Towards Thick Strategy Narratives

Narratives are getting to be a hot topic, so you'd think I'd be pleased that I've just published a book where they play a central role. A few people have even congratulated me on my timing. They think it is deliberate. Sadly, I am not so smart. In fact, if I'd seen this coming, I'd probably have picked something else to work on.

You see, I don't like working on popular, trendy things. I get anxious and irritable when I discover that others are working on the same ideas that I am. Call it intellectual agoraphobia, being an unsociable jerk, or just plain lack of competitive drive. So I haven't exactly been happy about narratives suddenly becoming a hot topic. It feels like I just went on a lovely solitary hike through some beautiful wilderness, and arrived at a great camping spot, only to discover that a whole noisy, partying crowd had also gotten there by a different route. Yeah, these are mean, turf-grubbing, selfish, uncharitable thoughts. It's not like I own Narrative National Park. Fortunately for my sanity, the big crowds seem to be headed along trails that don't interest me. In business and politics, much of the attention is on marketing and motivational narratives. In the broader cultural sphere, there's a lot of interest in identity narratives. There is also an anti-narrative movement focused on the problems of narrative approaches. If any of these topics interests you, here are some good starting points:

1. Marketing narratives: Made to Stick by Chip and Dan Heath
2. Motivational narratives: Squirrel Inc. by Stephen Denning
3. Identity narratives: The Redemptive Self by Dan McAdams
4. Anti-narrative movement: The Black Swan by Nicholas Nassim Taleb, Tyler Cowen's blog

My own interest is pretty narrowly focused on decision-narratives. The raw stream-of-consciousness story you tell yourself as you actually live through an experience and make your live, real-time decisions. The book is primarily about the fundamentals of the idea, but in this post, I want to explore an application of those fundamentals to the problem of crafting strategy narratives (the post is stand-alone; you don't need to have read the book). In particular, I am going to examine the limitations of existing varieties of strategy narratives, and argue that we need a new variety, thick strategy narratives. The Varieties of Strategy Narrative A strategy narrative is a story you tell to extract intelligence from real or imagined events (past, ongoing or predicted), in order to guide decision-making. I am going to limit myself to the business sector, but strategy narratives can be found in other domains like military thought, politics and scientific research as well (I might cover those domains in future posts). In business, case studies are one category of strategy narratives. The second major category is what I call is war stories, a category that includes many subcategories: biographies and autobiographies of key individuals, histories of companies or industry sectors, or the stories of significant trends. Each kind, even when done exceptionally well, has its limitations. In this post, I want to argue for a new, third kind of business strategy narrative: a thick strategy narrative, a mash-up of the idea of thick description in anthropology, and a certain approach to fictional storytelling around real events, best represented by David Milch's HBO show, Deadwood. Plus some extensions that don't exist in anthropology. To understand why we need a new variety, we have to understand the history of business strategy and the state of affairs it has led to. The Positioning-People Divide In the business strategy industry, the main historical divide has been between the so-called Positioning and People schools of thought. The Positioning School is all about theories of market structure and competition.  You can think of it as applied economics. The People School is all about theories of human motivation, interpersonal relationships and organizational psychology. Walter Kiechel's Lords of Strategy (the link is to my review/summary) is a magisterial study of the evolution of the field, and provides a great blow-by-blow account of how the field came to be. It explores the role played by academics like Michael Porter and Henry Mintzberg, the contributions of firms like BCG, Bain and McKinsey, the reasons why the Positioning School came to dominate the game, and why the People School might be making a comeback. In short, it is an excellent strategy narrative in its own right, with a self-referential focus on the strategy industry itself (very self-aware, these strategy types). To understand the role narratives have traditionally played in business strategy, and the expanded role that I hope they will start to play, you need to understand the divide between the two schools of thought. Kiechel's account of how the strategy industry grappled with the story of Honda's entry into the US market (yes, I know this is getting very meta: a story about a story about a story) illustrates the divide very well. The Honda Meta-Story In the early 80s when the entire American business strategy sector was trying to make sense of the Japanese threat with various models, Honda's successful entry into the motorcycle market was receiving particular attention.  Analysts were reading the Honda story in terms of things like experience curves, among the best conceptual tools available at that time. Everybody was applying their favorite theory. If the Honda story had happened today, people would be applying Christensen's theory of disruption or Friedman's theory of world-flattening or Gladwell's notion of tipping points. But in 1984 Robert Pascale did something that was unprecedented at the time for a strategist: he actually went and talked to the principals and learned the "real story" as the protagonists experienced it. He showed that the actual events and stream of decision-making had nothing to do with the models the analysts were using in after-the-fact analysis. The "real story" that Pascale dug out was messy and noisy, with many mis-steps and false starts. The protagonists were figuring stuff out as they went along, and had made no use of the elegant theories floating around (in fact, it is unclear whether they had even heard of them). A lot seemed to hinge on the specific personalities of the people who were leading the effort to enter the US market. The "real story" as it turned out, had a key "Aha!" moment. Honda had been attempting, in vain, to compete with Harley Davidson in the big machine market. Honda's initial products for that market were terrible. They were also using some light bikes for their own commuting needs while they were based in the US. The light bikes began attracting a lot of local attention, curiosity and demand. The Aha! moment was the protagonists recognizing the significance of this trickle of demand, and capitalizing on it. They did that by going after the small-bike market, carving out a position in that marginal market, using the secure marginal position to learn the skills involved in manufacturing larger motorcycles, and then attacking the main market. Today the Positioning School would probably explain this story using Christensen's disruption theory rather than BCG's experience curves, but Pascale's point was that the "real" story was a very path-dependent one involving real people, personalities, motivations and organizational cultures (in particular the startup culture of Honda's early outposts in the US). In short, he read the "real story" as a sort of vindication for the People School. The People-Positioning Détente But the People School reading of the story turned out to have its own limitations, since it did not adequately frame broader questions of industry structure or synthesize useful tools based on the analytical insights it offered. Unlike (say) Michael Porter's Five Forces model, or the idea of experience curves, the People School did not really tell you how to extract intelligence from the past and use it to solve your own problems better. The organizational and individual psychology models helped extract more nuanced insights from individual case studies, but the primary contribution of the People School was a negative one: driving home the point that reality is messy and that theories are neither necessary, nor sufficient. The Honda story seemed to show that real people and personalities matter, that local situational details and lucky breaks matter (like the Honda team using light bikes for their own needs in California, accidentally revealing latent demand), and perhaps most importantly, that you don't need to understand a situation with the "right" theoretical model in order to successfully navigate it (the "wrong" model or no model at all might do just as well). This seems completely obvious with hindsight. People ran successful businesses before the strategy industry came along to offer models. Alchemists were doing good chemistry with bad theories. Birds were flying around before Newton, Bernoulli and others created the theoretical knowledge that led to the modern theory of flight. But this understanding of the People-Positioning divide is disingenuous. Birds can fly without knowing the laws of flight. The Wrights could build the first airplane using only a very primitive version of the modern theory. But Boeing today cannot build a new airliner without applying all sorts of advanced theories. Abstract armchair theories matter. The Hegelian dialectic with its process of creative destruction of theories matters. The People School also has its own flaws. There is a reason I've been putting "real story" in scare quotes. The problem is that the story that you can dig out is full of noise, distractions, dead-end sub-plots, competing claims and people trying to make themselves look good. There are also the problems of poor recall, multiple confusing threads of action, unclear paths of causation and the effect of various cognitive biases (attribution error and survivorship bias in particular, are endemic diseases in business storytelling). The contentious nature of psychologically-based analyses is another problem. You will extract very different lessons out of the same events depending on whether you are a behaviorist or a neo-Freudian. Even the claim that real people, personalities, situational details and lucky breaks matter is actually suspect (that's why I said the Honda story only seems to demonstrate that). When conditions are right, the specific path-dependent course of events may be irrelevant. If enormous pressure builds up behind a dam, the specific crack that leads to the bursting of the dam, and the particular sequence of events that leads to devastation, is not particularly relevant. The important point is that pressure built up, and that there were cracks in the dam. There are cases where path-dependent histories actually matter, but they are not as common as you might think. None of these observations is particularly subtle. By the late 80s, the industry had recognized all these issues implicitly or explicitly, and a sort of détente evolved between the two schools. It was actually rather like the nuclear one: each school had the tools and critical arguments to completely undermine the other; a mutually-assured destruction scenario that nobody wanted (this is a more common situation in academia than is commonly recognized; the Hegelian dialectic is often deliberately slowed down for the collective benefit of a field). Over the long term, the Honda meta-story merely helped prevent overreach on the part of the Positioning School. The Positioning School had more internal consensus and a better repertoire of actual usable tricks, formulas and tools to offer, suitable for training armies of MBAs with some first-order thinking skills. The People School was more fragmented and contentious (Kiechel notes that while members of the Positioning School generally agree about what the top few concepts are, the members of the People School don't), and turned into a sort of diffuse and skeptical check-and-balance force that helped ensure that sufficient attention was being directed at psychological and situational factors. Thanks to the emergence of the case study as the main form of strategic "data," journals like Harvard Business Review managed to create a little cottage industry around using cases as the sporting arenas where theories could compete on relatively friendly terms (like the US and the Soviet Union at the Olympics) to offer insight into real or imagined situations. Sure, the characters and plots were rather anemic, and the drama rather tepid, but at least storytelling had carved out a formal role in the game. The two fields even managed to collaborate on a small side-project as it were, the study of leadership (this was rather like the Americans and Soviets with the Apollo-Soyuz Test Project). This led to a mini-boom in leadership studies (a somewhat disreputable subject academically speaking, but a huge money-spinner in the American business books market; the story of "leadership" as a subject is a fascinating one in its own right, and reveals some peculiarities of the American business world in particular). And in 2011, that's where the story still stands. The People School has perhaps been gaining a little more power, and the main tools in the Positioning School toolkit are starting to seem a little rusty, but the balance of power dynamic remains. The Limitations of Traditional Strategy Narratives The People-Positioning détente does not bother me. I think it is a healthy state of affairs.  It was an intellectual arms race that led to both schools getting much stronger, and businesses have benefited. Though I am personally partial to the People School (and my blogging claim to fame, The Gervais Principle, could be viewed as a contribution to the People School), I appreciate and use the tools from both sides. What bothers me is that this arms race between structural theories derived from economics and organizational/motivational theories derived from psychology has led to an overall strengthening of the role of theory over the role of narrative. The People School makes more use of narrative than the Positioning School, but for both, theory reigns supreme, and narrative is relegated to the status of "data." Fodder for analysis informed by different theories. This shows up in the format of the typical case study: a relatively bad (from a storyteller's point of view) narrative is laid out, and then the different theories engage in friendly combat over it, competing to offer more compelling insights and more persuasive recommendations. The dynamic also shows up in how a war story evolves. Once a significant story (like say the Southwest Airlines story) is discovered and a couple of rough versions are available, theoretical discussions take over. Let's look at the limitations of these two common types of narrative. Case Studies Case studies, especially the made-up ones, often leave me with an unresolved and inconclusive sense of "it's complicated." When I first started self-studying business ideas, I thought that perhaps I was the problem. Maybe as an engineer I was looking too hard for the "one right answer" to make sense of a business situation. Or perhaps, as Emerson said, consistency is the hobgoblin of small minds. But as I gained business experience, I realized that the problem was not with me. Case studies are actually written (or carefully chosen) to lead to inconclusive debates. In a détente, you do not want even friendly contests to lead to conclusive victories for any side. Unlike in the consulting industry, the point is not to "solve" a real case, but to use a carefully-engineered/selected dummy to practice skills. In real business situations, I found that there often is a right answer or at least a best gamble. Business consultants speak of "cracking a case" the way detectives speak of solving crimes. Only after I understood the Cold War contours of the world of case studies, did I understand that the primary purpose of the case study is a form of kata learning. They are intended as fertile opportunities that allow you to practice certain "moves," like applying disruption theory, and break away from the functional fixedness that can constrain your thinking if you only learn the few cherry-picked textbook examples that each theory comes packaged with (like disruption theory and the hard-drive industry). Like a good kata, a good case (fairly rare) is a richly layered construct. A novice MBA and Jack Welch may well form the same opinion about a case, and recommend the same course of action, but the latter's opinion will likely be much deeper, and more robust to variations and perturbations to the case. Cases, like kata, help you refine forms. Depending on your sophistication, the form you execute will be more legible to you, and represent more potential to accommodate perturbations. But kata cannot achieve what simulated or real combat can. In particular, live business situations may seem chaotic, but there is actually a good deal more clarity around motivations, priorities and commitments. There is a sense of what Mihalyi Csikzentmihalyi called "flow." In live situations, the unfocused and unresolved sense of "it's complicated" is actually a strong sign that you are not in flow, and are avoiding tough decisions and getting nowhere. War Stories In this category, as I mentioned, I include almost everything else besides case studies: biographies, autobiographies, corporate and industry histories, and the stories of key trends. The Honda story is an example. Each kind, of course, can be done very badly. Hagiographies by hacks infest each sub-category.  Marketing and brand narratives, and PR attempts to establish particular origin myths or spin negative events, confuse the situation (since genuine strategy narratives have a good deal more credibility than carefully manufactured marketing narratives, the latter often disguise themselves as the former). Even the non-business world intrudes, as insecure business leaders with low self-awareness attempt to shoehorn their personal identity and redemption narratives into the mold of the strategy narrative. There is of course, no formula for teasing out a usable strategy narrative from these contaminated available narratives. You just have to develop some taste, and read a good selection of well-done ones for calibration early on. My favorite benchmark examples, include Marc Levinson's history of the containerization and William Whyte's critical analysis of the Organization Man era (and of course, Kiechel's book). With a well-done war story, you do get a sense of real-time decision-making, the tempo of events as they played out, and at least a vague sense of when key insights allowed a stuck story get unstuck. Like a good detective story, a good war story shows you the clues as they were discovered, rather than in the most comprehensible order. It helps you recognize their significance in the same messy, out-of-sequence way that a fictional detective does. You get to experience key Aha! moments along with the protagonists. Those Aha! moments (what I call cheap tricks in the book) are central to strategy. They are what distinguish real-time decision-making based on an evolving data set from after-the-fact analysis. So this is a key advantage war stories have over case studies. Even when well done though, the problem with war stories as a whole is exactly the same one that limits the value of the "real" Honda story as told by Pascale: you don't get the benefit of available theories and structural models. Without a systematic approach to avoiding the various pathologies that can infect the story, even the most thoughtful and critical storytellers succumb to temptations such as redemptive plots, heroic archetypes and unexamined doctrines. Fortunately, there is a solution. Unfortunately, it is not a particularly easy one to use. The Anthropology of The Whole Elephant You could summarize the limitations of case studies as follows. Case studies function like the story of the three blind men and the elephant, except that only the parts that the blind men touch actually become accessible (the tail, trunk and one leg, or an ear in some versions). Manufactured case studies are worse: they are like fake elephants that are mostly wooden scaffolding, with more realistic fleshing out of the parts where the blind men are more likely to grope. The limitations of war stories are like a different mangling of the story. In this case, the blind men grope more of the elephant, but have a much smaller repository of useful notions to compare to their experience. Where the blind men, in the original fable, describe what they sense in terms of ropes, snakes, tree trunks and fans, in this version, the blind men have more data but are at a loss for words. Thick strategy narratives are an attempt to get at the whole elephant, without requiring the blind men to gain vision. They are a combination of two types of stories that are not commonly found in the business world:

• Thick descriptions: rich, critical ethnographies that consciously interleave a real-time narration of a fertile sequence of events with tasteful doses of models and theories. The classic thick description is the original one: Clifford Geertz' examination of the culture of cockfighting in Bali, Deep Play: Notes on the Balinese Cockfight. The narrative component is his own experience as an embedded ethographer (critically, one who has abandoned the detached-observer stance). The theoretical component is a set of ideas about the psychology of gambling due to Jeremy Bentham, the Utilitarian philosopher.
• Model-Driven Faction: "Faction" is a rather clumsy term in ethnography that is used to indicate ethnography in the form of imaginative fiction constructed out of facts (the fact that it also means "one side in a partisan conflict" makes it a Google-unfriendly term). By model-driven faction, I mean an attempt to use a theoretical model to weave a story around real events, and being willing to alter "facts" to serve the story. The best example (in fact the only one I know of) is David Milch's version of the story of the Black Hills Gold Rush in the 1870s, and in particular, the events in the mining town of Deadwood. Milch's stated goal was to use the story to explore classical political science models of how order emerges out of chaos.

Thick descriptions have three key elements: immersive and extensive participation as source material, blending of narration of live events with selected bits of history, and the interleaving of selected elements of one or more theoretical constructs, in a way that keeps them subservient to the narrative. Model-driven faction substitutes fictional reconstruction for immersive participation, and takes more liberties with the source material. It also integrates the theory into the plot development and archetypes, rather than offering interwoven meta-commentary. A key point to note is that model-driven faction does not mangle facts to serve the model. It mangles facts to enrich the story. In both thick description and model-driven faction, the quality of the story that can be told serves as a measure of the fertility of the theories that are mixed in, either organically, or as meta-commentary. This is a key characteristic that distinguishes both from a much more impoverished approach to mixing stories and theories, where facts are altered or manufactured to fit the theory rather than to improve the story itself, and make it more fertile. As a result, Deadwood has a compelling literary quality to it that (say) Ayn Rand's works or Eli Goldratt's The Goal do not (a dead giveaway is narrative anemia: when you fit a story to a theory, you will likely be forced to leave out bits that don't fit, leaving behind something that is impoverished). These approaches are not merely valuable because the resulting stories are more interesting to read. Fitting models "inside" stories rather than the other way around changes a potential Procrustean bed approach to one that explains the explainable and either leaves the unexplained intact as part of narrative tension, or manufactures new ideas to explain them. Bentham's notion of "deep play" and political science theories of "man in the state of nature" only explain part of the raw material in the respective stories of Balinese cockfighting or Deadwood. The leftover parts serve to provide the piquant colors and flavors that make for real stories, but also make the narrative a potential source of skeptical doubt and new models. Finally, the ability to alter facts allows the storyteller to prune away the irrelevant, compress multiple ideas into a single one, simplify the cast of characters and replace non-critical path-dependent subplots with ones that serve the story better. The net result is that a more potent narrative is distilled out of the "real story." A final point: the explicit and ironically self-aware insertion of theories (either Position or People theories) into the narrative prevents a common failure mode to which Hollywood often succumbs: creating faction to serve unexamined ordinary fictional goals, such as creating creating characters the audience can identify with, redemption arcs or projected plot conflicts that are chosen because they interest the audience rather than because they exist in the story itself.  Sorkin's version of the Facebook story illustrates these issues clearly, with its focus on the conflict with the Winkelvoss twins and an imaginary WASP-vs-Jewish-Kid in-group/out-group dynamic that Harvard alums point out bears no resemblance to the actual culture at Harvard. One reason I did not watch the movie was simply that the reviews told me the movie would not satisfy my business or technological curiosities. Thick strategy narratives are an attempt to get at the essence of the whole elephant without requiring non-blind men. The blind men will have to work harder, grope more of the elephant, use more reference concepts and manufacture new concepts in order to adopt a position, but the result will be vastly more useful. Towards Thick Strategy Narratives I haven't actually proposed a notion of "thick strategy narrative" that is distinct from the ethnographic parentage I am proposing, but it is already clear that ethnography is not what we need. Ethnographers attempt to understand the essence of something as a matter of particularist analysis. A Balinese cockfight qua Balinese cockfight. They certainly do not attempt to deeply examine theories or offer prescriptions. Rather than thinking of thick strategy narratives as a specific kind of construct, I prefer to think of as a predisposition to react in certain ways to new information, let's call it think-narrative thinking. What do I mean? Here are some examples:

1. Retelling: When you encounter a new theory, you might attempt to retell a story you know well (say Southwest Airlines) while the theory is fresh in your mind, taking whatever liberties you need to with the facts.
2. Transposition: When you encounter a new trend or market force, such as "globalization," you transpose some suitable old story to a modern setting and retell the story
3. Reverse cases: In a group, you come to a simplified consensual understanding of a theory (such as disruption) and compete to construct the best possible stories designed to illustrate the theory and its unexamined riches.
4. Farce: You take an existing telling of a story, and attempt to retell it as farce, turning Marx's observation ("history repeats itself, first as tragedy, then as farce") into a prescription
5. Joke Making: In a group, you share an incomplete Dilbert comic strip and propose a current, extant business theory or popular story, and attempt to come up with a joke about it.
6. Joke Deconstruction: You take actual business jokes, and attempt to deconstruct them and figure out why they are funny.
7. Archetype substitution: You take a story and replace a key protagonist with another (either a fictional archetype or a real person, such as "what if Steve Jobs had taken over from Bill Gates, instead of Ballmer?")
8. Doctrine substitution: You reduce the key protagonist of a story to a doctrine -- a few key beliefs about decision-making -- make some alterations, and retell the story.
9. Role playing: You pick a critical event, such as a meeting, based on a publicly-available description, and role-play the action, either attempting to act according to the character of the actual protagonists, or as yourself
10. Improvisation: You use a real business situation as the premise for an improv exercise (such as "You are George Bush reading to kids, and somebody has just informed you about 9/11")

I routinely try some of these methods, sometimes as private thinking exercises, and sometimes to generate writing ideas. Other methods just seem like good ideas to me right now. They might not be. Some of these ideas have already started percolating into the business world (simulation exercises and role-playing games are already common)  but the problem is that without a broader understanding of how and why such practices strengthen decision-making skills, and without mechanisms to integrate the experiences into everyday business thinking, they remain mere stimulating sideshows or end up serving other purposes such as trust-building in a group. Where can we go with these ideas? My broader hope, I suppose, is that business thinking can become a great deal more engrossing and more explosively dramatic  than the anemic parade of impoverished faddish models, weak storytelling and reactionary jokes that passes for discourse around what is probably the grandest of all human dramas. I am hoping to help steer business thinking in that direction with Tempo.