On Its Head

On Its Head

You Can Only Manage What You Can’t Measure

Inside the box thinking

Why Optimization Is Often Sub-Optimal

Simplicity Is Overrated

Right-Brain Analytics

The New LBO: Left-Brain Outsourcing

The Return of Human Judgment

Business Physics: The Law of Value Conservation Does Not Apply in An Open System

Homeland Security Is Too Serious To Be Left To Government

Augmented Paranoia

Packaging Hell: The Incredible Cost of Anti-Shoplifting Shells And User Manuals

The Valley Of Death

The Capacity Paradox

Adding capacity to a system often results in counter-intuitive behavior

Special Operations and the Non-Profit Sector have much more in common than they think –and what they could learn from each other

Resilience at the Edge

The Incentive Conundrum

Software As A Service, 21st Century Style (DayJets)

Discovering The Unexpected

Business Model Inversions: e.g., a health insurance company becoming a data company with an insurance business to feed the database; ...

Search vs Attract: crowdsourcing, Innocentive

Human computer

Incentives (and values, Israel day care center)

Power to the edge

Your people are not your greatest assets

We create things we don’t understand

Constraints can be liberating

http://techcrunch.com/2012/03/06/khan-academy-inspired-flip-of-doctor-office-visit/

Intuition

Noitiutin

In a complex world:

Too much of a good thing can be a bad thing (Barry Schwartz: a little bit of friction is good; our humanity is lost without the spaces between the words)

Optimization is often sub-optimal (can’t take everything into account)

Thinking outside is easier and less useful and thinking within the box. Constraints can be liberating.

Experts are not as good as groups of non-experts.

Cognitive heuristics fail us

Incentives backfire

Cambrian explosions and losses of diversity (financial markets, before/after iphone)

Tom Friedman + Richard Florida = Dan Pink

Left-brain interactions and flattening, while right-brain interactions are spike-inducing.

Liberating data

Last week, I contributed $5,000 to help publish a book by a friend, David Stephenson. A few years ago, we co-authored an article on why homeland security is too serious to be left solely in the hands of government.

Cyber

Behavior

Capacity paradox

Laws of value conservation

Business is not physics –and certainly not a closed, frictionless system anyway. Value is (probably almost) never conserved during industry disruption.

Inversions, Paradoxes and Crossovers

Citizen science goes 'extreme', Nature News 17 Feb 2012

One solution may be to develop a “science of citizen science”, suggests Francois Taddei, a molecular geneticist at the Centre for Research and Interdisciplinarity at Paris Descartes University. As the field develops, he says, researchers will need to figure out the optimal division of labour between citizens and professionals. The question is evolving, he says, owing to a slowly developing understanding that expert knowledge does not necessarily trump collective intelligence.

Inversions

My theme of the week on this blog is “inversions”. For our purposes, an inversion is the result of turning an aphorism, an expression or any example of “conventional wisdom” on its head. We tend to accept such oft-heard expressions as laws of nature. They have become so ingrained in our thinking that we forget that they are most often rules of thumb that work in a very specific, now forgotten, context. Reconsidering accepted wisdom is always a useful exercise, and what better way to reconsider it than to invert it? Sometimes it is simply fun inverting an expression, just to see what you get. My first example, “Decision-based evidence making”, is a case in point. Inversions appear at first unfamiliar and counter-intuitive, but often offer insights into the complexity of the world around us.

Decision-based evidence making

We have all heard about evidence-based X, where X stands for your favorite field, for example, evidence-based medicine, evidence-based policymaking, and more generally, evidence-based decisions. The slightly oxymoronic feeling these expressions may give you is not an accident. There is a very natural set of human biases that make us look for confirmatory evidence (confirmation bias). Once we have a theory in mind, it is difficult to move away from it (anchoring bias).

The prepared mind favors chance

This is evidently a play on the famous quote from Pasteur, “chance favors the prepared mind”. In fact, if it were translated literally from the French, the quote would read: “chance favors ONLY the prepared mind”, which I wish were true –but I know a lot of clueless idiots who have been favored by chance. Having said that, even if chance favored only the prepared mind, Pasteur’s aphorism is not helpful. Assuming I know how to prepare my mind (back to that in a minute), should I just sit around and wait for luck to hit? The inverted aphorism, “The prepared mind favors chance”, is much more useful and constructive –if you know how to use it. Although I wish I could lay claim to this wonderful inversion, it is my friend Michael Schrage who came up with it (http://www.leighbureau.com/speakers/MSchrage/essays/preparedminds.pdf). He rightly and modestly calls it “an upgrade” to Pasteur’s. While Michael focuses on how companies can look for insights in Big Data, I would like

Why optimization is sub-optimal

In manufacturing,

You Should Only Manage What You Can’t Measure

Peter Drucker was right when he wrote: "What gets measured gets managed." But why is this so often taken to a false corollary like "What can't be measured isn't worth managing"? [Prusak] The truth is, what’s measurable gets measured (and managed) but it is not necessarily what matters.

Judgment, Barry Schwartz: Practical Wisdom.

For measurable tasks that cannot be automated, people will manage themselves just fine to what gets measured. To a fault. Incentives are tied to measurements, people will optimize.

Intelligent design by means of evolution

Evolutionary algorithms are a fantastic success story of how a natural phenomenon (well, not everyone agrees that it is a real natural phenomenon but I am a believer) can be exploited for engineering purposes. One approach I find particularly promising was introduced by Richard Dawkins 20 years ago: interactive evolution, which is the* in silico* version of directed evolution.

Computational techniques known as artificial evolution or evolutionary computation replicate in silico the way that biological evolution works. Since its introduction thirty years ago, evolutionary computation has proven highly successful at solving a wide range of decision problems.

Now imagine the same process with the biological engine responsible for variation being replaced by a computing engine. The result is called interactive evolution or IE. IE is very useful when the space of potential solutions, designs or strategic options is large AND the goodness of a solution is difficult to formalize. For example, there are many situations where the decision maker doesn’t know ahead of time what the solution looks like – “I know it when I see it” kinds of situations. Starting with a more or less randomly generated population of solutions, the evolutionary technique will search the space of solutions by picking the fittest individuals as defined by the user, will mutate them and breed them, and the offspring will again be evaluated by the user, etc, until solutions emerge that satisfy the user. There have been a number of business applications over the last few years and the trend is accelerating.

Honda has used it to help its designers explore the space of car designs using interactive evolution. The problem with car design is that it is highly constrained: a designer has to satisfy hundreds of technological constraints simultaneously (such as wheelbase length, windshield angle, and size of engine compartment) while at the same time remaining creative. In other words, automobile designers must balance aesthetic considerations with technical specifications, an often frustrating juggling act resulting in a lengthy trial-and-error design process. The tool enables the designer to engage in a guided exploration of the design space: it is first presented with a number of initially random designs; the designer picks the ones that come the closest to what he is looking for –they are the fittest individuals; artificial evolution takes the fittest designs, mutates them and breeds them to create a new “virtual generation” of designs, which the designer evaluates again. The results are spectacular: in just a few iterations, a car designer can create any design he wants consistent with the constraints. Designers can create and compare a vast number of designs in a short time, greatly streamlining and accelerating the design process. Icosystem has applied IE to aircraft design, control system design, intrusion detection on computer networks, postal route optimization, drug discovery, exploratory data mining and model calibration. Companies like Procter and Gamble or Pepsi-Cola North America have harnessed the power of evolution combined with the collective brainpower of their customers with the help of Cambridge, MA-based Affinova to come up with new packaging or product designs. It works very much in the same way as the Honda tool but it is consumers who pick the designs they like as opposed to professional designers. Companies are able to directly translate their customers’ potentially non-verbal tastes and preferences into new products or designs that will please specific segments.

Why do I think IE is such a promising approach? Because it empowers us to explore and invent. IE helps you navigate the design space. You can never be sure you're exploring the space in an exhaustive manner. But you're navigating it in a way that is a lot smarter than a random walk and a lot more empowering than being forced into a solution.

Increasing capacity decreases throughput

Daniel Dennett: competence without comprehension

Daniel Dennett: the virus

Law of Small Numbers

Two Empowerment Projects

You sometimes come across people whose passion and purpose are contagious, people you want to be infected from. Two such people I know have embarked on two very different but inspiring journeys. I want to share a little bit of what I know about them and their empowerment projects, in the hope that their passion and purpose will infect you too.

“Swarm intelligence is the key to homeland security”, David Stephenson (http://stephensonstrategies.com/w-david-stephenson-bio/) told me six years ago as I was meeting him for the first time. What a crazy thing to say, I thought at the time –despite being the co-author of dozens of articles on swarm intelligence (http://openp2p.com/pub/a/p2p/2003/02/21/bonabeau.html), the collective intelligence that emerges out of the myriad interactions among large numbers of “things”. But what was crazy then quickly became a widespread meme. In February 2007, David and I co-authored an article (http://www.hsaj.org/?article=3.1.3) on why homeland security is too serious to be left solely in the hands of government –and how government can empower citizens for the benefit of all, leveraging the collective intelligence of its citizens (an excellent recent article in the same publication describes more recent developments, here http://www.hsaj.org/?fullarticle=7.1.3).

In the last three or four years, David has dedicated his life to a related empowerment project: data liberation (http://www.indiegogo.com/Data-Dynamite?c=home). Beyond the obvious transparency benefits of making data available to anyone, there are advantages to having “many eyeballs” involved in looking at the data. In addition to federal government sites such as recovery.gov (www.recovery.org) and transparency.gov (www.transparency.gov), a number of local governments have been creating websites offering datasets to users. According to the site DataSF.org (www.datasf.org), “DataSF is a clearinghouse of datasets available from the City & County of San Francisco. Our goal in releasing this site is: (1) improve access to data; (2) help our community create innovative apps; (3) understand what datasets you'd like to see ; (4) get feedback on the quality of our datasets.” If you haven’t already, it’s worth browsing the apps that have been created based on that data (here: http://datasf.org/showcase/). David’s book, entitled “Data Dynamite: Liberate Data to Transform our World” is a powerful manifesto for why we need to continue to make all sorts of data more available, not less.

Another empowerment project I discovered because I live in Santa Fe. I met Jamila Bargach at my neighborhood elementary school’s potluck in September 2010, as she was starting a Scholarship at the School for Advanced Research (http://sarweb.org/?resident_scholar_jamila_bargach) –one of Santa Fe’s many interesting institutions with uninformative names. Jamila is now back in her native Morocco, working to implement her idea: harvesting the clouds. From the website of the Foundation (http://www.darsihmad.ma/2/index_4.asp?n1=109&n2=380&n3=155) that she and her husband created, you can read:

“The region of Aït Baamrane is arid to semi-arid. Lack of water is a problem that profoundly affects the quotidian of the villages located in these Anti-Atlas mountains, and more particularly the lives of women and young women who are entrusted with the water-chore. They spend a yearly average of 3.5hours/day for fetching water, but this time constitutes a waste for opportunity, as they qualify it themselves, they could otherwise employ in more valorizing activities had they easier access to water. But interestingly, this region is particularly humid given the meteorological conditions which create thick and long lasting fog from December through June.”

Using fog collection technology (and the experience of FogQuest (www.fogquest.org))), Jamila is working on empowering women by changing the old model of water distribution and replacing it with a more rewarding one. Listen to Jamila talk about her project: Harvesting the Clouds: Fog Collection Technology and Gender Equality in a Berber Village, Morocco (http://www.darsihmad.ma/2/index_3.asp?n1=103&n2=404).

On August 29th, 2000, I had three babies. One girl named Capucine, one boy named Hippolyte and one company named Icosystem (well, technically, I hired the company’s CEO on August 30th, but you get the idea). Given that together the three have consumed all of my time and energy over the last ten and a half years, I do not feel qualified to write about anything else. And although I would love to write about my delightful parenting experience (and how amazing my kids are), this blog’s readers will probably be more interested in the things I have discovered on my journey with Icosystem. Luckily, Icosystem’s interests cover a lot of different domains, albeit through a common lens: how to make decisions in a complex world.

The ominous panda generator (http://solaas.com.ar/works/hipanda/panda2.htm) and the addictive baby naming site Nymbler (www.nymbler.com) are two examples of what we, at Icosystem, call the “hunch engine”. When searching for a baby name (or an ominous-looking panda), you don’t really know what you are looking for. Hopefully you’ll know it when you see it. But random-walking through the vast universe of baby names (and panda faces) can become rapidly boring, and unlikely to produce a real breakthrough. Have you noticed, for example, that most baby-naming books list names in alphabetical order? By the time you have reached the letter C, you no longer want to have a baby. Instead, Nymbler follows your hunches: I kind of like Ivy and Lily, but not enough to name my baby girl Ivy or Lily. On the other hand I do hate Amanda –reminds me of a pest back when I was in second grade. These are feelings and hunches. How can we leverage them?

The hunch engine uses a computational technique known as “interactive genetic algorithms” (the idea for which was first mentioned in Richard Dawkins’ The Blind Watchmaker (http://en.wikipedia.org/wiki/The_Blind_Watchmaker)). Genetic algorithms mimic the process of evolution by mutating and recombining the best members of a generation (Ivy and Lily) and making sure the worst has not offspring (sorry Amanda!). With interactive genetic algorithms, the user tells the machine what’s good and what’s not. For example, by picking Ivy and Lily, you are implicitly driving the hunch engine toward flower names –even though you may not have realized it before. And within a few clicks, you will have found Rosemary, the name you really wanted without knowing it. The secret, however, is that the hunch engine is constantly adding “noise” to your choices, trying to expand your horizon while at the same time offering variations on your selections. The whole becomes a source of serendipitous encounters.

Situations where we don’t know what we are looking for abound. Beyond baby names, you may be looking for a name for your pet, your company, your product or your website. You just want to escape for a few days but don’t know where, and you don’t know what might be available at a reasonable price. Or you’re in the mood to go out tonight and would like to explore options. Or you’d like to design your own wallpaper but don’t know how to explore all possibilities (http://www.icosystem.com/labsdemos/evodesign/). More generally, product configurators (for cellphones, computers or jeans) are usually limited or daunting. Or you’re looking for a tie to go with your new shirt: here again, have you ever tried shopping for ties online? Spend ten minutes and you’re left feeling dizzy. Yes, we even have a hunch engine for shopping for ties!

These are all examples where you would like to explore the space of the possible but it’s a really, really big space. The solution consists of outsourcing your left brain to the algorithm and maintaining right-brain control over what’s interesting –and what’s not. This division of labor between human and machine combines what machines are best at (sifting through lots and lots of stuff) with what humans are best at (finding patterns and using our experience and feelings). The hunch engine covers a continuum of situations, ranging from search (for example, finding a baby name) to design (for example, create your own wallpaper or a name for your company). As such, it is a form of intelligent design by means of evolution.

Competence Without Comprehension

In a previous post (http://www.theatlantic.com/technology/archive/2011/03/the-hunch-machine-how-to-make-better-choices/72797/) I described the hunch engine (http://www.wired.com/science/discoveries/news/2006/03/70388), an exploration-support tool based on interactive genetic algorithms.

There is an intriguing parallel I want to expose in more detail between biological evolution and decision making: search and evaluation in decision making are similar to variation and selection in evolutionary theory. Search is all about creating a variety of options and possible answers to a query; evaluation is the process through which some or none of the options are selected. Nature thus provides us with a powerful metaphor for decision making, and in that context genetic algorithms are decision-support tools. With interactive genetic algorithms, variation is performed by a non-human device while options generated by the device are evaluated by a human being.

http://www.technologyreview.com/video/?vid=103

In fact, we humans have been using this technique for hundreds of years, it is known under various names such as breeding, animal husbandry, or directed evolution. To name one famous example, corn was bred about 9000 years ago by Mexican farmers. Teosinte, the plant they started with, is so different from modern corn, that it was originally classified in a different genus. Teosinte is barely edible, while corn is today one of the leading sources of calories for humans. The story of how such a transformation was made possible, by the combination of careful selection by farmers with a genetic structure that enabled dramatic morphological changes, is still being uncovered by ongoing research. Which means that humans have been using a powerful biological engine called variation which they did not understand at all; all they knew was that it worked for producing the requisite amount of variation and they could provide selective pressure.

The philosopher Daniel Dennett uses the phrase “competence without comprehension” (http://www.pnas.org/content/106/suppl.1/10061.full) to describe the strange value proposition of Darwinian evolution, that “to make a perfect and beautiful machine, it is not requisite to know how to make it’’ [MacKenzie RB (1868) (Nisbet & Co., London), cited by Dennett]. Indeed, what McKenzie, a 19th century critic of Darwin, calls “a strange inversion of reasoning”, has been one of the weapons creationists have used, as in the pamphlet from the 1970s.

Creationist pamphlet from the 1970s (from Dennett).

But directed evolution is a highly successful embodiment of that inversion of reasoning –of competence without comprehension. Corn is the descendant through directed evolution of teosinte. The domestic dog, in its apparently infinite variety, is the product of many generations of breeding from just one common ancestor, the gray wolf. Examples abound.

In silico evolution, in the form of genetic algorithms, creates an opportunity for competence without [necessary] comprehension. You may be able to comprehend –either during or after the design process, but you don’t have to. The machine takes care of the variation process. This is a powerful concept: think about all the situations in which you have to rely on an expert to produce variations for you –an architect, a designer, a brand naming consultant, etc. The expert is the gatekeeper between you and your dreams, and defines the possible on the basis of her own biases. Your dreams are bounded by the expert. Yet, you are an expert on knowing whether you like something or not. You may not understand how the expert comes up with the variations, but you’re competent (in fact, you’re likely the most competent) when it comes to your own tastes. Competence without comprehension empowers you to innovate far beyond your comprehension. One caveat is obviously that whatever new stuff you produce be safe.

Consider the following situation: you took a picture with your camera-phone, but it’s too dark. You would like to make your photograph prettier, but you have no comprehension of digital photography. There is often a “magic button” you can use with digital photography software –yet you know that the outcome is arbitrary. One thing you are very competent about is deciding whether a picture is good or not. How do you exert your competence without comprehension of the underlying technology? The hunch engine offers you multiple choices, each of which is an image obtained from applying a random filter to your original picture. Following your competent hunches, you select one or more of the pictures that seem to offer some degree of improvement over the original. And you press “evolve”. The next iteration presents you with another set of choices, this time variants (mutated and recombined filters) of the filter(s) you selected in the previous iteration. Again, select the images that offer the most improvement. After two or three steps of this directed evolution, you end up with just the right picture –and you know it. At no time did you have to understand what was going on under the hood.

Beyond the breeding of natural organisms, there is evidence that the same processes were at play in aspects of technological evolution.

For example, Deborah Rogers and Paul Ehrlich of Stanford University have studied the evolution of Polynesian canoes (http://www.pnas.org/content/105/9/3416.full), and reached that conclusion –although in this case the mechanisms for producing variations are understood while the fitness of . Interestingly, they also cite a 1908 proposal by French philosopher Alain that boat design is subject to natural selection:

Every boat is copied from another boat. . . Let’s reason as follows in the manner of Darwin. It is clear that a very badly made boat will end up at the bottom after one or two voyages, and thus never be copied. . . One could then say, with complete rigor, that it is the sea herself who fashions the boats, choosing those which function and destroying the others.

Cyber-Security: Behavior Matters

In an earlier post, our beloved Jim Fallows wrote briefly about a DoD-funded cyber-security initiative named SENDS (http://www.theatlantic.com/technology/archive/2010/11/cyber-security-china-and-sends/67191/), for Science-Enhanced Networked Domains and Secure Social Spaces. The overall objective of SENDS is to promote and begin to demonstrate the concept of a science of cyberspace –with an initial focus on security. The vision for SENDS (http://sendsonline.org/), developed by Carl Hunt, Richard Raines and Craig Harm, is one that embraces the richness, diversity and messiness of cyberspace. Central to their vision is the idea that the social, economic and behavioral aspects of cyberspace, which are largely missing from the general discourse on cyber-security and are certainly under-funded and under-represented in government-sponsored programs, are at the core of what makes cyberspace the complex adaptive system that it is. An inclusive, multi-disciplinary, holistic approach that combines the technical and the behavioral is needed.

Being a founding member of the SENDS initiative, I am definitely partial to its vision. The extent to which research and development in cyber-security has been skewed toward “technical solutions” is mind-boggling. As an illustration, it seems surreal that in an otherwise excellent document, the authors of a 2009 manifesto from Sandia National Laboratories entitled “Complexity Science Challenges in Cybersecurity” (https://wiki.cac.washington.edu/download/attachments/7478403/Complexity+Science+Challenges+in+Cybersecurity.pdf?version=1&modificationDate=1238771914559) have not dedicated a single line to human behavior. For example, their main M&S thrust is entitled: “Modeling the behavior of programs, machines, and networks”. No humans necessary –although I concur with the authors that there is a need for a new “cyber-calculus” –just the ability to frame concepts and issues in modern mathematical terms would be of enormous help. Or in a recent report (http://www.fas.org/irp/agency/dod/jason/cyber.pdf) by a group of DoD-funded physicists, you can read:

On the positive side, the cyber-universe can be thought of as reduced to the 0s and 1s of binary data. Actions in this universe consist of sequences of changes to binary data, interleaved in time, and having some sort of locations in space. One can speculate as to why mathematics is so effective in explaining physics, but the cyber-world is inherently mathematical.

But cyberspace, although it is the result of tremendous technological progress, is not just a piece of technology: it is both an enabler and an amplifier of human nature, eliciting new manifestations of human nature. It feeds (and in many ways feeds on) one of the most fundamental needs of human beings: communication. That it has become such an integral part of our lives in such a short time shows how deeply it resonates with our need to communicate and be connected. It should come as no surprise, therefore, that the multifaceted dynamics of cyberspace be so strongly influenced, even defined, by the behavior of its participants. According to Mark Graff of Lawrence Livermore National Laboratory: [cyberspace] gives individuals and small groups unprecedented reach to affect others; it makes physical distance much less of an insulating factor; confuses us about what is permanent, or public, or safe; and largely operates insensibly to us. We feel safer if important data is near us, or some place we know, or with someone we’ve met, but these comfort factors make no “Internet” sense and don’t scale to Internet dimensions either. In matters of risk assessment, we feel pretty safe from attacks originating “far away”; we also tend to ignore “low and slow”— or sporadic—attacks; random, “pointless” attacks (like from Internet worms) mostly tend to be low on our worry list, too. No wonder that the intuition we have gained from the physical world over thousands of years of evolution leaves us ill prepared to deal with the new geography of cyberspace. We can’t hope to acquire this new kind of intuition overnight. The bad news is that we suffer from severe limitations in our understanding of a critical component of our lives. The good news is that we are all subject to the same limitations –good news only if we can regain a competitive advantage in what has been a level playing field. Understanding our own behavior and that of our enemies becomes the most viable defense and the most potent weapon we can develop.

Obviously it is essential to continue to improve the technical aspects of cyber-security and significant investments need to be made to ensure continuous progress –and to keep up with increasingly sophisticated enemies. But at the same time, human behavior is almost always the weakest link in security. The attacks on Google and other companies in China in 2009 were initiated through phishing –the underlying technical exploit is often trivial but social engineering is always the entry strategy. In the September/October 2010 issue of Foreign Affairs, Deputy Defense Secretary Lynn described the spread of a malicious worm on both classified and unclassified US Central Command systems (http://www.foreignaffairs.com/articles/66552/william-j-lynn-iii/defending-a-new-domain), which started with the insertion of an infected USB key into a US military laptop. Apparently it took the Pentagon 14 months to clean things up (http://www.wired.com/dangerroom/tag/operation-buckshot-yankee/). The worm would never have been able to infect any network without the help of someone –malicious insider or clueless insider. On the flip side, the recent Stuxnet worm that damaged the Iranian uranium enrichment infrastructure (http://en.wikipedia.org/wiki/Stuxnet), seems to have used the same entry strategy of USB key insertion to get started; once in a system, it would use multiple exploits to spread itself. Example after example of intrusions and attacks point to the fact that human behavior is the enabling factor. In the case of the leaks of diplomatic cables to Wikileaks by Private Manning, human behavior is at the core: no technology solution would on its own prevent it.

A small but growing community of scientists from academia and industry has emerged in the last few years (http://www.cl.cam.ac.uk/~rja14/shb09/schedule.html). They need encouragement and support.

On the Adoption of New Drugs: The Myth

In 1954, pharmaceutical giant Pfizer was interested in determining how physicians decide to adopt a new drug so that it could more effectively market its products through detailing and traditional media. By knowing how physicians acquire reliable information and who they trust, Pfizer could market its new drugs more effectively, optimizing the allocation of marketing resources among detailing, media advertisement, continuing medical education, etc. They funded a landmark social network study aimed at showing the effect of interpersonal influences on behavior change in relation to the adoption of Tetracycline, a powerful and useful antibiotic just introduced in the mid-1950s. Pfizer hoped tetracycline would diffuse rapidly because it was a tremendous improvement over existing antibiotics. The Pfizer-funded study contained two major advances over previous studies in that it relied on a behavioral measure of time of adoption by looking at prescription records and used network analysis to identify opinion leaders. However, numerous subsequent studies of this work revealed a number of weaknesses in the collection and analysis of the data and the study is inconclusive: the uptake in tetracycline adoption cannot be assigned with confidence to social network effects.

But the myth has persisted and the notion that drug adoption spreads over social networks through opinion leaders is driving a large amount of marketing spend. Social network scientists keep the myth alive. In a recent Marketing Science article (http://jhfowler.ucsd.edu/contagion_in_prescribing_behavior.pdf), two famous network scientists, Connected (http://connectedthebook.com/) authors Nicholas Christakis and James Fowler state that “Iyengar, Van den Bulte, and Valente, in their careful and insightful study, find that even after controlling for marketing effort and arbitrary system-wide changes, there is evidence for contagion in the prescribing patterns of doctors.” But they also cite a 2001 article by Van den Bulte and Lilien (Medical Innovation revisited: Social contagion versus marketing effort. Amer. J. Sociol.*106* 1409-1435) that clearly showed that contagion effects in the original Tetracycline study were not present –or at the very least significantly weaker than concluded. The more recent study by Iyengar and colleagues (sbm.temple.edu/dept/marketingscm/documents/Iyengar_2010.doc) on a completely different drug suggests some diffusion effects, but they are subtle and the concept of opinion leaders is murkier than one might think.

It is time for the myth to be debunked: the Tetracycline study is foundational is its methodology but it does NOT support contagion.

On the Adoption of New Drugs: One Reality

In a previous post, I briefly described the drug adoption contagion myth: the landmark Tetracycline study does not offer evidence that drug adoption spreads through social networks. But it is a widely-held belief and pharmaceutical companies still allocate a significant amount of their marketing resources based on this premise. Our experience is that the contagion effect, if it exists, tends to be small. Furthermore, opinion leaders (determined either through unreliable self-reporting or somewhat more reliable sociometric measures) have little impact on the speed of adoption –in other words, there are no real influencers. To be fair, however, our experience also suggests that each drug follows its own adoption path, which depends on a lot of factors such as the disease itself, the market at the time of introduction, and how it is prescribed and used.

For example, our team at Icosystem studied the adoption of a new drug used exclusively in intensive care units (ICUs), where the number of individuals (doctors, nurses, pharmacologists) involved in the decision to prescribe the drug is between 10 and 20 in a given hospital. The study revealed that the temporal structure of the social network is the key to prescription behavior. While a snapshot of the social network is unhelpful (we found no correlation between the topology of the network and probabilities of adoption), its dynamics over time is a great predictor of the speed with which the drug is adopted: this can be explained by the fact that in many hospital ICUs, physicians work only a few months per year and teach or work in other departments for the rest of the year, so that the only opportunities physicians have to interact is when their assignments overlap for a few days. We discovered that the degree of overlap correlates positively with the speed of adoption, suggesting that ICUs that are organized to provide more overlap between physicians are more favorable marketing targets. More broadly, some measure of the total exposure of a physician to the drug and its effects is a great predictor of adoption speed. Promoting the drug to ICUs with maximal overall exposure first accelerates adoption. ICUs that are more difficult to penetrate can be targeted in a second marketing wave as it is easier to market a product when you can tell your customers that their competitors or peers are already using it.

We must not lose sight of data in the excitement of playing with cool theories in simulated worlds. It is ok for the theory to be ahead of the data, but not by light-years. A case in point is the over-theorization of social networks in the last few years.

http://utminers.utep.edu/asinghal/Reports/EMR-Singhal-Quinlan-june19-07-DOI-Word-file-Stack-Salwen%5B1%5D.pdf

In my experience good business-related social network data (whatever that means) is a rarity. The data is often inadequate, ranges from incomplete to sparse, is noisy, sensitive to minute details and lacks such important characteristics as frequency, quality and nature of the interactions. In other words it is unusable in practice for predictive purposes.

One of the issues facing those who want to study the influence of social networks on the diffusion and adoption of innovations to design marketing interventions is the lack of reliable data. There are situations, however, where the community of adopters is sufficiently small that it can be mapped accurately.

Bad Decisions

Complexity and uncertainty impedes our ability to make sound decisions. You probably knew that, but the extent to which it is true is only beginning emerge. Consider the following examples.

Judges.

Discrimination and chaos.

401(k)

Hip replacement

“This is from a paper by Redelmeier and Schaefer. And they said, "Well, this effect also happens to experts, people who are well paid, experts in their decisions, do it a lot." And they basically took a group of physicians. And they presented to them a case study of a patient. Here is a patient. He is a 67 year old farmer. He's been suffering from a right hip pain for a while. And then they said to the physician, "You decided a few weeks ago that nothing is working for this patient. All these medications. Nothing seems to be working. So you refer the patient to hip replacement therapy. Hip replacement. Okay?" So the patient is on a path to have his hip replaced. And then they said to half the physicians, they said, "Yesterday you reviewed the patient's case and you realized that you forgot to try one medication. You did not try ibuprofen. What do you do? Do you pull the patient back and try ibuprofen? Or do you let them go and have hip replacement?" Well the good news is that most physicians in this case decided to pull the patient and try the ibuprofen. Very good for the physicians.

The other group of the physicians, they said, "Yesterday when you reviewed the case you discovered there were two medications you didn't try out yet, ibuprofen and piroxicam." And they said, "You have two medications you didn't try out yet. What do you do? You let them go. Or you pull them back. And if you pull them back do you try ibuprofen or piroxicam? Which one?" Now think of it. This decision makes it as easy to let the patient continue with hip replacement. But pulling them back, all of the sudden becomes more complex. There is one more decision. What happens now? Majority of the physicians now choose to let the patient go to hip replacement. I hope this worries you, by the way -- (Laughter) when you go to see your physician. The thing is is that no physician would ever say, "Piroxicam, ibuprofen, hip replacement. Let's go for hip replacement." But the moment you set this as the default it has a huge power over whatever people end up doing.”

Leaked US Diplomatic Cables a “credible source” of information in the Middle East

As I was exploring the timeline of events in the Middle East from recent months (a great visualization can be found here http://www.guardian.co.uk/world/interactive/2011/mar/22/middle-east-protest-interactive-timeline), it became clear to me that the publication of classified diplomatic cables by Wikileaks, rather than diminish the credibility of US diplomacy, has actually shown how credible its analyses are –at least to the populations of the Middle East. In an article last week (http://www.businessweek.com/magazine/content/11_12/b4220007540210.htm), Romesh Ratnesar of the New America Foundation (http://newamerica.net) describes how the extreme nepotism found in many of the toppled or endangered regimes was detailed in some of the cables. The cables thus provided validation, and perhaps even justification, in people’s revolts against these regimes. Writes Ratnesar:

“On June 23, 2008, a cable arrived at the U.S. State Dept. from the American ambassador to Tunisia, Robert F. Godec. Its subject was corruption, cronyism, and graft in the North African nation, as practiced by relatives of President Zine al-Abidine Ben Ali. "Whether it's cash, services, land, property, or, yes, even your yacht, President Ben Ali's family is rumored to covet it and reportedly gets what it wants," Godec wrote.”

and

“Godec's cable was supposed to remain classified until 2018, but last fall it surfaced in the cache of State Dept. documents made public by WikiLeaks. To Tunisians, the revelations of nepotism were hardly shocking, but never before had they been so publicly detailed by a credible source.”

The same credible source of information also played a role in Yemen, Egypt, and more. The situation in all these countries was arguable volatile (in Egypt in particular), but in an ironic twist, US Diplomacy’s reliable information may have been an unlikely catalyst to the chain reaction we have witnessed.

Has Complexity Peaked?

It was just days ago that I discovered Google’s Ngram viewer (http://grams.googlelabs.com) and started playing with it. For those unfamiliar with it, the Ngram viewer searches a huge database of books for occurrences of expressions that can contain multiple words, and plots the fraction of books in which the expression occurred in a given year. So I tried “Bonabeau”, one of the most important terms to me. No luck. I should instead keep Googling myself. I then tested “complexity science” and “complex adaptive systems”, which produced the fascinating diagram below: both terms take off in the early 1980’s, although complex adaptive systems have been around for longer, probably driven by the cybernetics movement. I suspect that the Santa Fe Institute (www.santafe.edu), founded in 1984, is the catalyst for this exponential increase in attention. But somewhere around 2005-2006, the trends is reversed for both terms.

Intrigued, I wanted to test the assumption that this might be due to fluctuating scientific fashion, since the two expressions embody a scientific community. But here is the diagram for “complexity” itself –the concept rather than the science, exhibits the same decay, around 1999. Note that the expression “complexity” has been used for much longer.

Now that’s interesting! It’s hard to argue that the world has become, and continues to become, a lot more complex over the last few decades, so why are mentions of complexity decreasing? First, it is important to remember that the diagram does not show absolute numbers but proportions: mentions of complexity may still be rising, but the relative popularity of the expression is decreasing. The number of books published per year has undoubtedly risen at accelerated pace. But the diagrams do suggest that the mindshare of complexity is diminishing. Are we repressing the complexity we’re facing? Are we craving simplicity? Have we had enough with complexity?

Complexity Is Us

An alternative, or perhaps complementary, explanation to the observation that mentions of complexity have been decreasing in recent years according to Google’s Ngram viewer, is that complexity concepts have become pervasive.

Complexity concepts have spread so successfully that they have even gone mainstream. For proof, look no further than George the cabdriver. One day during a taxi ride across Boston, I was struggling to get some work done while politely lending half an ear to my chatty driver, George. “They’re adding a lane to the highway,” George said, “but it’s not going to do any good. “Why not?” I asked. “Because traffic is an emergent phenomenon,” replied George, and suddenly he had my full attention. “Some very counterintuitive things can happen. You add a lane to a highway and actually traffic increases.” A Haitian immigrant whose formal education had ended with dropping out of high school, George went on to instruct me about how traffic patterns emerge from the interactions between drivers, and how the control and prediction of traffic requires a way of thinking that people aren’t naturally good at. For those who might dismiss George as a dubious N of 1, consider a few more examples.

The butterfly effect and “six degrees of separation” have become common parlance to the point of cliché. Both terms have been appropriated as titles of big-screen films led by A-list movie stars. Back in the mid 1990’s the Hollywood trivia game Six Degrees of Kevin Bacon became one of the first major internet memes.

Agent-based modeling and data mining show up in TV crime dramas. My own book, Swarm Intelligence, provided the inspiration for Michael Crichton’s bestselling nanotechnology-run-amok thriller, Prey.

The fabled “edge of chaos” is popularly (if simplistically) understood to be a wellspring of creative novelty. Bifurcations and phase transitions are well known among the laity nowadays in the form of tipping points, which came to popular awareness through an internationally bestselling book. A key concept in Al Gore’s global blockbuster documentary An Inconvenient Truth is the idea that the climate is a system in dynamic equilibrium which may abruptly reconfigure itself if perturbed too far.

The idea of collective intelligence and the “wisdom of crowds” – the title of another recently influential book – has become a basic tenet among the internet set. It can be seen everywhere, from the Wikipedia project to U.S. ruling class’s (recently shaken) quasi-spiritual respect for the judgments and choices of The Market.

And then of course there is social networking and peer-to-peer (P2P) file sharing, both of which achieved escape velocity into business-model space thanks to SFI impetus. Over the past few years web sites touting social network service have been springing up like mushrooms, and the swarm-based, self-organizing networks of P2P content-delivery systems now account for at least two-thirds of all internet traffic.

When at the Santa Fe Institute some 15 years ago, I remember this young postdoc, Duncan Watts who told me he was working on social networks. I told him this would never succeed as a business model because networks need a critical mass of people to achieve value, and that point is hard to achieve. But more than half a billion people are on Facebook –that’s critical mass!

And the list goes on. Strange attractors. Set points. “More is different.” Terms like these are widely bandied, if sparsely grokked, but the fact that they amount to little more than buzz phrases in most lay usage and business speak is beside the point. The amazing thing is that such arcane terms have any popular currency at all.

A lot of these ideas have become so familiar that people don’t even think about them as ideas anymore.

The future of complexity science is already here. There are still walls that need to be taken down, but there is already a new generation of scientists in place for whom thinking across disciplines is just part of the daily routine of being a scientist. That is the tremendous achievement of just 20 years of complexity science.

As to the practical applications of complexity thinking, what I call complexity engineering, we can already get a glimpse of the future, but in many subtle and sometimes hidden forms. For example, we are already seeing a broad familiarity with some of the fundamental concepts of complexity science in the general public in the form of social networking, peer-to-peer systems, wikipedia, web 2.0, and other decentralized systems. Web 2.0 in particular is a great playground for complexity engineers who recognize that their toolkit can help them make sense of the emergent phenomena that take place when millions of individuals start interacting. Google is probably the most successful “complexity company” on earth, although most people and financial commentators would probably drop “complexity”: Google understood early how to exploit the emergent knowledge base embedded in the interactions between web pages –a form of information social network of web pages created by users linking their own pages to others’. When I look at the techniques and algorithms used by Google’s search engine, they remind me of social network techniques developed in the 50’s and 60’s and transferred to other disciplines by complexity scientists in the 90’s before Google found the “killer app” for them. I am quite certain that other complexity killer apps will continue to emerge along the same lines.

Time, Space and Interactions: The Case for Agent-Based Modeling (ABM)

When a customer makes a purchase or switch decision, it is often the result of a history. Impulse decisions, while they do happen, are the exception rather than the rule. That does not mean that most decisions are rational, simply that they cannot be explained by just looking at the time they happen. When a wireless phone customer decides to switch carriers, such a decision is the result of all the interactions and experiences this customer had with his carrier as well as with other sources of information. Failing to recognize the temporal dimension of decision making can lead to dramatic prediction errors. ABM, and only ABM, can explicitly deal with all aspects of time: learning, waiting, simmering, habituation, forgetting, interacting with other customers, etc.

For example, in the casino industry, common wisdom holds that customers have a fixed budget and stop playing when their budget is exhausted. An ABM fed with real slot data from a loyalty card program showed that in reality customers stop playing when their total experience over time (TEOT) (a combination of the dynamics of their wins and losses weighted by demographic attributes and day of the week, and, yes, budget) reaches a threshold. TEOT is a much better predictor than budget or any combination of demographic attributes, which enables a major casino owner and operator to implement effective real-time marketing and promotional offers. Of course the dirty little secret is data and how to use it effectively to estimate complex time-dependent models of decision-making. When the data exists, in the absence of a coherent theoretical framework, not to mention theorems, one has to perform rigorous computational experiments based on statistical machine learning techniques.

Another example is health insurance, where a customer’s demographic attributes are not sufficient to predict which plan he or she will select. Instead, the characteristics of each plan are viewed through a looking glass that puts more weight on certain characteristics as a function of the customer’s experience with his current plan, which is a combination of his and his family’s health in the past year and satisfaction or dissatisfaction with the health care afforded by the plan. Furthermore, if specific adverse health events have happened in the recent past, they strongly affect the way the possibility of catastrophic losses is perceived. By using an ABM that explicitly deals with the effects of experience and recency, prediction error could be reduced by an order of magnitude at Humana, a leading US health insurer. No amount of traditional econometric modeling with demographic attributes as explanatory variables would have been able to achieve this level of accuracy.

In retail, the layout of a supermarket is known to be a key sales driver, yet shopper behavior is an emergent property with a strong spatio-temporal component that is never taken into account in traditional econometric modeling: while the trajectory of a shopper in a supermarket is influenced by the shopper’s shopping list, the trajectory in turn influences what the shopper buys beyond the shopping list. Through the use of a spatial behavioral model of shoppers in a supermarket, Pepsi was able to predict hot spots in any supermarket as a function of the supermarket’s layout and the demographic attributes of its shopper population. With the knowledge of hot spot locations, Pepsi could determine the best location not only for their products but also for promotional signs. Here again, the dirty little secret is data and how to use it. Not only did we have to develop special estimation techniques to infer trajectories, data collection itself was a challenge: shoppers were given “smart carts” equipped with tags for path tracking.

Customers experience, learn, adapt, adjust. Their decisions are path dependent: in other words, decisions are dependent upon a contingent history. Existing statistical or econometric techniques do not deal satisfactorily with path dependence. When done properly (and that’s a big IF) ABM combines the statistical rigor of existing techniques with the ability to accurately model the temporal components of decision-making. As a result, not only does predictability go through the roof, the outputs of what-if scenarios also become more reliable because behavioral models are fundamentally causal rather than correlational. Knowing that two variables are correlated is good enough to predict the past, but robustly predicting the future requires understanding the underlying causal mechanisms of decision making.

Evolved vs Designed (or Engineered)

The distinction is getting blurred:

Understandable vs opaque

• We design more and more complex systems, but that does not mean that we know everything about them. Examples: control systems, software.

• Paradox of human artifacts in the 21st century: they are products of our brains but our brains cannot begin to comprehend the universe of possibilities they’ve created. I worked in a software company some years ago, the software was so large that developers specialized in certain niches and noone had a holistic view of the code and its interdependencies.

• So complex engineered systems give us a false sense of control and security. We think we understand how they work but that is a superficial understanding. There are lots of engineering disasters to prove that. Bridges that collapse, aircraft that crash, etc.

• Managers would rather live with a problem they can’t solve than with a solution they don’t understand. But do you understand PowerPoint? Do you understand Windows or a computer? Do you think pilots understand an aircraft?

• There is a testing issue: evolving tests?

Efficient vs inefficient, redundant

• That may be right, but not necessarily. Optimization is often performed evolutionary algortithms.

• In fact, evolved systems are more robust. If you engineer you have to plan for all possible situations, which is of course impossible. So why not build in adaptivity.

Design and evolution on a continuum

• What is evolved does not evolve from scratch: one starts with a set of building blocks and hypotheses and an objective function, which embody design and engineering knowledge. Examples: evolving molecules, evolving parts.

• We humans have been using the forces of evolution successfully for hundreds of years for breeding purposes without a deep understanding of the underlying biology. There is a middle way: interactive evolution.

Complexity Risk at LinkedIn

I am a big fan of LinkedIn’s. I even PAY for the service from time to time, just to be a power user for a couple of months. There are many ways in which the LinkedIn experience could be improved –their search capability is appalling, exploration is exhausting, not to mention the lack of tools for managing and making sense of a large group of connections and their streams of updates of all sorts. But I had a surprise of a different kind this morning. In fact, I had two surprises, outlined by two black squares on the screen snapshot. The first one is funny and highlights the difficulty of managing a complex system: the black square on the left shows a Yahoo News story (http://news.yahoo.com/linkedin-opts-100-million-users-sharing-private-information-050409746.html) recommended to me by LinkedIn. The story is about how LinkedIn has surreptitiously coerced 100 million users into sharing private information. Not exactly a positive for LinkedIn, and an upsetting story for many users –their tactics may even be unlawful in some countries. Which leads me to the second surprise, this one not funny and rather worrisome. It is in some ways related to the first surprise, but not in a good way. Look at the square on the right, and you’ll see the “People You May Know” suggestions from LinkedIn. These are usually people who are second- or third-order connections and going through the list of suggestions has become part of my daily routine –about half of all my connections come from going through the list and indeed finding people I know and want to connect with. Here, Matt Flannery, Kiva.org’s CEO is a second-order connection, which means that we have several connections in common. I do not know Matt personally, but I am not surprised to see him on the list. The other two suggestions are very spooky. Here is why: we are not second- or third- or even fourth-order connections, but I do know them –sort of. They are both managers of hotels I stayed at in the last two months. I booked one through Expedia and the other one through Tablet Hotels (if you like design hotels and don’t know www.tablethotels.com, go now!). How does LinkedIn know about the hotels I booked through different providers? Are Tablet Hotels and Expedia both providing my private information to LinkedIn? I haven’t found any indication of it, but still exploring. If the data is not coming from Expedia and Tablet Hotels, then the only way that LinkedIn could have found out about these hotels is by spying on me, following my web surfing. So in a way I am hoping that Expedia and Tablet Hotels did share my data with LinkedIn, though I would be really upset. Let’s assume for a minute that LinkedIn did what I suspect they did (which is not totally unlikely given the story in the left square), why would they risk alienating their users? They recently beat the Street with their revenues and earnings, which might be a short term reward for tactics that might backfire. The problem for LinkedIn is that every action is connected to everything else, an obvious consequence of their very business model. They are not the only ones, of course, but their value comes from connectivity, and so might their undoing. They have created a complex ecosystem, whose fragility or robustness is unknown. Ecological complexity might provide not just a good metaphor but also a scientific framework for assessing the impact of perturbations to the ecosystem. In the meantime, I am disabing cookies.

Pharmaceutical compounds need to become portable information assets

The pharmaceutical industry’s megatrends do not bode well for major pharmaceutical and biotech companies: it takes more and more money and time to develop a new drug, the number of costly Phase III and post-launch failures is increasing, patent life is decreasing as generics reach the market earlier and more aggressively than ever, and the blockbuster model itself is under considerable pressure with advances in genomics and the discovery of drug response markers. In the short term, big pharma will remain partners of choice for scaling up –large scale clinical trials, manufacturing, sales and marketing, as long as the blockbuster model can be forced upon payers and patients. But innovation is no longer coming from big pharma, and the rise of tailored drugs that address smaller markets will make the economics of being a fully integrated pharmaceutical company untenable. The key to big pharma’s reinvention will be its ability to form fluid drug development networks. But despite advances in information and communications technologies, the coordination and transaction costs of operating in a networked mode remain high. The single most important reason for this is the absence of agreed upon standards for representing drug assets. Each company has its own idiosyncratic way of storing and exploiting information, making the hundreds of megabytes of documents and images difficult to share across companies. As a result, not only is collaboration between companies costly, market transactions are also highly inefficient. For example, many potentially valuable assets are left partially developed or undeveloped inside small and large pharma companies, VCs, foundations or academia because the time and resources needed to assess them in the absence of common standards make due diligence impractical on a large scale. Furthermore, assets originating from big pharma tend to have more rigorous and consistent data than those from other origins, compounding the issue.

Due diligence remains an ad hoc and idiosyncratic process. It is not uncommon for the seller of an asset to have to set up as many data rooms as there are potential acquirers, as each one requires its own format. If a common asset representation standard existed, it could be used not just to speed up transactions and reduce coordination costs –a big deal already, it would also make it possible to identify information gaps and promote the bridging of these gaps by anyone competent to do so. This relatively simple step of creating an information representation standard that would be accepted across the industry would unleash tremendous innovation in drug development: foundations or even patient groups could have drugs developed that target specific markets too small for big pharma; VCs would have the ability to develop languishing assets from failed portfolio companies; small pharma and biotech companies would ...

Despite a growing recognition by large pharmaceutical and biotechnology companies that they need to rely on an ecology of external partners for innovation, the rate at which partnerships form is not increasing as it should and the pace of innovation shows clear signs of decline. The transaction costs of coordinating the many developments necessary to take products to market remain high outside of the traditional vertical integration FIPCO model.

A large portion of the blockbuster drugs currently on the market are the results of partnerships or licensing deals. Beyond the blockbuster model, which itself is under attack, big pharma companies need to rely on nimbler partners.

Consider the following examples:

• Many compounds with great potential are not developed beyond the early clinical phases because they address markets that are too small. But the costs of shopping and transferring such compounds to other parties are too high for more than a tiny fraction to be sold or licensed out.

• Conversely, thousands of promising compounds are trapped in academia, government labs, foundations and venture capital firms. The cost of evaluating such compounds is simply too high.

• Even when partnerships are formed, making the necessary data and information available to the partners is a daunting task.

The number and speed of partnerships must increase by a significant amount. This will lead to the formation of fluid “plug-and-play” networks. Big pharma can have a major role to play as network orchestrators: FIPNET model. Businesses will emerge for the sole purpose of orchestrating development activities. As a result, many asset owners who until now did not have access to an integrated development capability will now be able to have their assets developed.

In addition, this will shift the way that the value of IP is captured, from a focus on composition of matter to a focus on indication (method of use or MOU) –in other words, from the protection of atoms to the protection of bits. That’s because the flow of information among partners requires clear mechanisms for assigning credit. Value will come from the interpretation of data, not from the data itself, so efforts of those who generate useful data have to be rewarded. This will enable other forms of innovation, such as MOU innovation, which are critical to the emergence of tailored therapeutics as well as the repurposing of existing, even off-patent, drugs. MOU innovation tends not to be a preferred path for most companies for reasons that have to do not just with patent protection but also with the lack of data standards.

When drugs become portable (inter-operable) information assets, information standards will unleash innovation.

Complexity in the Non-Profit Sector (1): Evaluation

Enormous amounts are invested by foundations each year into the non-profit sector. The traditional model of evaluation of foundation-funded programs is a very linear one:

inputs=>outputs=>outcomes=>impact

Human systems are inherently complex and require a different approach –a complex adaptive systems (CAS) approach. Foundation-funded interventions and programs do not take place in a vacuum but in a highly interconnected human ecosystem where one action can have multiple effects, can be amplified or be dampened, and can cascade into large events for the better or for worse. With a CAS approach, evaluation could become more than just an ineffective, backward-looking/post-mortem tool, it could also serve as a pro-active, forward-looking method for the assessment of future investment opportunities to maximize impact: what interventions can we design and/or fund that will have the most impact in a cost-effective manner? If success is appropriately defined and the parameters of success are better understood, more effort can be directed toward factors that improve the odds of success. The application of complexity science to the issue of evaluation in the non-profit sector has the potential to be a game changer.

A simple example, to provide the beginning of a framework, is the Earned Income Tax Credit (EITC) program (http://en.wikipedia.org/wiki/Earned_income_tax_credit), the adoption of which has been supported by multiple foundations, such as the Annie E. Casey Foundation. In principle, the success of the program can be assessed by looking at the numbers of people enrolled each year and the amount of money they represent. Following this metric, the diagram below from the Tax Policy Center, a joint-venture between the Urban Institute and the Brookings Institution (http://www.taxpolicycenter.org/taxfacts/displayafact.cfm?Docid=266) suggests that the program has been incredibly successful.

Other metrics have been used to assess the program’s reach (a 2006 report by EITC expert Steve Holt provides a great overview), but they all tend to be focus on numbers of people participating (including number of people in different socio-economic-ethnic categories) and amounts of money received.

It all sounds very one-dimensional. Does this money really go back into the community, and if so, how? Do the children of EITC families benefit? Does EITC really improve asset building over time? What happens to all the families that enroll only once, never to return? The simplistic diagram below illustrates what a systemic view of the EITC ecosystem may look like, with positive and negative synergies among the many different ecosystem components and stakeholders. That is the whole point: the EITC can have as many, if not more, unintended than intended consequences due to all these constituent units interacting. It raises many more questions than it answers and I view it as a starting point for thinking about evaluating a program like EITC by taking the entire ecosystem in which it operates into account. If the ultimate objective of the EITC is poverty alleviation, or, even more narrowly, asset development, we have to take many, many other things into account. The aggregate result is emergent.

A more recent, and equally insightful, report by Steve Holt (http://www.brookings.edu/~/media/Files/rc/papers/2011/0418_eitc_holt/0418_eitc_holt.pdf) suggests that the success of EITC goes beyond the simple quantitative metrics. But the evidence is more anecdotal than systematic. Even some of the quantitative metrics point to a more holistic view, for example, “having a tax refund directly deposited to a financial institution account”, which may be the beginning of asset development and savings. But data needs to be collected over time across multiple touchpoints to see if that is really the beginning of asset development or just an illusion. A CAS model will help determine what data needs to be collected to understand how the various factors contribute to the ultimate objectives, not just proxies that may or may not, in the end, reflect those objectives. Lastly, synergies among various components of the system have been discovered over time mostly by accident, and probably at a great cost. A systemic model can help uncover synergies in silico and dramatically accelerate the costly and lengthy trial-and-error process. Identifying leverage points means we can also determine where to invest, or how to design a more effective ecosystem.

Dealing with Budget Cuts in the Military: Reframing Budgetary Decision-Making

Drastic cuts in budgets are forcing US Government decision makers across all departments to radically change the way they allocate resources. The cognitive mechanisms involved in budgetary decisions in such a rapidly changing environment are likely to lead to sub-optimal resource allocation and severe performance degradation. The department that is perhaps the most vulnerable to these effects is, paradoxically, the one with the highest budget: the Department of Defense has benefited from a decade of budget growth and is now facing a dramatically different budget situation. Behavioral economics and cognitive psychology provide a powerful and practical framework for understanding and helping to improve decision making.

After a decade of fast-growing budgets, military decisions makers are now facing significant budget cuts and must allocate resources accordingly. A number of military officers and senior civilians are confronted with a down budget for the first time as decision makers: their mental framework for allocating resources and prioritizing needs is entirely defined by their experience of a growing budget and dramatically fewer constraints than today. As a result, the prevailing decision heuristics (the subconscious cognitive mechanisms, or “mental accounting”, by which humans make decisions) in use today are heuristics that worked in a very different environment and under very different constraints, and are unlikely to perform adequately in the new budget environment –particularly during a transition period.

Not only are these heuristics not adapted to the current environment, their effects are amplified by a second set of cognitive biases usually activated in a period of rapid change. For example, a set of heuristics called availability heuristics tend to focus a decision maker’s attention to the “stuff” that is easiest to comprehend, what’s “available” at the time the decision needs to be made, the last thing he or she may have been working on, or, in other words, the “cognitively obvious”. Availability heuristics can work well when the environment is stable: after all, if things don’t change too quickly, the focus of a decision maker’s attention today should probably be similar to yesterday’s focus of attention. But in a rapidly changing environment, things that mattered a lot yesterday may have to be relegated to the background for today’s decisions. A concrete example is the disproportionate amount of attention dedicated (and rightly so) to the force deployed and the multiple zones of US military presence, which produces, in a period of declining budget, a similar disproportionate focus on the out force when looking for cuts, while it represents only a fraction of the total budget.

The challenge of having to cut costs and continue to perform well is not unique to the military. A 2009 Booz & Co survey of executives at the largest corporations in the US shows that short-term, “available” approaches prevail (Banerjee et al., Cut costs, grow stronger, Strategy+Business, Autumn 2009; http://www.strategy-business.com/media/file/00001.pdf). The temptation to reduce costs across the board is dominant. The figure below, which summarizes the survey, shows that it is difficult to “be strategic when the clock is ticking.”

A number of cognitive heuristics likely play a significant role in shaping decisions in a sharply declining budget. A non-exhaustive list would include:

Identifying the nature, scope, frequency and intensity of the budgetary decision heuristics is crucially important to the implementation of mitigating measures. Once these have been identified, cognitive aids and decision frameworks, methods and processes can be designed to correct for the systematic biases resulting from decision heuristics. For example, default settings, red teaming and the reframing of budgetary decisions can dramatically reduce biases.

A two-stage “cognitive audit” approach would consist of:

• Designing and running behavioral economics experiments and adaptive survey-style exercises to assess, both qualitatively and quantitatively, the heuristics at work in a variety of decisions. Metrics can be derived from quantifying the degree to which heuristics are used. Multiple experiments should be carried out over time to measure any changes.

• Simulating the impact of the identified heuristics on budgetary decisions by building artificial agents endowed with the heuristics. The first stage identifies heuristics, while the second stage identifies heuristics with the most impact. Metrics can be derived in the simulation from measuring decision performance relative to near-optimal outcomes, which can be calculated.

Heuristics having the most impact are targets for interventions. If they can be mitigated, they constitute high return-on-investment, leverage points toward dramatically improved decisions in a sharply declining budget.

Decision-support frameworks and tools, both at the individual level and at the group level, exist to help correct the biases inherent in human decision making. The table below offers a number of examples of mitigation mechanisms for each of the biases cited earlier.

Cybersecurity: Behavior, Behavior, Behavior

Every day we hear reports of new cyber-threats, and every single time they point to the same culprit: people as the weakest link in cyber-security. In addition to my earlier rant (http://info.icosystem.com/Modeling-for-Business-and-Government-Applications-Bl/bid/56302/Cyber-Security-Human-Behavior-Matters), a great piece was posted a few weeks ago that articulates the issue very well (http://gcn.com/articles/2011/09/23/cybereye-security-singularity-human-threat.aspx). A case in point is the recent drone virus revealed by Wired (http://www.wired.com/dangerroom/2011/10/virus-hits-drone-fleet/). It is a great example of the lack of appreciation for the tradeoffs you need to make when running missions. After the 2008 incident in which an infected removable media drive was the vector of entry for a worm into an overseas secret-level DoD network, the use of USB drives has been severely restricted throughout the military. Predator and Reaper drone crews at the Creech Air Force Base in Nevada where a large number of drone missions are conducted, however, use removable hard drives to load map updates and transport mission videos from one computer to another, which is probably the entry vector for keylogger virus here again. But how else can you run drone missions? Wanna try flying without a map? It turns out that manned aircraft resort to the same approach: pilots upload maps to their flight computers using removable drives. What’s the alternative? The cost of doing business in an environment where the most efficient practices are prohibited is unbearable –especially if the other side does not have such restrictions. Another layer of absurdity was added to the drone case when it was revealed by Wired again (http://www.wired.com/dangerroom/2011/10/drone-virus-kept-quiet/) that even after the virus was discovered the Creech Air Force Base did not inform the Air Force cybersecurity unit –probably because they didn’t think it was a serious threat, or for fear of reprisal in case the infection was due to, say, military personnel playing Mafia Wars. Even though the Air Force denies this version (but many indicators suggest it is close to reality), it is again human behavior that amplified the potential security consequences of the threat. Ok, now what was the response from the DoD leadership? “Drone units at other Air Force bases worldwide have now been ordered to stop their use.” What about the missions? How do you upload the maps? What type of tricks do you think the pilots (of unmanned or manned aircraft) will use to do what they need to do? They will find a way around the ban because they have to, so the fix may end up being worse than the initial problem, except now we don’t know what the counter-trick is. Great.

Along the same lines, the excellent Security Innovation Network or SINET (http://www.security-innovation.org/), an organization dedicated to “advancing innovation and enabling global collaboration between the public and private sectors to defeat Cybersecurity threats”, recently announced their selection of 16 innovative security firms for 2011 (http://www.security-innovation.org/Showcase-2011-Innovators.htm):

I have nothing against these firms, and I am sure that the SINET team did a great job of vetting their technological capabilities. But, just in case you haven’t noticed, they provide very little in the way of mitigating or shaping behavior.

Examples where a CAS approach would benefit include:

• Community building/management

• Asset building

• Poverty alleviation

• Education

Issues that need to be addressed:

• Timescales: significant delays between intervention and success

• Complexity and multiplicity of pathways

• Inconsistent use of language

• Lack of common measures in the social sector

• Lack of quality data on social impacts, outcomes, outputs, and cost

• Lack of incentives for transparency

• Unintended consequences

• Inadequate utilization

• Cost of measurement

Scientific concepts of relevance:

• Complex adaptive systems

• Network science

• Adaptation and learning

• Econometrics (e.g., finding surrogates for variables that can’t be measured directly and/or now)

http://slashdot.org/story/11/08/10/0549213/IBM-Plays-emSimCityem-With-Portland-Oregon

"Portland, Oregon will be the first city to use IBM's new software called Systems Dynamics for Smarter Cities, containing 3,000 equations which collectively seek to model cities' emergent behavior and help them figure out how policy can affect the lives of their citizens. The program seeks to quantify the cause-and-effect relationships between seemingly uncorrelated urban phenomena. 'What's the connection, for example, between ... obesity rates and carbon emissions?' writes Greg Lindsay. 'To find out, simply round up experts to hash out the linkages, translate them into algorithms, and upload enough historical data to populate the model. Then turn the knobs to see what happens when you nudge the city in one direction.' One of the drivers of the 'Portland Plan' is the city's commitment to a 40 percent decrease in carbon emissions by 2030, which necessitates less driving and more walking and biking. After running the model, planners discovered a positive feedback loop: More walking and biking would lead to lower obesity rates for Portlanders. In turn, a fitter population would find walking and biking a more attractive option. But as the field of urban systems gathers steam, it's important to remember that IBM and its fellow technology companies aren't the first to offer a quantitative toolkit to cities. In the 1970s, RAND built models they thought could predict fire patterns in New York, and then used them to justify closing fire stations in NYC's poorest sections in the name of efficiency, a decision that would ultimately displace 600,000 people as their neighborhoods burned."

Article location:http://www.fastcompany.com/1772083/ibm-partners-with-portland-to-play-simcity-for-real

August 8, 2011

Tags: Innovation, Technology, Ethonomics, Social Responsibility

IBM Partners With Portland To Play SimCity For Real

By Greg Lindsay

Can the complexity of cities really be reduced to a single set of equations, as the physicist Geoffrey West claims [1], or even 3,000 of them? Is it really true, as West’s numbers would indicate, that Corvallis, Oregon--a city of 55,000 two hours’ drive south of Portland--is the most innovative city in America? Perhaps there’s something in the water, or it may have more to do with the fact that West's model loves patents and Hewlett Packard’s Advanced Products Division is based there [2], along with its patent portfolio, one developed by thousands of researchers worldwide.

West’s conclusions are only as good as the data and the models (patents equal innovation) he has to work with. This problem--if you can’t measure it, you can’t manage it--combined with the impulse to improve cities by models, is driving both IBM’s “smarter city [3]” strategy and the nascent “urban systems” movement, which seek to apply complexity science to cities. IBM sponsored the first Urban Systems Symposium [4] in May (where West co-starred in a show-stopping discussion [5] with Paul Romer and Stewart Brand) and today announced the latest plank in its smarter city platform: an “app” containing 3,000 equations which collectively seek to model cities’ emergent behavior. IBM also revealed its first customer, the City of Portland, Oregon.

Systems Dynamics for Smarter Cities, as the app is called, tries to quantify the cause-and-effect relationships between seemingly uncorrelated urban phenomena. What’s the connection, for example, between public transit fares and high school graduation rates? Or obesity rates and carbon emissions? To find out, simply round up experts to hash out the linkages, translate them into algorithms, and upload enough historical data to populate the model. Then turn the knobs to see what happens when you nudge the city in one direction.

“While other analytical approaches rely on breaking a problem down into smaller and smaller pieces, the model we've created recognizes that the behavior of a system as a whole can be different from what might be anticipated by looking at its parts,” Michael Littlejohn, vice president of strategy for Smarter Cities at IBM, said in this morning’s press release.

IBM pitched Portland in 2009 to assist in the creation of the “Portland Plan [6],” a 25-year road map for the city and its related government agencies, the first draft of which will be released later this month. The city’s goal was more modest than IBM’s, less a model of everything than a chance to “shine a light on the biggest drivers of change,” according to Joe Zehnder, the city’s chief planner. One of those drivers is the city’s commitment to a 40 percent decrease in carbon emissions by 2030, which necessitates less driving and more walking and biking. Running the model, Zehnder and his fellow planners found that obesity rates fell as the number of cyclists rose, which led to a further increase in biking. This knowledge proved useful both in formulating policy (more bicycle lanes, anyone?) and in creating metrics to measure their success down the road.

But Zehnder is also quick to point out the limitations of such software, both in terms of their ability to sway the public and to simulate reality. “As we sat down with the modelers, we had to make the point to them that we will not be able to convince our constituents to trust anything coming out of a black box,” he says, adding “the whole act of choosing variables is a political one, a value-laden one,” underscoring the fact that choosing what to measure and what not to measure not only compromises the integrity of any model’s ability to reflect reality, but also the prerogatives of the ones building the model.

As the field of urban systems gathers steam, it’s important to remember that IBM and its fellow technology companies aren’t the first to offer a quantitative toolkit to cities. In 1968, New York City mayor John Lindsay (no relation) announced the creation of the New York City-RAND Institute, an effort to apply the “RAND method” of game theory and “systems analysis” to city management, and as Lindsay put, complete “the introduction into city agencies of the kind of streamlined, modern management that Robert McNamara applied in the Pentagon with such success in the past seven years.”

As Joe Flood described in his book The Fires* [7],* it turned out the Vietnam-era Pentagon was not the best role model. RAND promptly began building models they thought could predict fire patterns in New York, and then used them to justify closing fire stations in the poorest sections of the Bronx, Brooklyn and Harlem in the name of efficiency, a decision that would ultimately displace 600,000 people as their neighborhoods burned.

New York was hardly alone [8]. Across the country, mayors appealed to the best and brightest of RAND, Lockheed, and McDonnell to apply themselves to the “urban crisis” of the 1960s, leading urban planners to co-opt their rationalist mindset and many of their technologies, most notably GIS and satellite imagery. And what was in it for RAND? An opportunity to diversify beyond their Air Force contracts.

None of this is to say Portland will burn because of its decision to let IBM help it create a few metrics. But with mayor Sam Adams earnestly declaring that its software “can help us become a Smarter City” (as per the press release) it’s important to keep in the mind the limitations of modeling cities, and the dangers of forgetting.

[Images: Top, Flickr user jcestnik [9]; bottom: IBM]

Links: [1] link [2] link [3] link [4] link [5] link [6] link [7] link [8] link [9] link

August 9, 2011 BySarah Rich

Sometimes the public expects government to see the future. Good decisions are to be expected while bad decisions — judged with the benefit of hindsight — are derided.

Some help is on the way. IBM is betting that software can give cities a better look into their future.

The company unveiled new software this week designed to help decision-makers forecast the consequences of actions that affect citizens. In turn, IBM said the software, called System Dynamics for Smarter Cities, can also help policymakers make decisions that generate positive results.

What could the software be used for?

Imagine if a city eliminated student subsidies for public transportation. The software’s analysis of results might show that a municipal transportation agency would reduce costs, thus improving its financial picture. On the other hand, because more students would bypass mass transit and instead drive themselves to school, the city might suffer more traffic congestion, worse air quality, and might be forced to deploy more first responders, said Naveen Lamba, IBM’s leader on intelligent transportation and smarter government efforts for GBS — IBM’s consulting section.

Or what if a city decided to add highway tolls where tolls previously didn’t exist?

The results could show that toll revenue could go to the coffers of agency responsible for the tolls, but could also result in less traffic congestion on the toll road, Lamba said. Fewer vehicles would improve air quality and result in more usage of public transit.

The first city to try IBM's new solution is Portland, Ore. The city used the new software to help develop a model that illustrates how the actions coming from the city’s Portland Plan — long-range sustainability goals over the next 25 years — ripple across the priority areas the city has identified, said Joe Zehnder, the chief planner for Portland’s Bureau of Planning and Sustainability.

By using the software, Portland confirmed its plan to reduce carbon emissions 40 percent by 2030 would on the whole be a positive outcome. The city had already figured in that reducing the amount of driving done in the city would encourage other modes of transportation, such as walking and biking. But city officials also appear to have discovered a positive feedback loop: More walking and biking would lead to lower obesity rates for Portlanders. In turn, a fitter population would find would find walking and biking a more attractive option.

The new software is built atop what IBM calls a “dynamic engine,” which contains 3,000 analytical equations derived from the company’s past work with cities. The system is customized for a particular city by gathering information from the city’s subject matter experts and combining it with existing government data, such as budget allocations, number of K-12 students, unemployment rates, population growth and many other indicators.

The result, according to IBM, is a “system of simultaneous differential equations” compared to as much as a decade’s worth of a city’s historic data.

In more basic terms, cities provide data from multiple domains — such as transportation, public safety and government services — that’s inputted into the software, which then generates “what-if” scenarios that predict positive and negative outcomes, Lamba said.

“The model is not predictive; it’s more illustrative of the kind of impacts — the direction of the impacts — positive or negative that you might see, and maybe the scale of impact,” said Portland’s Zehnder said. “But it’s also driven by the assumptions that you build into the model.”

To build the model, last year IBM started working with Portland subject-matter experts to learn about the city’s interconnections. Researchers from Portland State University and systems software company Forio Business Solutions. The stakeholders collected 10 years of city data, such as vehicle miles traveled per day per person, the Census population and the percent of populations in subsidized housing for the model.

Portland’s benefited from being an early adopter. The software development came at no cost to the city because IBM initiated the partnership as a test case.

The project is connected to IBM's Smarter Planet initiative, Lamba said. Other cities have shown interest in deploying the new predictive software, Lamba said, adding that the municipalities haven’t been publicly announced.

ANNALS OF TECHNOLOGY

THE TWEAKER

The real genius of Steve Jobs.

by Malcolm GladwellNOVEMBER 14, 2011

Jobs’s sensibility was more editorial than inventive. “I’ll know it when I see it,” he said.

Not long after Steve Jobs got married, in 1991, he moved with his wife to a nineteen-thirties, Cotswolds-style house in old Palo Alto. Jobs always found it difficult to furnish the places where he lived. His previous house had only a mattress, a table, and chairs. He needed things to be perfect, and it took time to figure out what perfect was. This time, he had a wife and family in tow, but it made little difference. “We spoke about furniture in theory for eight years,” his wife, Laurene Powell, tells Walter Isaacson, in “Steve Jobs,” Isaacson’s enthralling new biography of the Apple founder. “We spent a lot of time asking ourselves, ‘What is the purpose of a sofa?’ ”

It was the choice of a washing machine, however, that proved most vexing. European washing machines, Jobs discovered, used less detergent and less water than their American counterparts, and were easier on the clothes. But they took twice as long to complete a washing cycle. What should the family do? As Jobs explained, “We spent some time in our family talking about what’s the trade-off we want to make. We ended up talking a lot about design, but also about the values of our family. Did we care most about getting our wash done in an hour versus an hour and a half? Or did we care most about our clothes feeling really soft and lasting longer? Did we care about using a quarter of the water? We spent about two weeks talking about this every night at the dinner table.”

Steve Jobs, Isaacson’s biography makes clear, was a complicated and exhausting man. “There are parts of his life and personality that are extremely messy, and that’s the truth,” Powell tells Isaacson. “You shouldn’t whitewash it.” Isaacson, to his credit, does not. He talks to everyone in Jobs’s career, meticulously recording conversations and encounters dating back twenty and thirty years. Jobs, we learn, was a bully. “He had the uncanny capacity to know exactly what your weak point is, know what will make you feel small, to make you cringe,” a friend of his tells Isaacson. Jobs gets his girlfriend pregnant, and then denies that the child is his. He parks in handicapped spaces. He screams at subordinates. He cries like a small child when he does not get his way. He gets stopped for driving a hundred miles an hour, honks angrily at the officer for taking too long to write up the ticket, and then resumes his journey at a hundred miles an hour. He sits in a restaurant and sends his food back three times. He arrives at his hotel suite in New York for press interviews and decides, at 10 P.M., that the piano needs to be repositioned, the strawberries are inadequate, and the flowers are all wrong: he wanted calla lilies. (When his public-relations assistant returns, at midnight, with the right flowers, he tells her that her suit is “disgusting.”) “Machines and robots were painted and repainted as he compulsively revised his color scheme,” Isaacson writes, of the factory Jobs built, after founding NeXT, in the late nineteen-eighties. “The walls were museum white, as they had been at the Macintosh factory, and there were $20,000 black leather chairs and a custom-made staircase. . . . He insisted that the machinery on the 165-foot assembly line be configured to move the circuit boards from right to left as they got built, so that the process would look better to visitors who watched from the viewing gallery.”

Isaacson begins with Jobs’s humble origins in Silicon Valley, the early triumph at Apple, and the humiliating ouster from the firm he created. He then charts the even greater triumphs at Pixar and at a resurgent Apple, when Jobs returns, in the late nineteen-nineties, and our natural expectation is that Jobs will emerge wiser and gentler from his tumultuous journey. He never does. In the hospital at the end of his life, he runs through sixty-seven nurses before he finds three he likes. “At one point, the pulmonologist tried to put a mask over his face when he was deeply sedated,” Isaacson writes:

Jobs ripped it off and mumbled that he hated the design and refused to wear it. Though barely able to speak, he ordered them to bring five different options for the mask and he would pick a design he liked. . . . He also hated the oxygen monitor they put on his finger. He told them it was ugly and too complex.

One of the great puzzles of the industrial revolution is why it began in England. Why not France, or Germany? Many reasons have been offered. Britain had plentiful supplies of coal, for instance. It had a good patent system in place. It had relatively high labor costs, which encouraged the search for labor-saving innovations. In an article published earlier this year, however, the economists Ralf Meisenzahl and Joel Mokyr focus on a different explanation: the role of Britain’s human-capital advantage—in particular, on a group they call “tweakers.” They believe that Britain dominated the industrial revolution because it had a far larger population of skilled engineers and artisans than its competitors: resourceful and creative men who took the signature inventions of the industrial age and tweaked them—refined and perfected them, and made them work.

In 1779, Samuel Crompton, a retiring genius from Lancashire, invented the spinning mule, which made possible the mechanization of cotton manufacture. Yet England’s real advantage was that it had Henry Stones, of Horwich, who added metal rollers to the mule; and James Hargreaves, of Tottington, who figured out how to smooth the acceleration and deceleration of the spinning wheel; and William Kelly, of Glasgow, who worked out how to add water power to the draw stroke; and John Kennedy, of Manchester, who adapted the wheel to turn out fine counts; and, finally, Richard Roberts, also of Manchester, a master of precision machine tooling—and the tweaker’s tweaker. He created the “automatic” spinning mule: an exacting, high-speed, reliable rethinking of Crompton’s original creation. Such men, the economists argue, provided the “micro inventions necessary to make macro inventions highly productive and remunerative.”

Was Steve Jobs a Samuel Crompton or was he a Richard Roberts? In the eulogies that followed Jobs’s death, last month, he was repeatedly referred to as a large-scale visionary and inventor. But Isaacson’s biography suggests that he was much more of a tweaker. He borrowed the characteristic features of the Macintosh—the mouse and the icons on the screen—from the engineers at Xerox PARC, after his famous visit there, in 1979. The first portable digital music players came out in 1996. Apple introduced the iPod, in 2001, because Jobs looked at the existing music players on the market and concluded that they “truly sucked.” Smart phones started coming out in the nineteen-nineties. Jobs introduced the iPhone in 2007, more than a decade later, because, Isaacson writes, “he had noticed something odd about the cell phones on the market: They all stank, just like portable music players used to.” The idea for the iPad came from an engineer at Microsoft, who was married to a friend of the Jobs family, and who invited Jobs to his fiftieth-birthday party. As Jobs tells Isaacson:

This guy badgered me about how Microsoft was going to completely change the world with this tablet PC software and eliminate all notebook computers, and Apple ought to license his Microsoft software. But he was doing the device all wrong. It had a stylus. As soon as you have a stylus, you’re dead. This dinner was like the tenth time he talked to me about it, and I was so sick of it that I came home and said, “Fuck this, let’s show him what a tablet can really be.”

Even within Apple, Jobs was known for taking credit for others’ ideas. Jonathan Ive, the designer behind the iMac, the iPod, and the iPhone, tells Isaacson, “He will go through a process of looking at my ideas and say, ‘That’s no good. That’s not very good. I like that one.’ And later I will be sitting in the audience and he will be talking about it as if it was his idea.”

Jobs’s sensibility was editorial, not inventive. His gift lay in taking what was in front of him—the tablet with stylus—and ruthlessly refining it. After looking at the first commercials for the iPad, he tracked down the copywriter, James Vincent, and told him, “Your commercials suck.”

“Well, what do you want?” Vincent shot back. “You’ve not been able to tell me what you want.” “I don’t know,” Jobs said. “You have to bring me something new. Nothing you’ve shown me is even close.” Vincent argued back and suddenly Jobs went ballistic. “He just started screaming at me,” Vincent recalled. Vincent could be volatile himself, and the volleys escalated. When Vincent shouted, “You’ve got to tell me what you want,” Jobs shot back, “You’ve got to show me some stuff, and I’ll know it when I see it.”

I’ll know it when I see it. That was Jobs’s credo, and until he saw it his perfectionism kept him on edge. He looked at the title bars—the headers that run across the top of windows and documents—that his team of software developers had designed for the original Macintosh and decided he didn’t like them. He forced the developers to do another version, and then another, about twenty iterations in all, insisting on one tiny tweak after another, and when the developers protested that they had better things to do he shouted, “Can you imagine looking at that every day? It’s not just a little thing. It’s something we have to do right.”

The famous Apple “Think Different” campaign came from Jobs’s advertising team at TBWA\Chiat\Day. But it was Jobs who agonized over the slogan until it was right:

They debated the grammatical issue: If “different” was supposed to modify the verb “think,” it should be an adverb, as in “think differently.” But Jobs insisted that he wanted “different” to be used as a noun, as in “think victory” or “think beauty.” Also, it echoed colloquial use, as in “think big.” Jobs later explained, “We discussed whether it was correct before we ran it. It’s grammatical, if you think about what we’re trying to say. It’s not think the same, it’s think different. Think a little different, think a lot different, think different. ‘Think differently’ wouldn’t hit the meaning for me.”

The point of Meisenzahl and Mokyr’s argument is that this sort of tweaking is essential to progress. James Watt invented the modern steam engine, doubling the efficiency of the engines that had come before. But when the tweakers took over the efficiency of the steam engine swiftly quadrupled. Samuel Crompton was responsible for what Meisenzahl and Mokyr call “arguably the most productive invention” of the industrial revolution. But the key moment, in the history of the mule, came a few years later, when there was a strike of cotton workers. The mill owners were looking for a way to replace the workers with unskilled labor, and needed an automatic mule, which did not need to be controlled by the spinner. Who solved the problem? Not Crompton, an unambitious man who regretted only that public interest would not leave him to his seclusion, so that he might “earn undisturbed the fruits of his ingenuity and perseverance.” It was the tweaker’s tweaker, Richard Roberts, who saved the day, producing a prototype, in 1825, and then an even better solution in 1830. Before long, the number of spindles on a typical mule jumped from four hundred to a thousand. The visionary starts with a clean sheet of paper, and re-imagines the world. The tweaker inherits things as they are, and has to push and pull them toward some more nearly perfect solution. That is not a lesser task.

Jobs’s friend Larry Ellison, the founder of Oracle, had a private jet, and he designed its interior with a great deal of care. One day, Jobs decided that he wanted a private jet, too. He studied what Ellison had done. Then he set about to reproduce his friend’s design in its entirety—the same jet, the same reconfiguration, the same doors between the cabins. Actually, not in its entirety. Ellison’s jet “had a door between cabins with an open button and a close button,” Isaacson writes. “Jobs insisted that his have a single button that toggled. He didn’t like the polished stainless steel of the buttons, so he had them replaced with brushed metal ones.” Having hired Ellison’s designer, “pretty soon he was driving her crazy.” Of course he was. The great accomplishment of Jobs’s life is how effectively he put his idiosyncrasies—his petulance, his narcissism, and his rudeness—in the service of perfection. “I look at his airplane and mine,” Ellison says, “and everything he changed was better.”

The angriest Isaacson ever saw Steve Jobs was when the wave of Android phones appeared, running the operating system developed by Google. Jobs saw the Android handsets, with their touchscreens and their icons, as a copy of the iPhone. He decided to sue. As he tells Isaacson:

Our lawsuit is saying, “Google, you fucking ripped off the iPhone, wholesale ripped us off.” Grand theft. I will spend my last dying breath if I need to, and I will spend every penny of Apple’s $40 billion in the bank, to right this wrong. I’m going to destroy Android, because it’s a stolen product. I’m willing to go to thermonuclear war on this. They are scared to death, because they know they are guilty. Outside of Search, Google’s products—Android, Google Docs—are shit.

In the nineteen-eighties, Jobs reacted the same way when Microsoft came out with Windows. It used the same graphical user interface—icons and mouse—as the Macintosh. Jobs was outraged and summoned Gates from Seattle to Apple’s Silicon Valley headquarters. “They met in Jobs’s conference room, where Gates found himself surrounded by ten Apple employees who were eager to watch their boss assail him,” Isaacson writes. “Jobs didn’t disappoint his troops. ‘You’re ripping us off!’ he shouted. ‘I trusted you, and now you’re stealing from us!’ ”

Gates looked back at Jobs calmly. Everyone knew where the windows and the icons came from. “Well, Steve,” Gates responded. “I think there’s more than one way of looking at it. I think it’s more like we both had this rich neighbor named Xerox and I broke into his house to steal the TV set and found out that you had already stolen it.”

Jobs was someone who took other people’s ideas and changed them. But he did not like it when the same thing was done to him. In his mind, what he did was special. Jobs persuaded the head of Pepsi-Cola, John Sculley, to join Apple as C.E.O., in 1983, by asking him, “Do you want to spend the rest of your life selling sugared water, or do you want a chance to change the world?” When Jobs approached Isaacson to write his biography, Isaacson first thought (“half jokingly”) that Jobs had noticed that his two previous books were on Benjamin Franklin and Albert Einstein, and that he “saw himself as the natural successor in that sequence.” The architecture of Apple software was always closed. Jobs did not want the iPhone and the iPod and the iPad to be opened up and fiddled with, because in his eyes they were perfect. The greatest tweaker of his generation did not care to be tweaked.

Perhaps this is why Bill Gates—of all Jobs’s contemporaries—gave him fits. Gates resisted the romance of perfectionism. Time and again, Isaacson repeatedly asks Jobs about Gates and Jobs cannot resist the gratuitous dig. “Bill is basically unimaginative,” Jobs tells Isaacson, “and has never invented anything, which I think is why he’s more comfortable now in philanthropy than technology. He just shamelessly ripped off other people’s ideas.”

After close to six hundred pages, the reader will recognize this as vintage Jobs: equal parts insightful, vicious, and delusional. It’s true that Gates is now more interested in trying to eradicate malaria than in overseeing the next iteration of Word. But this is not evidence of a lack of imagination. Philanthropy on the scale that Gates practices it represents imagination at its grandest. In contrast, Jobs’s vision, brilliant and perfect as it was, was narrow. He was a tweaker to the last, endlessly refining the same territory he had claimed as a young man.

As his life wound down, and cancer claimed his body, his great passion was designing Apple’s new, three-million-square-foot headquarters, in Cupertino. Jobs threw himself into the details. “Over and over he would come up with new concepts, sometimes entirely new shapes, and make them restart and provide more alternatives,” Isaacson writes. He was obsessed with glass, expanding on what he learned from the big panes in the Apple retail stores. “There would not be a straight piece of glass in the building,” Isaacson writes. “All would be curved and seamlessly joined. . . . The planned center courtyard was eight hundred feet across (more than three typical city blocks, or almost the length of three football fields), and he showed it to me with overlays indicating how it could surround St. Peter’s Square in Rome.” The architects wanted the windows to open. Jobs said no. He “had never liked the idea of people being able to open things. ‘That would just allow people to screw things up.’ ” ♦

ILLUSTRATION: ANDRÉ CARRILHO

Read more http://www.newyorker.com/reporting/2011/11/14/111114fa_fact_gladwell#ixzz1d2AFu9MK