it so we both get a two."

If you look at the world multiplicatively not additively, those people who cooperate and agree to share losses and share wins will actually grow at a faster rate and be more successful than the people who say, "Nope. Whatever I captures mine." 

By the way, this also mathematically explains the insurance industry if you think about it because the insurance industry is essentially a way of pooling… So taking risks which people wouldn't be happy to take in a time series and pooling them with a whole bunch of people. It explains the extraordinary dilemma… How many people have read Daniel Kahneman Thinking Fast and Slow? Anybody here? A few, good. Excellent. He had this extraordinary thing where he spoke to the board of directors of what I think was GE, and he went to the managers of the various divisions. And he said, "If I offered you a business decision where there was a 60% chance that you'd increase your profits by 50% at a 40% chance that your profits would fall by 30%, would you take those odds?" And all but two of them said, "No because there's a 40% chance I'd lose my job." The Chief Executive at the end of the table said, "But I'd want all of you to take that bet because net as a Chief Executive, we'd win out." So there's a fundamental misalignment when you aggregate probabilities versus when you treat them as if they're all in a row.

This comes down to the origins of meat sharing in primitive societies, as variance reduction. It comes down to ergodicity. Essentially this is the kind of physical definition where there's a distinction between what happens to a group of people at one moment of time and what happens over time. It's really useful to think about this by the way because most mathematical models and most spreadsheets don't capture it at all. They don't capture the distinction. 

If you think about it, it's because of this that Amazon Prime is necessary because 10 people don't mind paying £3 once a month to have something delivered, but one person isn't going to pay £3 10 times a month. So if Amazon wants to have frequent customers, it has to have something like Amazon Prime because the same thing that's acceptable for 10 people to do occasionally isn't acceptable for one person to do a lot. I went to the train companies in the UK and said, "Have you ever thought about train overcrowding like this?" They said, "What do you mean?" I said, "Well, 100 people who have to stand 10% of the time don't really matter that much. If you're a commuter, that's just the kind of crap that happens. But 10 people who have to stand 100% of the time, which in your model would look exactly the same, now you've got 10 really, really angry people." If you actually differentiate between those two, you'll find a way of solving the problem. What I suggested is you run two trains each day for annual season ticket holders only. So if you're actually making the journey every day, you'll be pretty much sure of a seat.

So those two concepts of ergodicity and also variance reduction. If you think about it, I had the most surreal conversation in my life. A dinner with an esteemed Tallab where we talked about the reason that McDonald's is the most popular restaurant in the world isn't because it's very, very good. It's because it's really good at not being terrible. If you think about it, I mean, all of us here have probably eaten at a McDonald's, and we've eaten at Michelin starred fancy restaurants. And let's be honest, the Michelin starred restaurant at it's best is quite a bit better. But I'd argue you're more likely to get ill at a Michelin star restaurant than you are at the Golden Arches. If you're looking at downside minimization, then McDonald's has rather a lot to say for it.

If you think about our approach to risk and uncertainty, the crazy thing about economics is it tries to be behavioral science without understanding biology. It tries to make the science of human behavior resemble physics, which is a ridiculous thing to try to do. You'll never get that degree of certainty because people's behavior will always vary depending on context. But it will also vary depending on just perception. You can describe the same thing in two different ways, and people will like it or they'll hate it. I'll show you a few examples of that later.

Does anybody here shop on Ebay? Yeah. A few of you do. Now if you were completely economically rational, which given you're a room full of people working in banking, maybe you are. If someone was selling a pair of let's say RayBan sunglasses for 20% less than everybody else but they had a feedback rating of 96%, rationally, that's a good deal, isn't it? Because he's 4% unreliable, but the price is 20% lower. Therefore you're getting 96% of a pair of sunglasses for 80% of the price. And an economically rational person should buy those sunglasses. What happens if you look at Ebay, nobody does. If you got a 96% approval rating, if you're selling sunglasses at all, it's probably at 50% of the price than the people with the 100% rating will pay. That's why it's perfectly normal for premium broadband from a well-established brand to sell very well because it isn't a commodity. What people are trying to do is they're going to say, "This is the most expensive broadband in the market by the most respected brand in the market. The one thing I can confidently say about that for which I will happily pay a bit extra is it's not going to be terrible."

The other thing economics doesn't understand is trust. Now I think banking lost enormous amount of trust when it ceased to be local. It gained efficiency but it lost trust. Let me explain why. Again, with reference to animal behavior. I think these are some sort of antelope. Now if you look at antelope, when they go grazing on the Serengeti or wherever they are, they could get better grass if they wandered off on their own, if they tried to become optimizers, maximizers. And each antelope went to the best grass there was, you'd get better grass. Why didn't they do this? The reason they don't do this is very simple, which is that if you're out on your own as an antelope, you have to spend about 50% of your time looking out for lions so you can only spend 50% of your time eating. If you're in a herd, you can spend 95% of your time eating and 5% of your time keeping an eye on your fellow antelope. And generally if all the other antelope are eating, you don't need to worry too much about lions. Your benefiting from 100 pairs of eyes, not one because the second one of the antelope gets jumpy, now you know you need to stop eating. 

In the same way when banking was local, you could trust your bank manager completely. Not only because everybody knew him and he knew everybody and he was a pillar of the community. There was the additional second order dimension to this. All his customers knew each other, which meant that they knew that you only had to be called out once cheating on person on their home contents insurance in a small town, and your reputation was toast throughout the town. You were no longer a member of the golf club. You were no longer respected. And that kept it honest. Not just the fact that it was local in terms of he knew the customers, the customers knew him. It was the fact that he knew the customers all knew each other, which made it potent.

I think the future of selling pensions for that reason because I believe instinctively humans are a bit of a herd species. And in the UK recently they did a thing where essentially it was a behavioral economics experiment where your employer put money into a pension for you, and you had to opt-out. And they expected the opt-out rate to be about 40%. It was about 5%. Nearly everybody who was enrolled in a pension with their employer by their employer stayed in the pension their employer had put them in. Why'd you think that might be? My theory is because when everybody in your company has the same pension as you do, you're an antelope in a herd.

I'll be realistic. I'm an advertising copywriter. I'm never going to go home in the evenings and check the performance of my pension plan. You could get away with ripping me off gloriously if you wanted to. Putting out the charges while I wasn't looking. I wouldn't notice for 15 years. And in my company, if the people who provide my company pension suddenly put out the charges by 2%, 95% of us won't notice. But some sad booger in the finance department will, and we'll all get to hear about it. So in the same way I think what a lot of antelope do while they're grazing is they just keep an eye on the most neurotic antelope. And you effectively outsourced the vigilance to him while you get on with eating the grass.

Once you understand that the mathematics and dynamics of making decisions in low trust environments is different, suddenly lots and lots of things which seem weird to us suddenly start to make sense. Just to give an example, brands I think mostly exist. This is an extraordinary story. I met a man who's now 90. He's based in Chicago. He founded the Oval of the Office in Chicago. He was talking to David Ogilvy in the 1950s, late '50s, early '60s. Joel Raphaelson was his name. He's the son of the script writer for all sorts of the… His father was a script writer for the Jazz singer out of Hollywood royalty. And he said to David Ogilvy, he said, "I don't think people pay a premium for brands. I don't think people choose brand B rather than brand A because it's better. I think they choose brand B rather than brand A because they're more certain that it's good." And it was an extraordinary advanced question in decision science to understand that people are paying a premium not necessarily for product superiority.

If you think about it, no brand can guarantee that their product to always superior. That would be an impossible task. Samsung makes very good televisions. They make very good phones, but they can't guarantee that every single one they bring out will be the best. What you might be paying premium for when you buy a brand is the reliable assurance since they got a huge amount of reputational skin in the game, it's much more costly for them to produce a bad product than it is someone you've never heard of. So our preference for famous brands shows quite a lot of second order, high second order social intelligence. Someone who like Samsung who's invested probably about £8 billion building up their reputation from being the poor man's alternative to Sony to being as good as Sony in the television space. They're not going to squirm to that sum cost by quickly selling a few crap televisions to make a quick buck. Does that make sense?

So I think what's fascinating about this is most of this we do not consciously, we do it unconsciously. When we buy an unbranded piece of equipment, we don't rationalize, "I don't really like buying this because I don't think the maker's got reputational skin in the game." We just feel a bit uneasy. And so in a way we pay a premium for brands for reasons we ourselves don't consciously understand. It's just that when we're buying non-brands, we feel more scared. I noticed that if you travel in France if you're British, you get weird gas stations. And it's your car but you got to fill up with a brand of gasoline you've never seen before. And deep down your lizard brain really doesn't like it.

You can even look at really small things, right? & Sons, what does that say? It says this company… or & Daughters. This company has been set up with one eye on posterity. A company that's old is going to be a bit more trustworthy than the company that's brand new because it's acquired more reputational capital over the years. Established 1707. Again, why do companies always put this information? Because actually it's not irrelevant. If you're looking for someone to trust, someone who's been in business 150 years has more to throw away. They've got more reputational capital at stake than someone who's just started. Why are bank buildings always incredibly big, marble, exotic things? Because someone who spent that much money on an immovable asset isn't planning to skip town. If you had a bank in a Winnebago caravan, going and depositing $50,000 in cash would make you pretty nervous. What this says is nobody who builds a building like that is interested in short term gain, and they're not planning to disappear any moment soon.

Bank architecture is very interesting for a second reason, by the way. Do you know what the second important facet? When you design a bank building, it has to look really permanent. It has to look expensive, and it also has to be really big inside because the last thing you want in a bank is a queue of people on the street because then everybody thinks there's a run on the bank. So if there is a long queue to withdraw money, a bank needs to keep it invisible.

Now we do the same thing, if you think about it, up front expense is proof of long term intention is a form of signaling which exists not just in finance, exists in marriage. That's what an engagement ring is, right? Every man in this room, if you get engaged, whatever you spend on an engagement ring, it has to be an amount of money that kind of hurts. There's also, by the way, a particular reason why it's very potent if you're a man giving jewelry or flowers to women as a present, which is that men by and large don't care at all about jewelry or flowers. So it's a very visible clear sign of actual sacrifice and generosity. So if you buy your wife flowers or you buy your wife jewelry, it involves some degree of sacrifice. There's very little self interest in it. Whereas if you buy your wife a drone or a quad bike for Christmas, there could be a high suspicion of self interest.

So these are things which are actually in weird kind of way they're meaningful because they're pointless. The bank doesn't need to spend all that money to function as a bank, but the very fact that it does suggests that it's a bank that in game theory terms is playing the repeat game, the long game, which tends to be positive sum in corporative, not the short one off game, which tends to involve deception, disappearance, and negative sum outcomes.
    
So in the same way, by the way, when I talked about that business of a bank talking to customers who all know each other. When you get married, you don't give your vows door to door. You don't email your vows to people. You gather everybody you know together in one place, and you make your vows there in the open to everybody all at once. Because only one person has to say, "He's already married," and the wedding's off. So a promise you make in mass media or in public is actually much, much more robust than a promise you make one person at a time because it only needs one person to know the truth, and your promise is now void. So I just think that understanding the whole concept of trust is really, really vital in finance. Because if you don't have trust, doesn't matter how good your product is, you won't be able to sell it. You can produce a product which objectively is fabulous if people don't trust you, nobody's buying it.
    
I always wanted an experiment to take a brilliant product and to sell it in an incredibly untrustworthy way to see what would happen. Now obviously nobody's going to volunteer to do this, but two Melbourne comedians for some reason got hold of Ed Sheeran, who is one of the hottest entertainment properties in the world. And they got hold of him for a morning in Melbourne, and they decided to sell tickets to a very, very short Ed Sheeran concert in the most untrustworthy way possible.

Andy:    Hello. Hi. 

Peep shows, they have a pretty bad name, normally associated with lewd content. But by definition, they don't have to be. So in an attempt to change that, we took one of the world's biggest performing artists, kept all his clothes on, and set up an Ed Sheeran peep show. Would anyone dare to believe what was written outside and come in to our very dungy looking venue?

How are you feeling? 

Ed Sheeran:    I don't really know what's going on to be honest.

Andy:    That was fair enough because we dressed Hamish as a fairly shady looking sperka in charge of getting customers. 

Hamish:    I got some Sheeran. Who wants some Sheeran.

Andy:    All I can hear, Hamish.

Ed Sheeran:    You think he can get anyone?

Andy:    I don't think we'll… It's going to be a brave soul.

Ed Sheeran:    I wouldn't come in to… If there was a dude with a beard and a hat saying, "Come in and see this."

Andy:    Ed was right. This was going to be tough.

Hamish:    You want to pay for Ed Sheeran for $2.

Speaker 5:    Insurance?

Hamish:    Do you want to pay for Ed Sheeran? Your loss.

What do you recon, big fellow? Got Ed Sheeran in here. Beautiful ginger head man sitting on a stool. What do you recon? $2. Got Ed Sheeran just sitting on a stool in there. You want him? $2. $2 for a peep. Think about it. It's actually a pretty good value.

Andy:    Despite trying, we had a total lack of interest for over 50 minutes.

It's been some time. We should've got you a more comfy chair I think.

Ed Sheeran:    I'm all right. I'm all right. 

Hamish:    Hey, big fellow. I got Ed Sheeran in there. $2. 

Speaker 6:    What is it? 

Hamish:    All the Ed Sheeran you need.

Speaker 6:    All the what? 

Hamish:    Ed Sheeran.

Speaker 6:    Oh, I don't know who that is. 

Hamish:    He's a singer. Yeah? Is that a yes? No. 

I think one of the big problems is people think Ed Sheeran's a code word for a new drug. 

How's it going? You guys like Ed Sheeran? $2. $2 for a 30 second peep. 

Ed Sheeran:    What? Are they just saying no?

Andy:    Yeah.

Ed Sheeran:    That's great.

Hamish:    Dirt cheap peep. Dirt cheap peep. There we go. $2.

Andy:    I think we're pricing it too high and that's why we're not getting people coming in.

Ed Sheeran:    I think $2 is pretty fair.

Hamish:    There we go boys. It's a Friday. Get your Ed Sheeran peep show. $2. Sitting on a stool, play you a song.

Andy:    If someone actually does think it's a peep show, I might quickly give you the go ahead to take off all your clothes. You're willing to do that.

Ed Sheeran:    I've been drinking a lot of beer recently.

Andy:    Oh right. You're not-

Ed Sheeran:    A couple of months ago maybe, but yeah, I'm in shape. It's just the shape of a potato.

Andy:    Two hours in, and Hamish was getting more desperate.

Hamish:    Ed Sheeran is literally sitting in there on the stage waiting for your $2. 

Andy:    We have a feeling that it's well… But just when we thought this had been a giant waste of everyone's time.

Hamish:    You guys like Ed Sheeran? You like Ed Sheeran? $2, peep show. Let's go. He's sitting on stage. $2.

Speaker 7:    Are you for real?

Hamish:    It's going to cost you $2. You only get 30 seconds though. You want to come in? 

Speaker 7:    I don't believe you.

Hamish:    Well, there's only one way to find out. 

Andy:    We might be on here.

Hamish:    There we go, Ed Sheeran peep show. He's there til midday. All right. Your choice.

Andy:    No. She did this smile thing and walked away. 

Ed Sheeran:    Oh.

Hamish:    Listen, if you guys want to have a go, he's sitting in there by himself. It'll probably get busy later on. $2, 30 seconds. I mean, you both can come if you want. Just $2 a head. 

Everything's above board I can assure you.

Speaker 8:    We're not going to get (bleep) are we?

Hamish:    No. Absolutely not. I can't guarantee what Ed will do, but yeah. $2.

Andy:    After two hours and 23 minutes, including some final hesitation, we finally found people brave enough to take a peep.

Hamish:    Do you guys go to peep shows a lot or-

Speaker 8:    No. (bleep). (bleep).

Hamish:    There you go, man. Keep your clothes on. Stay on the seat. Behave yourselves. Just listen to the announcement. Have a good one. Enjoy your peep.

Andy:    Welcome to the peep show. Your time will start in five seconds.

Speaker 7:    I'm actually a bit scared. 

Ed Sheeran:    (singing).

Speaker 7:    Oh my god. 

Ed Sheeran:    (singing).

Andy:    Your time is finished. Thank you very much.

Hamish:    All right, guys. There you go. It's been fun.

Rory Sutherland:    Now after the financial crisis, that must be how it feels selling a financial services product. Does that make sense? It's also worth remembering, if you just placed one ad in the middle of the newspaper and you charged $100 for it, you would have a queue around the block. But the context in which things are sold affects the value we attach to them. Now the only people who understood this by and large, they're probably about four of you. Anybody in the room who's a fan of Austrian School of Economics? They're always sitting at the back. Exactly. There's always about three, and they always sit at the back because they're terrified of the near liberals ganging up on them and bullying them. Austrian School of Economics had something a little like behavioral science called praxiology, which is the study of human behavior. And they see economics as the study of praxiology under conditions of scarcity. But they also understand that marketing and presentation are as much a part of value creation as manufacturing is. 

There's a great quote from Ludwig von Mises where he says, "There's no useful distinction to be made in economics between the value created in a restaurant by the man who makes the food and the value created by the man who sweeps the floor. One of them produces the food. The other one produces the context in which it's possible to enjoy the food." If you've got a great restaurant with a Michelin star chef but the restaurant smells slightly of sewage and there are mouse droppings on the floor, no one would enjoy the meal. So the notion of value is being independent from perception, as far as the Austrians is concerned, is complete nonsense. That's just a weird thing that essentially near liberal economists created as a way of making economics look like physics. Because the one thing physics needs is a context free law in order to look mathematical. But what that takes away is the possibility to use context to create value. You can create value by creating a great meal at a restaurant, but you can also create value by making the restaurant a fantastic place.

Now in physics, I accept that. We go the second law of thermal dynamics, energy cannot be created or destroyed. I accept that. There's no magic in physics. Economics, Milton Friedman got in on the act by saying there's no such thing as a free lunch. Well that's rubbish. I've had loads of them, right? And the point is that when you understand that unlike physics, economics because it involves human beings with their own sense of perception, allows you to create value through context or through meaning, not just through objective reality. It frees you up to perform a kind of magic. I'll give you just a few examples of this. Anybody familiar with… I'll give you this example. Save more tomorrow pension. Shlomo Benartzi with Richard Thaler's, now got a Nobel Prize for economics ironically, created a new form of pension saving. Instead of signing up to say $300 a month for your pension, you signed up a percentage of all your future pay rises. See, when you signed up, you might choose 20%, and the great thing is first of all didn't cost you anything immediately. Secondly, it never actually cost you downside money. You didn't get poorer, you got less richer as a way of paying for a pension.

Now to an economist, there's no distinction between those states because again they assume ergodicity, and they don't understand downside variance. To a human brain, there's a huge difference between less richer and poorer. These were the results. So the average savings rate was increased from 3.5% to 13.6% over the course of 40 months. Not only that but the 13.6% of people ended up saving at twice the rate that the 3.5% did. Take basically the same thing and present it in a different way and the behavior will be completely different.

One little tip for all of you in banking, every time you sign onto online banking or onto a mobile banking app, what's the first thing it shows you? Your balance. People don't want to see that. People would use their apps much more often if seeing your balance was an option, not a necessity. Generally as humans we don't willingly expose ourselves to disappointment and pain. So putting it on the first page may not be the best idea.
    
Now I'm not suggesting by the way that you don't need people with economic rigor and very, very good math skills in the financial industries. What I would say is there's a bias between the two. So that if you're rational, you're effectively unpleased. All creative people if I suggest something counterintuitive, if I suggest something a bit unusual, I have to present it to lots and lots of rational people for a cost benefit analysis, a feasibility study, et cetera. The same law does not apply the other way around. Once someone decides something using economic theory, they don't say, "Let's present it to some weird people first to see if they've got a different solution." 
    
So if crazy people are very heavily policed, I discovered this very interestingly quite often you can increase sales of things by putting the price up. KFC in South Africa, they had a product that wasn't selling. I said, "Before you put the price down, why don't you try putting the price up?" If you'd merely suggested the product's not selling, so we're going to drop the price by 20%. In a board meeting, that's approved in seven seconds because it's consonant with economic theory. If you go in and say, "Product's not selling, so we're going to put the price up and do an advertising campaign," that involves six months. For some reason, the barrier of proof for imaginative ideas is set about 10 yards higher than it is for boring, logical ideas. Even though in that case the logical idea is much the most expensive. Sure enough, we put the price up, demand went up. I think there's a very simply reason, by the way. You go to KFC for two reasons. You go for a bargain, or you go for a treat. And something that's priced in the middle isn't a bargain, and it isn't a treat. So nobody wants to buy it. 
    
But when I say you can use context to create value, you know that feeling you get when you land at an airport and the planes engines wind down. And you're still about a mile from the terminal building, and everybody on the plane has the same reaction, which is, "Oh, crap. It's going to be a bus." We just don't like being bused to the terminal. Now I'd always thought that. "Oh, god. It's one of those buses." And then one day I land and the pilot says, "I've got some bad news and some good news." He says over the intercom, he says, "The bad news is we won't be able to get you an air bridge because there's a plane blocking the gate. The good news is that the bus will take you all the way to passport control so you won't have far to walk to get to your bags." We all looked at each other and went, "Hold on. That's always true, isn't it? Actually I'm quite pleased there's a bus." Suddenly we saw it as a conveyance, not as an inconvenience. Next time you're on a plane and that happens, just say very loudly, "Actually, I'm glad there's a bus." You've synthesized happiness in everybody within an earshot. Because by getting people to think of something in a different way, you can change what it's worth. 
    
Who's got an espresso machine? Anybody? Yeah. Now they are insanely expensive, objectively. If you have to buy an espresso coffee in a jar like Nescafe, like Maxwell House, like Folgers, it would cost you about $40 for an equivalent dosage of an espresso coffee. And you'd look at it in a jar, and go, "That is bat shit crazy. I can't pay that. It's 10 times as expensive as the coffee I drink out of a can." But the thing is it doesn't come in a jar, it doesn't come in a big tin. It comes in an individual pod. And we don't know what an individual Maxwell House costs. So when we put that pod into our machine, we think, "Well, it's 39 cents, that would've been $2.80 at a Starbucks. This machine's basically making me money." And if you change the frame of reference, something that's expensive can become cheap. 
    
Rolls Royce and Maserati stopped selling their cars at car shows because a $400,000 car looks really expensive at a car show. They started selling them at yacht and aircraft shows because if you're looking at leer jets all afternoon, a $400,000 car is an impulse buy, right? And because evolution, evolution will sacrifice 25% of accuracy in order to gain 2% of fitness. It's not in our evolutionary interest to perceive the world objectively for whatever reason. So we perceive color. We perceive it contextually. We perceive price contextually. Taste contextually. Wine taste better if you pour it from a heavier bottle. Pain killers are more effective if they're expensive. Pain killers are more effective if they're branded. Your car drives better when you've had it cleaned, have you noticed that? It's not just a cleaner car, it's a smoother, quieter, better cornering, better performing car.
    
If you look at this little thing, if you use your hands just to cover up the joint on the left. I've done it for you on the right with a black bar. But if you just hold up your hands so you can't see the joint between the top and bottom of that thing, you'll see the bottom isn't white. It's exactly the same shade as gray as the top. Your brain has made the bottom white because it thinks it might be a shadow, and it's helpfully trying to reconstruct the correct color for you. Here's an even weird experiment. 

Speaker 9:    At any one moment, we are being bombarded by sensory information. Our brains do a remarkable job of making sense of it all. It seems easy enough to separate the sounds we hear from the sights we see. But there is one illusion that reveals this isn't always the case. 

Lawrence R.:    Bah, bah, bah, bah-

Speaker 9:    Have a look at this. What do you hear? 

Lawrence R.:    Bah, bah, bah. Bah, bah, bah. 

Speaker 9:    But look what happens when we change the picture.

Lawrence R.:    Bah, bah, bah. Bah, bah, bah. Bah, bah, bah-

Speaker 9:    And yet, the sound hasn't changed. In every clip, you are only ever hear bah with a B. 

Lawrence R.:    Bah.

Speaker 9:    It's an illusion known as the McGurk Effect. Take another look.

Lawrence R.:    Bah, bah, bah. 

Speaker 9:    Concentrate first on the right of the screen. Now to the left of the screen.

Lawrence R.:    Bah, bah, bah.

Speaker 9:    The illusion occurs because what you are seeing clashes with what you are hearing.

Lawrence R.:    In the illusion, what we see overrides what we hear. So the mouth movements we see as we look at a face can actually influence what we believe we're hearing. 

Rory Sutherland:    Every time you got a business problem, always look at it and say, "Is this best solved with objective reality or do we solve it with perception? Do we solve it with coming up with a new word or phrase? Do we solve it with describing something differently?" A brilliant example of that was the phrase designated driver was invented by marketers who said, "If there's a name for the behavior, it immediately becomes a social norm." So they went to Hollywood and they said, "Can you use plots where you use the phrase designated driver once or twice?" Because once there's a name for this behavior, it'll become much more common.

If you can invent a useful name for a behavior, you make it much easier for people to adopt because now it's normal. Weirdest one, for example, is does anybody here, if I gave you $500 and said to you, "You can keep that money if you put it into your pension before you get home this evening," how many of you would know how to do that? Top up your pension by $500 right now using a mobile phone or a laptop. Okay. Now I did this in London. One person put up their hand, and they worked for Goldman Sachs. Now a very sensible thing for people to do when they have a windfall is to put some of it into a pension. But it's difficult to do. I would have to go home, ask my wife for a strange filing cabinet where my pension records are kept, find an address, write a check. I'm not even sure I have a checkbook anymore, right? I have to post a check. I wouldn't get a text to say the money had arrived either. I'd have to remember to check in six months time when I next got my pension statement that the money had been credited. 

Now two things. One, if it's difficult to do, nobody's going to do it. But two, if it's difficult to do, people automatically infer that they're not supposed to do it. Because if I was supposed to top up my pension on an ad-hoc basis, people would've made it easy for me to do. So I won't do this because it's obviously weird. It's why we don't buy moist lavatory paper. Because most of the brands on the shelf for dry lavatory paper, and there's one brand of moist paper on the top shelf. And we go, "It's obviously for like weirdos and people with a medical condition." And in the same way, topping up your pension with a one off payment, even though it makes fantastic sense, feels weird. 

Remember what I said about uncertainty, we really, really hate uncertainty. When I pay my American Express card, I get a text about 10 seconds later saying, "We've received £200 or £1000 pounds or whatever." What do I do then? I usually send some more because I go, "Well, you got the first lot." If I had to pay money to my pension, it's like sending money into the ether. I've got no idea what's happened to it. 

Now I'd almost go as far as to say when you look at human psychology, one of the things you got to do is work out what a sensible person would do and ask, what would happen if we did the opposite? Now imagine you wanted to compete with Coca Cola, and you said, "For 100 years, it's been the most popular cold, non-alcoholic drink in the world," apart from water. How do we compete with Coke? And the logical people would immediately say, "We need a drink that tastes nicer than Coke. It needs to cost less than Coke, and we'll put it in a really big can so people get great value for money." 

Now no one would get fired for doing that. Everybody would feel completely sensible. People in research would say, "That's a great idea." There's only one problem. The most successful attempt to compete with Coke in the last 75 years is this, it comes in a tiny can. It costs a fortune, and it tastes disgusting. Now the reason it works is because context determines perception. When you position something as being medicinal or psychoactive or psychotropic, it's got to taste a bit weird. The fact that it's expensive makes it potent. The fact that it's in a small can says, "This is so powerful we've had to put it in a specially small container because if you have the full 330 milliliters, you'd go postal." Right? And so what something is is not so much dependent on what it actually objectively is. When they research the taste, everybody said things like, "I wouldn't drink this piss if you paid me to." And yet people pay a huge amount of money for this drink all the time.

Always ask the question, before we ask a question objectively, before we set a target objectively, is there a subjective question we could be asking which is more interesting? I work quite a bit with train companies. I said, "Here's a question. You want to reduce train overcrowding." I mentioned that earlier. Here's the most interesting question to ask, which is how can you get people to choose to stand on a train? Because if they choose to stand up on a train, train overcrowding isn't a problem anymore, right? I said, "Actually, you could do that." At the moment, on any commuter train, if you get a seat, you get a seat. You get a table. You get a plug. You get a view out of the window. You get everything, right? If you're standing on a train, you can't even read a book or use your mobile phone because one hand is needed to stop yourself falling over. You don't get a view. You don't anywhere to charge your phone, and you can't do anything. You feel cheated. 

I said, "Redesign a train so the seats are down the middle. They don't get a view, and they don't get a table. You do get a seat. Along the edge of the train, you get loads of little bum rests for people to lean on, and they get a little shelf to put their tablet on, and they get two USB chargers for their phone." Loads and loads of people under those conditions will actually… It's called adaptive preference formation. They'll actually say, "I'm glad I'm standing up because I get this view." 

We do this with pizza delivery. We said every single person seems to want their pizza delivered straight away. If you could get people to give you early notice or to be happy waiting a bit longer, you could put three deliveries on one bike instead of going deliver a pizza, back to base, deliver a pizza, back to base, deliver a pizza, back to base. You could go leave bike, leave base with three delivers, deliver one, deliver two, deliver three, back to base. So it'd be about twice as efficient, not three times as efficient maybe. But twice as efficient as doing it the present way. 

But they said, "Everybody wants their pizza delivered immediately." We said, "How would you get people to want their pizza a bit later." Every single app you produce, the default is ASAP. Every single website, the default is ASAP. If you ring up, they don't say, "When would you like it," they assume you want it as soon as possible. Change the default on the app, and also make it a time of day not a duration. So say 8:45 instead of 45 minutes. Make the default 8:45, 8:50, 8:55. People can still ask for it as soon as possible if they want, but change the norm.

You can push the default out as far as an hour without any loss in sales. You can put three pizzas on one bike. The weirdest thing of all is customer satisfaction went up by 50%. Before you look at a thing as a logistics question… Now if you're designing an aircraft, if you're working in physics, in engineering, you're dealing with objective reality. There's a single right answer. If you're dealing with human beings, there are always multiple right answers, and you can get to change the question. The Boeing 787 is designed brilliantly by engineers. But it's also designed by brilliant psychologists. If you make the entry way to the plane particularly spacious, it makes the whole of the plane feel more spacious. 

Here are five things humans really, really care about that economics doesn't understand. David Rock, he's a New Zealander based in New York, and he calls it SCARF. Status, certainty, autonomy, relatedness. They like to do business with the same people repeatedly. In other words, they don't see economic transactions as a single shot gain. And fairness. Now if you don't believe that we instinctively care about fairness, something economics doesn't understand at all. It just understands individual utility. It's not just humans who care about fairness. Have a quick look at this.
Speaker 11:    So final experiment that I want to mention to you is our fairness study, and so this became a very famous study. And there's now many more because after we did this about 10 years ago, it became very well known. We did that originally with Capuchin monkeys, and I'm going to show you the first experiment that we did. It has now been done with dogs and with birds and with chimpanzees. But we started out with Capuchin monkeys.

So what we did is we put two Capuchin monkeys side by side. Again, these animals, they live in a group. They know each other. We take them out of the group, put them in a test chamber, and there's a very simple task that they need to do. And if you give both of them cucumber for the task, the two monkeys side by side, they're perfectly willing to do this 25 times in a row. So cucumber, even though it's really only water in my opinion, but cucumber is perfectly fine for them.

Now if you give them partly grapes, the food preferences of my Capuchin monkeys correspond exactly with the prices in the supermarket. So if you give them grapes, there's a far better food, then you create inequity between them. So that's the experiment we did.

Recently, we videotaped it with new monkeys who never done the task thinking that maybe they would have a stronger reaction, and that turned out to be right. The one on the left is a monkey who gets cucumber. The one on the right is the one who gets grapes. The one who gets cucumber noted the first piece of cucumber is perfectly fine. The first piece he eats. Then she sees the other one getting grape, and you will see what happens. So she gives a rock to her, that's the task, and we give her a piece of cucumber and she eats it. The other one needs to give a rock to us, and that's what she does. She gets a grape, and she eats it. The other one sees that. She gives a rock to us now, gets again cucumber. 
    
She tests her rock now against the wall. She needs to give it to us, and she gets cucumber again. 

So this is basically the Wall Street protest that you see here. 

Rory Sutherland:    Now one of my contentions is that the really successful weird companies Dyson, Apple, Uber, in a sense are weirdly successful because they answered a psychological question where all their competitors were trying to answer a practical one, an objective one. No one was there saying, "I really want to spend a lot of money on a vacuum cleaner." But somehow Dyson tapped into the psychology. Apple, everybody else was asking, "What can this mobile phone do?" And Jobs asked the question, "What does it feel like while you're doing it?" And in the third case, I think Uber I think plays to about three things. It manages expectation. Before you book a car, it actually gives you a rough estimate as to how long you'll have to wait. So when the reality then becomes clear, you're much less upset. We really care about certainty. 

If you remember that on the SCARF model. The best thing London Underground did to improve passenger satisfaction per pound spend wasn't fastest, more frequent, longer trains. It was dot matrix displays on the platform. We'd rather wait nine minutes for a train knowing it's coming in nine minutes then wait five minutes for a train not knowing.

The Uber map is brilliant because the Uber map means that the normal uncertainty of, "Is it here yet? Where is he? They said 10 minutes. Why isn't he here?" Which we find painful is replaced by just looking at the map and going, "Oh. He's stuck at those traffic lights. He'll be here in a minute or two." There's even a bit of status in it. I don't know if anybody does this. I time my walk onto the sidewalk to coincide with my car pulling up because it makes you feel like Keyser Soze at the end of The Usual Suspects. You can't imagine Keyser Soze standing in the rain going, "Maybe that's my car over there." Right? It makes you feel like Louie the 14th. 

There's a bit of relatedness in the fact that they rate you and you rate them. And the really successful companies aren't the ones that look great on an economists spreadsheet. They're the ones that understand the things that economists don't.

The joy of this for me as an advertising person is first of all you can solve problems for many, many people, which doesn't necessarily involve bought media. For every person with a media budget and a problem, I always say there are 1000 people who've got a problem but no media budget, and we can talk to them. It also means, fascinatingly, you can ask really, really interesting questions not only about consumer behavior but about institutional behavior. 

So one of my contentions, which just bear it in mind because I think it's interesting. If you want diversity in employment, you can achieve it without quotas or imposition automatically by hiring people in groups. If you hire people one at a time, think about that variance reduction thing. When people only had one car… In my childhood, everybody had a saloon car. They had a sedan. Now when people have a two car household, they don't have two sedans. In fact, they probably don't even have one. They have like a small town car and an SUV. 

When you hire people in three's, four's, five's, six's, you look for complementarity. When you hire people one at a time, you look for narrow conformity. So you can create social, gender, ethnic, every other form of diversity, completely automatically if you hire people in groups. And there's proof of that, just to end, that's how I got my job. They had four jobs going. They decided on three people for the first three, and they couldn't decide who to give the fourth job to. And eventually someone I learnt years later said, "Why don't we just take a chance on the weirdo?" When someone's got one job, no one takes a chance on the weirdo. If there'd been four jobs going and I'd applied to all four of them and they'd been given to people seprately, I wouldn't have got any of them. I'm still there 30 years later. 

So the way in which we perceive things, the way in which choices presented to us, the language that's used, the frame of reference that's used affects our perception of value and our decision making just as much as what the objective thing is. That's why we're human. That's why my book's called Alchemy. It's out on about May the 9, available on Amazon.com and even other book sellers. It's why whenever you're given a question to answer, try answering it twice. Once in objective terms, and the second time subjectively just to see what happens. 

Thank you very much indeed.