Tag: Complexity

Rory Sutherland on The Psychology of Advertising, Complex Evolved Systems, Reading, Decision Making

“There is a huge danger in looking at life as an optimization problem.”

***

Rory Sutherland (@rorysutherland) is the Vice Chairman of Ogilvy & Mather Group, which is one of the largest advertising companies in the world.

Rory started the behavioral insights team and spends his days applying behavioral economics and evolutionary psychology to solve problems that conventionally advertising agencies haven't been able to solve.

In this wide-ranging interview we talk about: how advertising agencies are solving airport security problems, what Silicon Valley misses, how to mess with self-driving cars, reading habits, decision making, the intersection of advertising and psychology, and so much more.

This interview was recorded live in London, England.

Enjoy this amazing conversation.

“The problem with economics is not only that it is wrong but that it's incredibly creatively limiting.”

Listen

Transcript
A lot of people like to take notes while listening. A transcription of this conversation is available to members of our learning community or you can purchase one separetly.

***

If you liked this, check out all the episodes of the knowledge project.

Principles for an Age of Acceleration

MIT Media Lab is a creative nerve center where great ideas like One Laptop per Child, LEGO Mindstorms, and Scratch programming language have emerged.

Its director, Joi Ito, has done a lot of thinking about how prevailing systems of thought will not be the ones to see us through the coming decades. In his book Whiplash: How to Survive our Faster Future, he notes that sometime late in the last century, technology began to outpace our ability to understand it.

We are blessed (or cursed) to live in interesting times, where high school students regularly use gene editing techniques to invent new life forms, and where advancements in artificial intelligence force policymakers to contemplate widespread, permanent unemployment. Small wonder our old habits of mind—forged in an era of coal, steel, and easy prosperity—fall short. The strong no longer necessarily survive; not all risk needs to be mitigated; and the firm is no longer the optimum organizational unit for our scarce resources.

Ito's ideas are not specific to our moment in history, but adaptive responses to a world with certain characteristics:

1. Asymmetry
In our era, effects are no longer proportional to the size of their source. The biggest change-makers of the future are the small players: “start-ups and rogues, breakaways and indie labs.”

2. Complexity
The level of complexity is shaped by four inputs, all of which are extraordinarily high in today’s world: heterogeneity, interconnection, interdependency and adaptation.

3. Uncertainty
Not knowing is okay. In fact, we’ve entered an age where the admission of ignorance offers strategic advantages over expending resources–subcommittees and think tanks and sales forecasts—toward the increasingly futile goal of forecasting future events.”

When these three conditions are in place, certain guiding principles serve us best. In his book, Ito shares some of the maxims that organize his “anti-disciplinary” Media Lab in a complex and uncertain world.

Emergence over Authority

Complex systems show properties that their individual parts don’t possess, and we call this process “emergence”. For example, life is an emergent property of chemistry. Groups of people also produce a wondrous variety of emergent behaviors—languages, economies, scientific revolutions—when each intellect contributes to a whole that is beyond the abilities of any one person.

Some organizational structures encourage this kind of creativity more than others. Authoritarian systems only allow for incremental changes, whereas nonlinear innovation emerges from decentralized networks with a low barrier to entry. As Stephen Johnson describes in Emergence, when you plug more minds into the system, “isolated hunches and private obsessions coalesce into a new way of looking at the world, shared by thousands of individuals.”

Synthetic biology best exemplifies the type of new field that can arise from emergence. Not to be confused with genetic engineering, which modifies existing organisms, synthetic biology aims to create entirely new forms of life.

Having emerged in the era of open-source software, synthetic biology is becoming an exercise in radical collaboration between students, professors, and a legion of citizen scientists who call themselves biohackers. Emergence has made its way into the lab.

As a result, the cost of sequencing DNA is plummeting at six times the rate of Moore’s Law, and a large Registry of Standard Biological Parts, or BioBricks, now offers genetic components that perform well-understood functions in whatever organism is being created, like a block of Lego.

There is still a place for leaders in an organization that fosters emergence, but the role may feel unfamiliar to a manager from a traditional hierarchy. The new leader spends less time leading and more time “gardening”—pruning the hedges, watering the flowers, and otherwise getting out of the way. (As biologist Lewis Thomas puts it, a great leader must get the air right.)

Pull over Push

“Push” strategies involve directing resources from a central source to sites where, in the leader’s estimation, they are likely to be needed or useful. In contrast, projects that use “pull” strategies attract intellectual, financial and physical resources to themselves just as they are needed, rather than stockpiling them.

Ito is a proponent of the sharing economy, through which a startup might tap into the global community of freelancers and volunteers for a custom-made task force instead of hiring permanent teams of designers, programmers or engineers.

Here's a great example:

When the Fukushima nuclear meltdown happened, Ito was living just outside of Tokyo. The Japanese government took a command-and-control (“push”) approach to the disaster, in which information would slowly climb up the hierarchy, and decisions would then be passed down stepwise to the ground-level workers.

It soon became clear that the government was not equipped to assess or communicate the radioactivity levels of each neighborhood, so Ito and his friends took the problem into their own hands. Pulling in expertise and money from far-flung scientists and entrepreneurs, they formed a citizen science group called Safecast, which built its own GPS-equipped Geiger counters and strapped them to cars for faster monitoring. They launched a website that continues to share data – more than 50 million data points so far – about local environments.

To benefit from these kinds of “pull” strategies, it pays to foster an environment that is rich with weak ties – a wide network of acquaintances from which to draw just-in-time knowledge and resources, as Ito did with Safecast.

Compasses over Maps

Detailed maps can be more misleading than useful in a fast-changing world, where a compass is the tool of choice. In the same way, organizations that plan exhaustively will be outpaced in an accelerating world by ones that are guided by a more encompassing mission.

A map implies a straightforward knowledge of the terrain, and the existence of an optimum route; the compass is a far more flexible tool and requires the user to employ creativity and autonomy in discovering his or her own path.

One advantage to the compass approach is that when a roadblock inevitably crops up, there is no need to go back to the beginning to form another plan or draw up multiple plans for each contingency. You simply navigate around the obstacle and continue in your chosen direction.

It is impossible, in any case, to make detailed plans for a complex and creative organization. The way to set a compass direction for a company is by creating a culture—or set of mythologies—that animates the parts in a common worldview.

In the case of the MIT Media Lab, that compass heading is described in three values: “Uniqueness, Impact, and Magic”. Uniqueness means that if someone is working on a similar project elsewhere, the lab moves on.

Rather than working to discover knowledge for its own sake, the lab works in the service of Impact, through start-ups and physical creations. It was expressed in the lab’s motto “Deploy or die”, but Barack Obama suggested they work on their messaging, and Ito shortened it to “Deploy.”

The Magic element, though hard to define, speaks to the delight that playful originality so often awakens.

Both students and faculty at the lab are there to learn, but not necessarily to be “educated”. Learning is something you pursue for yourself, after all, whereas education is something that’s done to you. The result is “agile, scrappy, permissionless innovation”.

The new job landscape requires more creativity from everybody. The people who will be most successful in this environment will be the ones who ask questions, trust their instincts, and refuse to follow the rules when the rules get in their way.

Other principles discussed in Whiplash include Risk over Safety, Disobedience over Compliance, Practice over Theory, Diversity over Ability, Resilience over Strength, and Systems over Objects.

Frozen Accidents: Why the Future Is So Unpredictable

“Each of us human beings, for example, is the product of an enormously long
sequence of accidents,
any of which could have turned out differently.”
— Murray Gell-Mann

***

What parts of reality are the product of an accident? The physicist Murray Gell-Mann thought the answer was “just about everything.” And to Gell-Mann, understanding this idea was the the key to understanding how complex systems work.

Gell-Mann believed two things caused what we see in the world:

  1. A set of fundamental laws
  2. Random “accidents” — the little blips that could have gone either way, and had they, would have produced a very different kind of world.

Gell-Mann pulled the second part from Francis Crick, co-discoverer of the human genetic code, who argued that the code itself may well have been an “accident” of physical history rather than a uniquely necessary arrangement.

These accidents become “frozen” in time, and have a great effect on all subsequent developments; complex life itself is an example of something that did happen a certain way but probably could have happened other ways — we know this from looking at the physics.

This idea of fundamental laws plus accidents, and the non-linear second order effects they produce, became the science of complexity and chaos theory. Gell-Mann discussed the fascinating idea further in a 1996 essay on Edge:

Each of us human beings, for example, is the product of an enormously long sequence of accidents, any of which could have turned out differently. Think of the fluctuations that produced our galaxy, the accidents that led to the formation of the solar system, including the condensation of dust and gas that produced Earth, the accidents that helped to determine the particular way that life began to evolve on Earth, and the accidents that contributed to the evolution of particular species with particular characteristics, including the special features of the human species. Each of us individuals has genes that result from a long sequence of accidental mutations and chance matings, as well as natural selection.

Now, most single accidents make very little difference to the future, but others may have widespread ramifications, many diverse consequences all traceable to one chance event that could have turned out differently. Those we call frozen accidents.

These “frozen accidents” occur at every nested level of the world: As Gell-Mann points out, they are an outcome in physics (the physical laws we observe may be accidents of history); in biology (our genetic code is largely a byproduct of “advantageous accidents” as discussed by Crick); and in human history, as we'll discuss. In other words, the phenomenon hits all three buckets of knowledge.

Gell-Mann gives a great example of how this plays out on the human scale:

For instance, Henry VIII became king of England because his older brother Arthur died. From the accident of that death flowed all the coins, all the charters, all the other records, all the history books mentioning Henry VIII; all the different events of his reign, including the manner of separation of the Church of England from the Roman Catholic Church; and of course the whole succession of subsequent monarchs of England and of Great Britain, to say nothing of the antics of Charles and Diana. The accumulation of frozen accidents is what gives the world its effective complexity.

The most important idea here is that the frozen accidents of history have a nonlinear effect on everything that comes after. The complexity we see comes from simple rules and many, many “bounces” that could have gone in any direction. Once they go a certain way, there is no return.

This principle is illustrated wonderfully in the book The Origin of Wealth by Eric Beinhocker. The first example comes from 19th century history:

In the late 1800s, “Buffalo Bill” Cody created a show called Buffalo Bill's Wild West Show, which toured the United States, putting on exhibitions of gun fighting, horsemanship, and other cowboy skills. One of the show's most popular acts was a woman named Phoebe Moses, nicknamed Annie Oakley. Annie was reputed to have been able to shoot the head off of a running quail by age twelve, and in Buffalo Bill's show, she put on a demonstration of marksmanship that included shooting flames off candles, and corks out of bottles. For her grand finale, Annie would announce that she would shoot the end off a lit cigarette held in a man's mouth, and ask for a brave volunteer from the audience. Since no one was ever courageous enough to come forward, Annie hid her husband, Frank, in the audience. He would “volunteer,” and they would complete the trick together. In 1880, when the Wild West Show was touring Europe, a young crown prince (and later, kaiser), Wilhelm, was in the audience. When the grand finale came, much to Annie's surprise, the macho crown prince stood up and volunteered. The future German kaiser strode into the ring, placed the cigarette in his mouth, and stood ready. Annie, who had been up late the night before in the local beer garden, was unnerved by this unexpected development. She lined the cigarette up in her sights, squeezed…and hit it right on the target.

Many people have speculated that if at that moment, there had been a slight tremor in Annie's hand, then World War I might never have happened. If World War I had not happened, 8.5 million soldiers and 13 million civilian lives would have been saved. Furthermore, if Annie's hand had trembled and World War I had not happened, Hitler would not have risen from the ashes of a defeated Germany, and Lenin would not have overthrown a demoralized Russian government. The entire course of twentieth-century history might have been changed by the merest quiver of a hand at a critical moment. Yet, at the time, there was no way anyone could have known the momentous nature of the event.

This isn't to say that other big events, many bad, would not have precipitated in the 20th century. Almost certainly there would have been wars and upheavals.

But the actual course of history was in some part determined by small chance event which had no seeming importance when it happened. The impact of Wilhelm being alive rather than dead was totally non-linear. (A small non-event had a massively disproportionate effect on what happened later.)

This is why predicting the future, even with immense computing power, is an impossible task. The chaotic effects of randomness, with small inputs having disproportionate and massive effects, makes prediction a very difficult task. That's why we must appreciate the role of randomness in the world and seek to protect against it.

Another great illustration from The Origin of Wealth is a famous story in the world of technology:

[In 1980] IBM approached a small company with forty employees in Bellevue, Washington. The company, called Microsoft, was run by a Harvard dropout named bill Gates and his friend Paul Allen. IBM wanted to talk to the small company about creating a version of the programming language BASIC for the new PC. At their meeting, IBM asked Gates for his advice on what operating systems (OS) the new machine should run. Gates suggested that IBM talk to Gary Kildall of Digital Research, whose CP/M operating system had become the standard in the hobbyist world of microcomputers. But Kildall was suspicious of the blue suits from IBM and when IBM tried to meet him, he went hot-air ballooning, leaving his wife and lawyer to talk to the bewildered executives, along with instructions not to sign even a confidentiality agreement. The frustrated IBM executives returned to Gates and asked if he would be interested in the OS project. Despite never having written an OS, Gates said yes. He then turned around and license a product appropriately named Quick and Dirty Operating System, or Q-DOS, from a small company called Seattle Computer Products for $50,000, modified it, and then relicensed it to IBM as PC-DOS. As IBM and Microsoft were going through the final language for the agreement, Gates asked for a small change. He wanted to retain the rights to sell his DOS on non-IBM machines in a version called MS-DOS. Gates was giving the company a good price, and IBM was more interested in PC hardware than software sales, so it agreed. The contract was signed on August 12, 1981. The rest, as they say, is history. Today, Microsoft is a company worth $270 billion while IBM is worth $140 billion.

At any point in that story, business history could have gone a much different way: Kildall could have avoided hot-air ballooning, IBM could have refused Gates' offer, Microsoft could have not gotten the license for QDOS. Yet this little episode resulted in massive wealth for Gates and a long period of trouble for IBM.

Predicting the outcomes of a complex system must clear a pretty major hurdle: The prediction must be robust to non-linear “accidents” with a chain of unforeseen causation. In some situations this is doable: We can confidently rule out that Microsoft will not go broke in the next 12 months; the chain of events needed to take it under quickly is so low as to be negligible, no matter how you compute it. (Even IBM made it through the above scenario, although not unscathed.)

But as history rolls on and more “accidents” accumulate year by year, a “Fog of the Future” rolls in to obscure our view. In order to operate in such a world, we must learn that predicting is inferior to building systems that don't require prediction, as Mother Nature does. And if we must predict, must confine our predictions to areas with few variables that lie in our circle of competence, and understand the consequences if we're wrong.

If this topic is interesting to you, try exploring the rest of the Origin of Wealth, which discusses complexity in the economic realm in great (but readable) detail; also check out the rest of Murray Gell-Mann's essay on Edge. Gell-Mann also wrote a book on the topic called The Quark and the Jaguar which is worth checking out. The best writer on randomness and robustness in the face of an uncertain future, is of course Nassim Taleb, whom we have written about many times.

The Need for Biological Thinking to Solve Complex Problems

“Biological thinking and physics thinking are distinct, and often complementary, approaches to the world, and ones that are appropriate for different kinds of systems.”

***

How should we think about complexity? Should we use a biological or physics system? The answer, of course, is that it depends. It's important to have both tools available at your disposal.

These are the questions that Samuel Arbesman explores in his fascinating book Overcomplicated: Technology at the Limits of Comprehension.

[B]iological systems are generally more complicated than those in physics. In physics, the components are often identical—think of a system of nothing but gas particles, for example, or a single monolithic material, like a diamond. Beyond that, the types of interactions can often be uniform throughout an entire system, such as satellites orbiting a planet.

Biology is different and there is something meaningful to be learned from a biological approach to thinking.

In biology, there are a huge number of types of components, such as the diversity of proteins in a cell or the distinct types of tissues within a single creature; when studying, say, the mating behavior of blue whales, marine biologists may have to consider everything from their DNA to the temperature of the oceans. Not only is each component in a biological system distinctive, but it is also a lot harder to disentangle from the whole. For example, you can look at the nucleus of an amoeba and try to understand it on its own, but you generally need the rest of the organism to have a sense of how the nucleus fits into the operation of the amoeba, how it provides the core genetic information involved in the many functions of the entire cell.

Arbesman makes an interesting point here when it comes to how we should look at technology. As the interconnections and complexity of technology increases, it increasingly resembles a biological system rather than a physics one. There is another difference.

[B]iological systems are distinct from many physical systems in that they have a history. Living things evolve over time. While the objects of physics clearly do not emerge from thin air—astrophysicists even talk about the evolution of stars—biological systems are especially subject to evolutionary pressures; in fact, that is one of their defining features. The complicated structures of biology have the forms they do because of these complex historical paths, ones that have been affected by numerous factors over huge amounts of time. And often, because of the complex forms of living things, where any small change can create unexpected effects, the changes that have happened over time have been through tinkering: modifying a system in small ways to adapt to a new environment.

Biological systems are generally hacks that evolved to be good enough for a certain environment. They are far from pretty top-down designed systems. And to accommodate an ever-changing environment they are rarely the most optimal system on a mico-level, preferring to optimize for survival over any one particular attribute. And it's not the survival of the individual that's optimized, it's the survival of the species.

Technologies can appear robust until they are confronted with some minor disturbance, causing a catastrophe. The same thing can happen to living things. For example, humans can adapt incredibly well to a large array of environments, but a tiny change in a person’s genome can cause dwarfism, and two copies of that mutation invariably cause death. We are of a different scale and material from a particle accelerator or a computer network, and yet these systems have profound similarities in their complexity and fragility.

Biological thinking, with a focus on details and diversity, is a necessary tool to deal with complexity.

The way biologists, particularly field biologists, study the massively complex diversity of organisms, taking into account their evolutionary trajectories, is therefore particularly appropriate for understanding our technologies. Field biologists often act as naturalists— collecting, recording, and cataloging what they find around them—but even more than that, when confronted with an enormously complex ecosystem, they don’t immediately try to understand it all in its totality. Instead, they recognize that they can study only a tiny part of such a system at a time, even if imperfectly. They’ll look at the interactions of a handful of species, for example, rather than examine the complete web of species within a single region. Field biologists are supremely aware of the assumptions they are making, and know they are looking at only a sliver of the complexity around them at any one moment.

[…]

When we’re dealing with different interacting levels of a system, seemingly minor details can rise to the top and become important to the system as a whole. We need “Field biologists” to catalog and study details and portions of our complex systems, including their failures and bugs. This kind of biological thinking not only leads to new insights, but might also be the primary way forward in a world of increasingly interconnected and incomprehensible technologies.

Waiting and observing isn't enough.

Biologists will often be proactive, and inject the unexpected into a system to see how it reacts. For example, when biologists are trying to grow a specific type of bacteria, such as a variant that might produce a particular chemical, they will resort to a process known as mutagenesis. Mutagenesis is what it sounds like: actively trying to generate mutations, for example by irradiating the organisms or exposing them to toxic chemicals.

When systems are too complex for human understanding, often we need to insert randomness to discover the tolerances and limits of the system. One plus one doesn't always equal two when you're dealing with non-linear systems. For biologists, tinkering is the way to go.

As Stewart Brand noted about legacy systems, “Teasing a new function out of a legacy system is not done by command but by conducting a series of cautious experiments that with luck might converge toward the desired outcome.”

When Physics and Biology Meet

This doesn't mean we should abandon the physics approach, searching for underlying regularities in complexity. The two systems complement one another rather than compete.

Arbesman recommends asking the following questions:

When attempting to understand a complex system, we must determine the proper resolution, or level of detail, at which to look at it. How fine-grained a level of detail are we focusing on? Do we focus on the individual enzyme molecules in a cell of a large organism, or do we focus on the organs and blood vessels? Do we focus on the binary signals winding their way through circuitry, or do we examine the overall shape and function of a computer program? At a larger scale, do we look at the general properties of a computer network, and ignore the individual machines and decisions that make up this structure?

When we need to abstract away a lot of the details we lean on physics thinking more. Think about it from an organizational perspective. The new employee at the lowest level is focused on the specific details of their job whereas the executive is focused on systems, strategy, culture, and flow — how things interact and reinforce one another. The details of the new employee's job are lost on them.

We can't use one system, whether biological or physics, exclusively. That's a sure way to fragile thinking. Rather, we need to combine them.

In Cryptonomicon, a novel by Neal Stephenson, he makes exactly this point talking about the structure of the pantheon of Greek gods:

And yet there is something about the motley asymmetry of this pantheon that makes it more credible. Like the Periodic Table of the Elements or the family tree of the elementary particles, or just about any anatomical structure that you might pull up out of a cadaver, it has enough of a pattern to give our minds something to work on and yet an irregularity that indicates some kind of organic provenance—you have a sun god and a moon goddess, for example, which is all clean and symmetrical, and yet over here is Hera, who has no role whatsoever except to be a literal bitch goddess, and then there is Dionysus who isn’t even fully a god—he’s half human—but gets to be in the Pantheon anyway and sit on Olympus with the Gods, as if you went to the Supreme Court and found Bozo the Clown planted among the justices.

There is a balance and we need to find it.

The Map Is Not the Territory

Map and Territory

“(History) offers a ridiculous spectacle of a fragment expounding the whole.”
— Will Durant in Our Oriental Heritage

“All models are wrong but some are useful.”
— George Box

***

The Relationship Between Map and Territory

“That’s another thing we’ve learned from your Nation,” said Mein Herr, “map-making. But we’ve carried it much further than you. What do you consider the largest map that would be really useful?”

“About six inches to the mile.”

“Only six inches!” exclaimed Mein Herr. “We very soon got to six yards to the mile. Then we tried a hundred yards to the mile. And then came the grandest idea of all! We actually made a map of the country, on the scale of a mile to the mile!”

“Have you used it much?” I enquired.

“It has never been spread out, yet,” said Mein Herr: “the farmers objected: they said it would cover the whole country, and shut out the sunlight! So we now use the country itself, as its own map, and I assure you it does nearly as well.
Sylvie and Bruno Concluded

In 1931, in New Orleans, Louisiana, mathematician Alfred Korzybski presented a paper on mathematical semantics. To the non-technical reader, most of the paper reads like an abstruse argument on the relationship of mathematics to human language, and of both to physical reality. Important stuff certainly, but not necessarily immediately useful for the layperson.

However, in his string of arguments on the structure of language, Korzybski introduced and popularized the idea that the map is not the territory. In other words, the description of the thing is not the thing itself. The model is not reality. The abstraction is not the abstracted. This has enormous practical consequences.

In Korzybski’s words:

A.) A map may have a structure similar or dissimilar to the structure of the territory.

B.) Two similar structures have similar ‘logical’ characteristics. Thus, if in a correct map, Dresden is given as between Paris and Warsaw, a similar relation is found in the actual territory.

C.) A map is not the actual territory.

D.) An ideal map would contain the map of the map, the map of the map of the map, etc., endlessly…We may call this characteristic self-reflexiveness.

Maps are necessary, but flawed. (By maps, we mean any abstraction of reality, including descriptions, theories, models, etc.) The problem with a map is not simply that it is an abstraction; we need abstraction. Lewis Carroll made that clear by having Mein Herr describe a map with the scale of one mile to one mile. Such a map would not have the problems that maps have, nor would it be helpful in any way.

(See Borges for another take.)

To solve this problem, the mind creates maps of reality in order to understand it, because the only way we can process the complexity of reality is through abstraction. But frequently, we don’t understand our maps or their limits. In fact, we are so reliant on abstraction that we will frequently use an incorrect model simply because we feel any model is preferable to no model. (Reminding one of the drunk looking for his keys under the streetlight because “That’s where the light is!”)

Even the best and most useful maps suffer from limitations, and Korzybski gives us a few to explore: (A.) The map could be incorrect without us realizing it(B.) The map is, by necessity, a reduction of the actual thing, a process in which you lose certain important information; and (C.) A map needs interpretation, a process that can cause major errors. (The only way to truly solve the last would be an endless chain of maps-of-maps, which he called self-reflexiveness.)

With the aid of modern psychology, we also see another issue: the human brain takes great leaps and shortcuts in order to make sense of its surroundings. As Charlie Munger has pointed out, a good idea and the human mind act something like the sperm and the egg — after the first good idea gets in, the door closes. This makes the map-territory problem a close cousin of man-with-a-hammer tendency.

This tendency is, obviously, problematic in our effort to simplify reality. When we see a powerful model work well, we tend to over-apply it, using it in non-analogous situations. We have trouble delimiting its usefulness, which causes errors.

Let’s check out an example.

***

By most accounts, Ron Johnson was one the most successful and desirable retail executives by the summer of 2011. Not only was he handpicked by Steve Jobs to build the Apple Stores, a venture which had itself come under major scrutiny – one retort printed in Bloomberg magazine: “I give them two years before they're turning out the lights on a very painful and expensive mistake” – but he had been credited with playing a major role in turning Target from a K-Mart look-alike into the trendy-but-cheap Tar-zhey by the late 1990s and early 2000s.

Johnson's success at Apple was not immediate, but it was undeniable. By 2011, Apple stores were by far the most productive in the world on a per-square-foot basis, and had become the envy of the retail world. Their sales figures left Tiffany’s in the dust. The gleaming glass cube on Fifth Avenue became a more popular tourist attraction than the Statue of Liberty. It was a lollapalooza, something beyond ordinary success. And Johnson had led the charge.

With that success, in 2011 Johnson was hired by Bill Ackman, Steven Roth, and other luminaries of the financial world to turn around the dowdy old department store chain JC Penney. The situation of the department store was dour: Between 1992 and 2011, the retail market share held by department stores had declined from 57% to 31%.

Their core position was a no-brainer though. JC Penney had immensely valuable real estate, anchoring malls across the country. Johnson argued that their physical mall position was valuable if for no other reason that people often parked next to them and walked through them to get to the center of the mall. Foot traffic was a given. Because of contracts signed in the ’50s, ’60s, and ’70s, the heyday of the mall building era, rent was also cheap, another major competitive advantage. And unlike some struggling retailers, JC Penney was making (some) money. There was cash in the register to help fund a transformation.

The idea was to take the best ideas from his experience at Apple; great customer service, consistent pricing with no markdowns and markups, immaculate displays, world-class products, and apply them to the department store. Johnson planned to turn the stores into little malls-within-malls. He went as far as comparing the ever-rotating stores-within-a-store to Apple’s “apps.” Such a model would keep the store constantly fresh, and avoid the creeping staleness of retail.

Johnson pitched his idea to shareholders in a series of trendy New York City meetings reminiscent of Steve Jobs’ annual “But wait, there’s more!” product launches at Apple. He was persuasive: JC Penney’s stock price went from $26 in the summer of 2011 to $42 in early 2012 on the strength of the pitch.

The idea failed almost immediately. His new pricing model (eliminating discounting) was a flop. The coupon-hunters rebelled. Much of his new product was deemed too trendy. His new store model was wildly expensive for a middling department store chain – including operating losses purposefully endured, he’d spent several billion dollars trying to effect the physical transformation of the stores. JC Penney customers had no idea what was going on, and by 2013, Johnson was sacked. The stock price sank into the single digits, where it remains two years later.

What went wrong in the quest to build America’s Favorite Store? It turned out that Johnson was using a map of Tulsa to navigate Tuscaloosa. Apple’s products, customers, and history had far too little in common with JC Penney’s. Apple had a rabid, young, affluent fan-base before they built stores; JC Penney’s was not associated with youth or affluence. Apple had shiny products, and needed a shiny store; JC Penney was known for its affordable sweaters. Apple had never relied on discounting in the first place; JC Penney was taking away discounts given prior, triggering massive deprival super-reaction.

In other words, the old map was not very useful. Even his success at Target, which seems like a closer analogue, was misleading in the context of JC Penney. Target had made small, incremental changes over many years, to which Johnson had made a meaningful contribution. JC Penney was attempting to reinvent the concept of the department store in a year or two, leaving behind the core customer in an attempt to gain new ones. This was a much different proposition. (Another thing holding the company back was simply its base odds: Can you name a retailer of great significance that has lost its position in the world and come back?)

The main issue was not that Johnson was incompetent. He wasn’t. He wouldn’t have gotten the job if he was. He was extremely competent. But it was exactly his competence and past success that got him into trouble. He was like a great swimmer that tried to tackle a grand rapid, and the model he used successfully in the past, the map that had navigated a lot of difficult terrain, was not the map he needed anymore. He had an excellent theory about retailing that applied in some circumstances, but not in others. The terrain had changed, but the old idea stuck.

***

One person who well understands this problem of the map and the territory is Nassim Taleb, author of the Incerto series – Antifragile , The Black SwanFooled by Randomness, and The Bed of Procrustes.

Taleb has been vocal about the misuse of models for many years, but the earliest and most vivid I can recall is his firm criticism of a financial model called Value-at Risk, or VAR. The model, used in the banking community, is supposed to help manage risk by providing a maximum potential loss within a given confidence interval. In other words, it purports to allow risk managers to say that, within 95%, 99%, or 99.9% confidence, the firm will not lose more than $X million dollars in a given day. The higher the interval, the less accurate the analysis becomes. It might be possible to say that the firm has $100 million at risk at any time at a 99% confidence interval, but given the statistical properties of markets, a move to 99.9% confidence might mean the risk manager has to state the firm has $1 billion at risk. 99.99% might mean $10 billion. As rarer and rarer events are included in the distribution, the analysis gets less useful. So, by necessity, the “tails” are cut off somewhere and the analysis is deemed acceptable.

Elaborate statistical models are built to justify and use the VAR theory. On its face, it seems like a useful and powerful idea; if you know how much you can lose at any time, you can manage risk to the decimal. You can tell your board of directors and shareholders, with a straight face, that you’ve got your eye on the till.

The problem, in Nassim’s words, is that:

A model might show you some risks, but not the risks of using it. Moreover, models are built on a finite set of parameters, while reality affords us infinite sources of risks.

In order to come up with the VAR figure, the risk manager must take historical data and assume a statistical distribution in order to predict the future. For example, if we could take 100 million human beings and analyse their height and weight, we could then predict the distribution of heights and weights on a different 100 million, and there would be a microscopically small probability that we’d be wrong. That’s because we have a huge sample size and we are analysing something with very small and predictable deviations from the average.

But finance does not follow this kind of distribution. There’s no such predictability. As Nassim has argued, the “tails” are fat in this domain, and the rarest, most unpredictable events have the largest consequences. Let’s say you deem a highly threatening event (for example, a 90% crash in the S&P 500) to have a 1 in 10,000 chance of occurring in a given year, and your historical data set only has 300 years of data. How can you accurately state the probability of that event? You would need far more data.

Thus, financial events deemed to be 5, or 6, or 7 standard deviations from the norm tend to happen with a certain regularity that nowhere near matches their supposed statistical probability.  Financial markets have no biological reality to tie them down: We can say with a useful amount of confidence that an elephant will not wake up as a monkey, but we can’t say anything with absolute confidence in an Extremistan arena.

We see several issues with VAR as a “map,” then. The first that the model is itself a severe abstraction of reality, relying on historical data to predict the future. (As all financial models must, to a certain extent.) VAR does not say “The risk of losing X dollars is Y, within a confidence of Z.” (Although risk managers treat it that way). What VAR actually says is “the risk of losing X dollars is Y, based on the given parameters.” The problem is obvious even to the non-technician: The future is a strange and foreign place that we do not understand. Deviations of the past may not be the deviations of the future. Just because municipal bonds have never traded at such-and-such a spread to U.S. Treasury bonds does not mean that they won’t in the future. They just haven’t yet. Frequently, the models are blind to this fact.

In fact, one of Nassim’s most trenchant points is that on the day before whatever “worst case” event happened in the past, you would have not been using the coming “worst case” as your worst case, because it wouldn’t have happened yet.

Here’s an easy illustration. October 19, 1987, the stock market dropped by 22.61%, or 508 points on the Dow Jones Industrial Average. In percentage terms, it was then and remains the worst one-day market drop in U.S. history. It was dubbed “Black Monday.” (Financial writers sometimes lack creativity — there are several other “Black Monday’s” in history.) But here we see Nassim’s point: On October 18, 1987, what would the models use as the worst possible case? We don’t know exactly, but we do know the previous worst case was 12.82%, which happened on October 28, 1929. A 22.61% drop would have been considered so many standard deviations from the average as to be near impossible.

But the tails are very fat in finance — improbable and consequential events seem to happen far more often than they should based on naive statistics. There is also a severe but often unrecognized recursiveness problem, which is that the models themselves influence the outcome they are trying to predict. (To understand this more fully, check out our post on Complex Adaptive Systems.)

A second problem with VAR is that even if we had a vastly more robust dataset, a statistical “confidence interval” does not do the job of financial risk management. Says Taleb:

There is an internal contradiction between measuring risk (i.e. standard deviation) and using a tool [VAR] with a higher standard error than that of the measure itself.

I find that those professional risk managers whom I heard recommend a “guarded” use of the VAR on grounds that it “generally works” or “it works on average” do not share my definition of risk management. The risk management objective function is survival, not profits and losses. A trader according to the Chicago legend, “made 8 million in eight years and lost 80 million in eight minutes”. According to the same standards, he would be, “in general”, and “on average” a good risk manager.

This is like a GPS system that shows you where you are at all times, but doesn’t include cliffs. You’d be perfectly happy with your GPS until you drove off a mountain.

It was this type of naive trust of models that got a lot of people in trouble in the recent mortgage crisis. Backward-looking, trend-fitting models, the most common maps of the financial territory, failed by describing a territory that was only a mirage: A world where home prices only went up. (Lewis Carroll would have approved.)

This was navigating Tulsa with a map of Tatooine.

***

The logical response to all this is, “So what?” If our maps fail us, how do we operate in an uncertain world? This is its own discussion for another time, and Taleb has gone to great pains to try and address the concern. Smart minds disagree on the solution. But one obvious key must be building systems that are robust to model error.

The practical problem with a model like VAR is that the banks use it to optimize. In other words, they take on as much exposure as the model deems OK. And when banks veer into managing to a highly detailed, highly confident model rather than to informed common sense, which happens frequently, they tend to build up hidden risks that will un-hide themselves in time.

If one were to instead assume that there were no precisely accurate maps of the financial territory, they would have to fall back on much simpler heuristics. (If you assume detailed statistical models of the future will fail you, you don’t use them.)

In short, you would do what Warren Buffett has done with Berkshire Hathaway. Mr. Buffett, to our knowledge, has never used a computer model in his life, yet manages an institution half a trillion dollars in size by assets, a large portion of which are financial assets. How?

The approach requires not only assuming a future worst case far more severe than the past, but also dictates building an institution with a robust set of backup systems, and margins-of-safety operating at multiple levels. Extra cash, rather than extra leverage. Taking great pains to make sure the tails can’t kill you. Instead of optimizing to a model, accepting the limits of your clairvoyance.

The trade-off, of course, is short-run rewards much less great than those available under more optimized models. Speaking of this, Charlie Munger has noted:

Berkshire’s past record has been almost ridiculous. If Berkshire had used even half the leverage of, say, Rupert Murdoch, it would be five times its current size.

For Berkshire at least, the trade-off seems to have been worth it.

***

The salient point then is that in our march to simplify reality with useful models, of which Farnam Street is an advocate, we confuse the models with reality. For many people, the model creates its own reality. It is as if the spreadsheet comes to life. We forget that reality is a lot messier. The map isn’t the territory. The theory isn’t what it describes, it’s simply a way we choose to interpret a certain set of information. Maps can also be wrong, but even if they are essentially correct, they are an abstraction, and abstraction means that information is lost to save space. (Recall the mile-to-mile scale map.)

How do we do better? This is fodder for another post, but the first step is to realize that you do not understand a model, map, or reduction unless you understand and respect its limitations. We must always be vigilant by stepping back to understand the context in which a map is useful, and where the cliffs might lie. Until we do that, we are the turkey.

Making Decisions in a Complex Adaptive System

complexadaptive

In Think Twice: Harnessing the Power of Counterintuition, Mauboussin does a good job adding to the work we've already done on complex adaptive systems:

You can think of a complex adaptive system in three parts (see the image at the top of this post). First, there is a group of heterogeneous agents. These agents can be neurons in your brain, bees in a hive, investors in a market, or people in a city. Heterogeneity means each agent has different and evolving decision rules that both reflect the environment and attempt to anticipate change in it. Second, these agents interact with one another, and their interactions create structure— scientists often call this emergence. Finally, the structure that emerges behaves like a higher-level system and has properties and characteristics that are distinct from those of the underlying agents themselves. … The whole is greater than the sum of the parts.

***

The inability to understand the system based on its components prompted Nobel Prize winner and physicist Philip Anderson, to draft the essay, “More Is Different.” Anderson wrote, “The behavior of large and complex aggregates of elementary particles, it turns out, is not to be understood in terms of the simple extrapolation of the properties of a few particles. Instead, at each level of complexity entirely new properties appear.”

Mauboussin comments that we are fooled by randomness:

The problem goes beyond the inscrutable nature of complex adaptive systems. Humans have a deep desire to understand cause and effect, as such links probably conferred humans with evolutionary advantage. In complex adaptive systems, there is no simple method for understanding the whole by studying the parts, so searching for simple agent-level causes of system-level effects is useless. Yet our minds are not beyond making up a cause to relieve the itch of an unexplained effect. When a mind seeking links between cause and effect meets a system that conceals them, accidents will happen.

***
Misplaced Focus on the Individual

One mistake we make is extrapolating the behaviour of an individual component, say an individual, to explain the entire system. Yet when we have to solve a problem dealing with a complex system, we often address an individual component. In so doing, we ignore Garrett Hardin's first law of Ecology, you can never do merely one thing and become a fragilista.

That unintended system-level consequences arise from even the best-intentioned individual-level actions has long been recognized. But the decision-making challenge remains for a couple of reasons. First, our modern world has more interconnected systems than before. So we encounter these systems with greater frequency and, most likely, with greater consequence. Second, we still attempt to cure problems in complex systems with a naïve understanding of cause and effect.

***

When I speak with executives from around the world going through a period of poor performance, it doesn't take long for them to mention they want to hire a star from another company. “If only we had Kate,” they'll say, “we could smash the competition and regain our footing.”

At first, poaching stars from competitors or even teams within the same organization seems like a winning strategy. But once the star comes over the results often fail to materialize.

What we fail to grasp is that their performance is part of an ecosystem and removing them from that ecosystem — that is isolating the individual performance — is incredibly hard without properly considering the entire ecosystem. (Reversion to the mean also likely accounts for some of the star's fading as well).

Three Harvard professors concluded, “When a company hires a star, the star’s performance plunges, there is a sharp decline in the functioning of the group or team the person works with, and the company’s market value falls.”

If it sounds like a lot of work to think this through at many levels, it should be. Why should it be easy?

Another example of this at an organizational level has to do with innovation. Most people want to solve the innovation problem. Ignoring for a second that that is the improper framing, how do most organizations go about this? They copy what the most successful organizations do. I can't count the number of times the solution to an organization's “innovation problem” is to be more like Google. Well-intentioned executives blindly copy approaches by others such as 20% innovation time, without giving an ounce of thought to the role the ecosystem plays.

Isolating and focusing on an individual part of a complex adaptive system without an appreciation and understanding of that system itself is sure to lead to disaster.

***
What Should We Do?

So this begs the question, what should we do when we find ourselves dealing with a complex adaptive system? Mauboussin provides three pieces of advice:

1. Consider the system at the correct level.

Remember the phrase “more is different.” The most prevalent trap is extrapolating the behavior of individual agents to gain a sense of system behavior. If you want to understand the stock market, study it at the market level. Consider what you see and read from individuals as entertainment, not as education. Similarly, be aware that the function of an individual agent outside the system may be very different from that function within the system. For instance, mammalian cells have the same metabolic rates in vitro, whether they are from shrews or elephants. But the metabolic rate of cells in small mammals is much higher than the rate of those in large mammals. The same structural cells work at different rates, depending on the animals they find themselves in.

2. Watch for tightly coupled systems.

A system is tightly coupled when there is no slack between items, allowing a process to go from one stage to the next without any opportunity to intervene. Aircraft, space missions, and nuclear power plants are classic examples of complex, tightly coupled systems. Engineers try to build in buffers or redundancies to avoid failure, but frequently don’t anticipate all possible contingencies. Most complex adaptive systems are loosely coupled, where removing or incapacitating one or a few agents has little impact on the system’s performance. For example, if you randomly remove some investors, the stock market will continue to function fine. But when the agents lose diversity and behave in a coordinated fashion, a complex adaptive system can behave in a tightly coupled fashion. Booms and crashes in financial markets are an illustration.

3. Use simulations to create virtual worlds.

Dealing with complex systems is inherently tricky because the feedback is equivocal, information is limited, and there is no clear link between cause and effect. Simulation is a tool that can help our learning process. Simulations are low cost, provide feedback, and have proved their value in other domains like military planning and pilot training.

Still Curious? Think Twice: Harnessing the Power of Counterintuition.