Can Marketing Be Saved?

If the digital age has done nothing else for productivity, it has proved spectacularly effective at generating a supply of mantras.

For those of us old enough to remember the inception of the web, it is occasionally worth pausing to remember a few mantras often heard in the mid-90s, and to ask how well they have survived the test of time.

For example, in 1999 Wired's US edition predicted the rise of "infomediaries". I remember that. These seem not to have fully emerged. Nor, in truth, has the "hyperlocal" web really taken shape.

But another prediction has always stuck with me because it was made with so much confidence:

"There will be no customer loyalty on the internet, since a cheaper price is always just one click away."

That statement was almost universally believed at the time. In the late 90s it was rare to read an article in a business magazine that did not predict that online shopping would degenerate into an orgy of price-comparison sites linked to a plethora of online retailers. What do we see in reality? A kind of inverse-square law is in operation: the largest online retailer, Amazon, sells nearly five times more than its nearest US competitor, Staples.com, and ten times more than Walmart.com. In the UK its dominance is even more pronounced (I am excluding the online grocery services here, since the supplying of perishable goods is unavoidably different).

It is difficult for any physical retailer to enjoy physical proximity to all its customers, but mental proximity is a different matter – you can effectively monopolize that.

Why were all the economists so wrong on this question? Or, to paraphrase the Queen on the financial crisis, "Why did nobody see this coming?"

Well, it's only fair to say there are some perfectly conventional economic reasons why Amazon enjoys this supremacy. Its prices by and large are highly competitive.

It offers enormous choice.

Its service is good.

And it benefits from scale in its warehousing and in its negotiations with suppliers (though certainly no more than Walmart/Asda). It also profits from other network effects, for instance, in the volume of customer reviews it attracts and in its appeal to marketplace sellers.

All these are plausible reasons which conventional economists would advance to explain its success. And it seems to maintain its share price without making much of a profit.

But are these the only reasons? None of these post-rationalisations can on their own adequately explain Amazon's dominance. If, as economists believe, people really do check competitive prices before buying a book from Amazon, it doesn't explain why no one much even bothers to compete in the categories Amazon dominates. Try searching book titles on Google if you doubt this.

What if the biggest reasons for Amazon's supremacy are not economic but psychological? Or the product of "heuristics and biases", to use fashionable psychology lingo. These are the cognitive short-cuts we adopt while facing an influx of data presented to us in everyday situations.

Let me give you a few examples:


No 1- The Mere Availability Effect

This is a mental bias whereby we are inclined to adopt a course of action simply because it easily comes to mind. Note that this is not the same as simple "awareness" or "fame". After all, Ryman and Asda both enjoy equivalent name recognition to Amazon in the UK. "A brand's mental availability refers to the probability that a buyer will notice, recognise and/or think of a brand in buying situations," says Byron Sharp, professor of marketing science at the University of South Australia. "It depends on the quality and quantity of memory structures related to the brand".

He continues, "This is much more than awareness, whether that is top-of-mind awareness, recognition or recall. Indeed, all of these [conventional marketing] measures are flawed by the use of a single, a-situational cue".

So, regardless of overall fame or reputation, context matters. When I am in a mall, I am content to wander into Waterstones; when online, it takes less cognitive effort to think of Amazon. It is a mental default.

When I'm offline, whether I go to Tesco or Waitrose may depend on the time of day, my location, mood and a host of other variables, all of which cause me to distribute my retail spending across a plethora of brands. When online, these variables are far fewer; hence the winner-takes-all effect (sometimes called the "Matthew Effect", named after the passage in that gospel where it is suggested that those who already have, inevitably get more) is not diminished by the ease of comparison and switching. On the contrary, it's more extreme.

It is difficult for any physical retailer, even Asda, to enjoy physical proximity to all its customers. But mental proximity is a different matter. You can effectively monopolise that.


No 2 – Habituation and Defaults

Yes, sometimes we do what economists believe, and assiduously check every possible retail outlet to find the lowest cost. But, let's face it, our lives would be intolerable if we evaluated every alternative before doing anything. We rarely test-drive more than two cars before we commit to buy one, so why should we search hundreds of web pages before buying a DVD? Instead we fall back on a simple, default behaviour: "If nothing bad happened last time, do what I did last time".

The evolutionary basis for this default behaviour does not need much explanation. In risk-averse modes (all mail order has an element of uncertainty and risk) it is sensible to be conservative. This conservatism applies to our taste in food: we feel very comfortable eating food which tastes identical to food we have eaten in the past, since the very fact that we are contemplating the decision is evidence that it did not kill us in the past. Ray Kroc, the former owner of McDonald's, spotted this tendency: "People don't want the best burger in the world," he stated, "they want one which tastes just like the one they had last time".

I understand KFC: you go to the counter and tell them what you want. I understand The Ivy: you are seated at a table and someone brings you a menu. But Nando’s puts me in a complete funk. Where am I supposed to go? Where’s the bloody cutlery?

In a desktop internet scenario the force of habit is strengthened by a significant factor. First of all there are none of the other confounding factors (such as proximity or happenstance) that nudge us out of our well-trodden paths. But there is also a further boost to the "familiarity breeds contentment" in online shopping: the cognitive burden of using any unfamiliar website is really quite high. A site you have used 20 times in the last year can be used with a degree of unconscious fluency - whereas a new site requires painful, conscious mental effort, generating uncertainty and doubt. (Think of the first few kilometres you drive in a hire car, where all the controls are in a different place to the car you own at home - or the bemusement you experience in a traditional shop when they rearrange the shelving).

I call this feeling of unease "Nando's Syndrome" - from the uncomfortable feeling I experience as a 47-year-old man when dining at the eponymous South African restaurant chain. I understand KFC; you go to the counter and tell them what you want. I understand The Ivy; you are seated at a table and someone brings you a menu. But Nando's puts me in a complete funk. Is it self-service or not? Where am I supposed to go? How do I get a drink? Where is the bloody cutlery? Compared to, say, waterboarding or colonoscopy it is not an extreme form of mental torture, but these feelings of unease are the mental equivalent of a hangnail or a blister - we do everything we can to make them go away.

Again, habit amplifies the abundance of the Matthew Effect. The more often that people go to your site, the more likely they are to come back.


No 3 – Social Proof and Contagion

The other default by which we live all the time is "do what everyone else does". Economists choose to ignore this behaviour, or deride it as "irrational" for a rather self-serving reason: once it is accepted that one person's behaviour may affect that of another, it makes a horrendous mess of their orderly mathematical models. It takes you away from the field of modelling human behaviour as though it were neat, Newtonian physics and brings you into the Wieneresque world of complex systems, emergence and feedback loops, for which they are mathematically ill-equipped. It also throws a spanner into their beloved, Panglossian idea that markets are perfectly efficient.

But we would not be recognisably human at all had we not developed the sensible instinct of copying the behaviour of others. Given limited time and energy, to go with the flow of mass behaviour is both necessary for survival (your tribe can either all fight or all run away - both are preferable to half of you doing one thing, half the other), and cognitively efficient (it draws on collective knowledge and experience, rather than individual enquiry, and allows learned behaviour to spread much faster).

Going to the same chip shop as everyone else may not be a perfect solution but it is unlikely to be terrible. Going out on a limb is risky and mentally demanding. That slight feeling of unease you have on entering an unfamiliar restaurant when you discover you are the only diner. That’s the heuristic at work.

Gerd Gigerenzer, the world's foremost expert on heuristics, would refer to this copying approach as "ecologically rational". The dissident economist Alan Kirman might call it social rationality. This is not "rationality" in the narrow, individualistic sense defined by economists and rational-choice theorists, but a good strategy under the circumstances. Going to the same chip shop as everyone else may not be a perfect solution, but it is unlikely to be terrible. Going out on a limb is risky and mentally demanding. And many human accomplishments depend on social norms - no one can decide on their own what it means to be fashionable. That slight feeling of unease you have on entering an unfamiliar restaurant when you discover you are the only diner? That's the herd heuristic at work.

There are many areas of expenditure that are disproportionately affected by this heuristic - since people are inclined to settle upon a suboptimal but universal consensus rather than ploughing their own lonely perfectionist furrow. Being a vegan is just damn difficult - every time you have to dine out with carnivores you have to explain yourself. Smoking is now mildly embarrassing in all middle-class circles. Coke is not my favourite carbonated drink (I prefer Dr Pepper), but it is a social norm to offer it: you cannot offer only Dr Pepper to guests. To use the language of blood groups, Coke is the type O negative of carbonated drinks - the kind you can give to anyone.

Once again this heuristic tends to increase the Matthew Effect: "To him that hath, more shall be given" (Matthew 13:12).


No 4 – To a Game Theorist Customer Loyalty is Not Irrational

In any conventional economic model, loyalty is irrational. In fact a rational economist would never get married. ("Why should we get married?" an economist asks his girlfriend in psychoanalyst Stephen Grosz's The Examined Life, a collection of case histories. "I choose you every day."He then goes on to insist on a prenup).

This failure to understand the ecological rationality underlying loyalty is a product of the bizarre behavioural model that underlies conventional economics. In the economic model of rational behaviour, a series of essentially anonymous economic actors engage in one-off exchanges, with each possessed of perfect information and complete trust. This is a situation which happens in the real world somewhere between "rarely" and "never". There are and have always been asymmetries of information, questions of trust, and also future unknowns.

In these circumstances, when the identities of the buyer and seller are known to each other, and are known to be known to each other, use of game theory may recommend that you might be better off paying more to buy disproportionately from one supplier rather than spreading your custom meanly and thinly across several.

First of all, the kind of exchanges that take place within a non-anonymous iterated-game framework tend to be more trusting and mutually beneficial than those that take place as a one-off transaction. Economic study of actions in, of all places, the Marseilles fish market has consistently demonstrated this instinctive preference repeatedly to deal with the same suppliers, even when it costs you a little more. The difference in optimal strategies between one-off exchanges and repeated exchanges is worth stressing. It is this distinction that essentially differentiates the capitalism of the functioning market from the capitalism of the tourist souk.

The Souk-vendor also has no prospect of reputational damage through selling sub-standard merchandise (nor of any reputational gain through selling something especially good).

When you buy from a tourist souk, you are participating in a zero-sum game. There is almost no prospect of repeat purchase, since the tourist in the souk will be back on a plane home to Luton the following week. Hence the sole interest of the vendor is to sell at as high a price as possible a carpet of as execrable a quality as possible. He has no interest in the future satisfaction of his customer, since any satisfaction or disappointment will have no economic gain or loss to his future business, which depends simply on a steady stream of new suckers passing his stall every day.

The souk vendor also has no prospect of reputational damage through selling substandard merchandise (nor of reputational gain through selling something especially good). Tourists generally do not fly back to Luton on easyJet and spend the next week recommending that their friends avoid the third stall along in the Marrakech market. Even if they did, it is unlikely that this reputational retaliation would affect the carpet seller's business very much. (TripAdvisor and other rating technologies may be changing this. In certain categories, such as small hotels, TripAdvisor may be of considerable economic importance)

It is for this reason that any economic transaction with strangers in a strange place is so mentally harrowing. The usual reputational and repeat mechanisms which ensure fair dealings (and even a kind of altruism) no longer function; suddenly you are operating in a low-trust, high-deception marketplace. As I often advise people, "If you want a bad meal, go to a tourist restaurant - and if you want a really bad meal go to a tourist restaurant with a view." The worst service I have ever received anywhere in the world was in Granada, where the entire customer base for restaurants consists of transients - tourists visiting the Alhambra and a large body of students. Without the prospect of repeat custom or the risk of reputational retaliation, markets simply do not function very well. (It's one of the reasons why, in any tourist area, it often makes sense, in game-theory terms, to go to McDonald's - since they will care about your custom back in Luton, and about their reputation overall).

Contrast this souk-like uncertainty and mistrust with the relationship that exists between you and, say, your local butcher. He is anxious that you are not disappointed by his sausages, since he stands to lose your future custom if they are found to contain large quantities of horse. Hence, when your local butcher sells to you, he is mindful of your future satisfaction. He also stands to suffer significant collateral reputational damage, because you can go around the neighbourhood mouthing off about your disappointing cheval-burgers, and hurt his business still more than by simply boycotting him on the quiet.

And the butcher, since he knows who you are, can engage in a form of reciprocation, rewarding you in ways that transcend simple monetary exchange. If you shop at the same butcher each week, you can reasonably expect a slightly better turkey at Christmas, for instance, or the best cuts of steak.

Tourists generally do not fly back to Luton on Easyjet and spend the next week recommending that their friends avoid the third stall along in the Marrakech market.

But how does Amazon operate in this way? Well, unlike a large impersonal bookshop, Amazon does know who you are. It knows your transaction history. And you know that it knows. Hence if you have satisfactorily bought 20 books from them over the last year and then the twenty-first book fails to arrive in the post, you can reasonably expect that Amazon accords you "the benefit of the doubt" and resends the book. If I am a one-off customer, I can have no such expectation. What indication do they have that I am not a con artist? After all, they have no record of any past transaction with me where I have not issued a complaint.

For the same reason, aside from the obvious economic incentive of air miles, I prefer flying with airlines I fly with quite a lot when I belong to their frequent-flier programme. Because of my frequent-flier record, I know that they know I am fairly valuable to them. If there is only one seat left on the last flight out of Stockholm before the snows come down at Heathrow, I can reasonably expect that the seat goes to me and not to some random backpacker. On an airline I rarely use I have no such hope or expectation.

These social currencies of exchange are usually not mathematically expressible (there is no mathematical notation for the "tit" or the "tat") but they do exist in our heads. In evolutionary terms, tit-for-tat reciprocation predated the invention of money by a million years or more, and it is necessary in many social situations. It would be seen as ethically intolerable for anyone besides a trained economist if British Airways were to hold a spot auction at Stockholm airport to determine who gets that last seat - and so past value is probably the only socially acceptable and economically intelligent way of allocating it. In any non-anonymous exchange, loyalty buys you a kind of insurance. Do we understand this instinctively? Once you look, real life, as opposed to theoretical "markets", is full of game theory.

An engagement ring or an expensive wedding is, in effect, a game-theoretic device, an up-front expense that indicates long-term commitment. It is through these sunk costs that we come to trust one another. A market where we "rationally" shift our loyalties according to short-term expediency may work less well than a more loyal market when a mental or digital record exists of each individual's transactions over time.

It is this instinctive urge to reciprocate and to retaliate that may allow markets with imperfect or asymmetrical information to work. It is one way we get around the problem of information asymmetry: how can you trust the seller of a car when he knows the quality of his car far better than you do? You look to futurity (by repeatedly using the same local car dealer) or to reputational collateral (by buying from someone local, whose reputation you can hurt). Unless you are desperate, you do not buy a second-hand car from a bloke called Dave living 200km away, for whom the only contact information you have is the number of a pay-as-you-go mobile.

While living in London, I noticed how many of my friends, on buying their first car, would return to their homes in Yorkshire, Wales or somewhere else far away and arrange the purchase through their father or friend. In retrospect the reasons for this seem obvious: London is too fluid, large and anonymous a place for anyone selling a dodgy second-hand car to worry about endangering his reputation. In Yorkshire, the private seller of the car drank in the same pub as the buyer's father.

Since advertising is expensive and difficult to do well, the cost of advertising is also a virtual engagement ring, proffered to the potential consumer: the upfront expense entailed being proof of long-term intention for the product, the brand and the relationship.

I agree with the social scientist Jon Elster when he says it is difficult to understand social behaviour without an understanding of game theory. Yet most economic and marketing models are almost totally blind to it and its effects. But it again helps explain why the Matthew Effect is so pronounced in online retailing. In an uncertain field such as mail order, you are more likely to be trusted and respected when things go wrong if there exists a record of your previous transactions - as an indicator of your trustworthiness and a hint to the likely loss of future value should you choose to defect.

In my view much advertising expenditure probably works this way. Since advertising is expensive - and difficult to do well - the cost of advertising is also a virtual engagement ring proffered to the potential consumer; the upfront expense entailed being proof of long-term commitment to the product, the brand and the relationship. Advertising sometimes conveys information, of course. But much of it ostensibly conveys really very little that is new or compelling. But the act of advertising, especially in expensive media, is a form of information in itself. Since it takes time to recoup the cost of an advertising campaign, it only pays to run one when the advertiser has reasonable expectations of the long-term, widespread popularity of the product being advertised. The act of advertising your product is hence a valuable signal that the manufacturer has faith in its own product - equivalent to a racehorse owner betting heavily on his own horse. It is not irrational that we're influenced by such as an action - on the contrary, it shows a high degree of instinctive social intelligence.

Therefore, the idea that brand loyalty is irrational is not true. In real life, we cannot assume the good intentions of the people we deal with. We need to look for relationships we can trust. A preference for dealing repeatedly with people who have reputations to lose by ill-treating us seems far from irrational - it is the very basis on which all human dealings rest.

When that economist said to his girlfriend "I choose you every day", he fundamentally misunderstood the nature of human relationships, which require something more enduring than a commitment that is reassessed on a daily basis. She - rightly, I think - ditched him soon afterwards.

What I've listed here are just four psychological theories as to why Amazon enjoys unrivalled success - alongside its obvious virtues as a highly efficient and well-run business.

Now you may not agree with all of these. In fact you may not agree with any of them. But bear with me. You see, what is really significant about these mental biases is that they are for the most part completely neglected by business and government in their decision-making. Businesses today may be making decisions - to compete, to undercut, to launch products, to enter a market, to advertise - with no understanding of the powerful effect these biases may have on success or failure. Governments may define policy or tax schemes in a way which makes sense to economists, but which is psychologically blind.

Take, for instance, the recent riots over student fees. Part of the reason for the fury aroused by student tuition fees is that they were paid for by "a loan" - a concept that involves a student labouring for years under a debt the size of which is depressing enough to people who are rich. For the poorer student, the sums are enough to discourage all but the most confident from going to university at all. Yet you could have instigated exactly the same financial mechanism but called it a graduate tax, which would have meant no-one would have needed to know of the depressing debt figure at all.

This simple reframing technique was probably considered by nobody, since all the effort was put into designing the scheme, and no attention at all paid to its psychological effects.

Why do businesses and governments make such egregious errors in understanding these heuristics and biases? Because they are unaware of them. Why? Because, frankly, we are largely unaware of them ourselves.

In understanding customer behaviour, companies have traditionally used two tools to determine likely purchasing intent. One is the standard assumptions of neoclassical economic theory, which is psychologically blind: the kind of "perfect" transactions it models are almost never found in reality. In the economist's mind, people are calculating rationalists, merely seeking to maximise their own utility in a world of perfect information, and devoid of such concepts as uncertainty, mistrust, fear and regret. Yet the human is far less a rational calculating machine than a kind of anxious, moralising, herd-like, reciprocating, image-conscious, story-telling game theorist.

The other tool most widely used in ascertaining and predicting consumer preference is market research. This, in truth, is little better than the economic model, as we don't really know why we do what we do. We're good at pretending to know, or constructing plausible-sounding post-rationalisations of our behaviour. But the heuristics that influence and, at times, determine our actions operate on us at a purely unconscious level.

No cricketer knows the heuristic they use to catch a ball. People come up with explanations, but all are wrong. Only observation reveals how people catch balls - and it is the same heuristic for everyone. But we lack the introspective mental mechanisms to understand and relate what we are doing. It is instinctive and tacit.

Hence research cannot really uncover much in the way of explanation for people's behaviour, since the people you are researching often do not understand why they behave as they do.

Moreover, when in groups, the explanations people contrive may be more motivated by the urge to impress the strangers in the room than to give the researcher any true insights. (It isn't only companies which have marketing departments - the brain has a pretty active marketing function of its own.)

So what I am saying here is that both the standard tools we use to predict and model human behaviour are really quite bad. Our ability to model and predict may therefore be little better than, say, weather forecasting in the 1820s. Now this is the critical question. We shall, I suspect, never be able to predict human behaviour exactly - any more that we can ever issue a perfect weather forecast for next year. But what if we could just develop models that improved it by just a bit? We are, after all, starting from a low base…

Much as we all love to decry it, weather forecasting has improved markedly. A four-day forecast today is as accurate as a two-day forecast in 1985. The path of a hurricane can be predicted now within a course over 65% narrower than 40 years ago.

More important perhaps, weather forecasting has improved in another way too - it now acknowledges the limits to its powers. It has learned that the ability to forecast anything more than ten days ahead may be computationally impossible. Knowing the limits to your understanding is a form of intellectual progress in itself.

What if economics and market research were to achieve a similar leap forward both in its ability to predict and in its understanding of its limitations? What would the implications be for business and government efficiency and for economic growth and well-being?

Books such as Nudge and Thinking, Fast and Slow have topped the bestseller lists. "Big data" now provides real-time behavioural information that will make it easier to test smaller behavioural experiments. Computer modelling - the very technology that has transformed weather forecasting - better depicts the complexity of actual market behaviour than the naïve Newtonian models of conventional economics. Darwinian psychology provides insights into those parts of our instinctive nature that we can change - and those we can't. And I haven't even mentioned neuroscience. Will these ever provide us with a perfect means to predict and adjust human tastes and actions? Not a chance. On the other hand, how much would our understanding need to improve before we would notice a significant improvement in economic effects, in quality of life - and in the intelligence of political debate? Remember, we are starting from a very low base - where intelligent people can be diametrically wrong in their predictions of the likely shape of online retail.

If the success of new product launches were to grow from the current six percent to ten, how much more could profitably be spent on Research and Development? Is it from psychological progress, not from physical technology, that we should expect the greater parts of the next 20 years?

I hold out rather small hopes for the gains for further scientific innovation. Once you've progressed from horsepower to the Boeing 747, it is far harder to enjoy a similar increase in velocity again. But what if the next century were marked as the golden age of progress in social sciences? That seems far more feasible and desirable. Or, as the great Robert Trivers puts it: "In short, Darwinian social theory gives us a glimpse of an underlying symmetry and logic in social relationships which, when more fully comprehended by ourselves, should revitalise our political understanding and provide the intellectual support for a science and medicine of psychology. In the process it should also give us a deeper understanding of the many roots of our suffering."

It is not to the next Steve Jobs or Bill Gates that we should look for the next great revolution in economic life, but rather to thinkers such as Daniel Kahneman, Robert Kurzban, Dan Ariely, Robert Trivers, Gerd Gigerenzer, Timothy D Wilson and John Tooby.