Surfing the Edge of Chaos

Every decade or two during the past one hundred years, a point of inflection has occurred in management thinking. These breakthroughs are akin to the S-curves of technology that characterize the life cycle of many industrial and consumer products: Introduction ? Acceleration ? Acceptance ? Maturity. Each big idea catches hold slowly. Yet, within a relatively short time, the new approach becomes so widely accepted that it is difficult even for old-timers to reconstruct how the world looked before.

The decade following World War II gave birth to the “strategic era.” While the tenets of military strategy had been evolving for centuries, the link to commercial enterprise was tenuous. Before the late 1940s, most companies adhered to the tenet “make a little, sell a little, make a little more.” After the war, faculty at the Harvard Business School (soon joined by swelling ranks of consultants) began to take the discipline of strategy seriously. By the late 1970s, the array of strategic concepts (SWOT analysis, the five forces framework, experience curves, strategic portfolios, the concept of competitive advantage) had become standard ordnance in the management arsenal. Today, a mere twenty years later, a grasp of these concepts is presumed as a threshold of management literacy. They have become so familiar that it is hard to imagine a world without them.

It is useful to step back and reflect on the scientific underpinnings to this legacy. Eric Beinhocker writes:

“The early micro-economists copied the mathematics of mid-nineteenth century physics equation by equation. [‘Atoms’] became the individual, ‘force’ became the economists’ notion of ‘marginal utility’ (or demand), ‘kinetic energy’ became total expenditure. All of this was synthesized into a coherent theory by Alfred Marshall — known as the theory of industrial organization.”1

Marshall’s work and its underpinnings in nineteenth century physics exert a huge influence on strategic thinking to this day. From our concept of strategy to our efforts at organizational renewal, the deep logic is based on assumptions of deterministic cause and effect (i.e., a billiard ball model of how competitors will respond to a strategic challenge or how employees will behave under a new incentive scheme). And all of this, consistent with Newton’s initial conceptions, is assumed to take place in a world where time, space (i.e., a particular industry structure or definition of a market), and dynamic equilibrium are accepted as reasonable underpinnings for the formulation of executive action. That’s where the trouble begins. Marshall’s equilibrium model offered appropriate approximations for the dominant sectors of agriculture and manufacturing of his era and are still useful in many situations. But these constructs run into difficulty in the far from equilibrium conditions found in today’s service, technology, or communications-intensive businesses. When new entrants such as Nokia, Amazon.com, Dell Computer, or CNN invade a market, they succeed despite what traditional strategic thinkers would write off as a long shot.

During the 1980s and 1990s, performance improvement (e.g., total quality management, kaizen, just-intime, reengineering) succeeded the strategic era. It, too, has followed the S-curve trajectory. Now, as it trails off, an uneasiness is stirring, a feeling that “something more” is required. In particular, disquiet has arisen over the rapidly rising fatality rates of major companies. Organizations cannot win by cost reduction alone and cannot invent appropriate strategic responses fast enough to stay abreast of nimble rivals. Many are exhausted by the pace of change, and their harried attempts to execute new initiatives fall short of expectations.

The next point of inflection is about to unfold. To succeed, the next big idea must address the biggest challenge facing corporations today — namely, to dramatically improve the hit rate of strategic initiatives and attain the level of renewal necessary for successful execution. As in the previous eras, we can expect that the next big idea will at first seem strange and inaccessible.

Here’s the good news. For well over a decade, the hard sciences have made enormous strides in understanding and describing how the living world works. Scientists use the term “complex adaptive systems” (“complexity” for short) to label these theories. To be sure, the new theories do not explain everything. But the work has identified principles that apply to many living things —amoebae and ant colonies, beehives and bond traders, ecologies and economies, you and me.

For an entity to qualify as a complex adaptive system, it must meet four tests. First, it must be comprised of many agents acting in parallel. It is not hierarchically controlled. Second, it continuously shuffles these building blocks and generates multiple levels of organization and structure. Third, it is subject to the second law of thermodynamics, exhibiting entropy and winding down over time unless replenished with energy. In this sense, complex adaptive systems are vulnerable to death. Fourth, a distinguishing characteristic, all complex adaptive systems exhibit a capacity for pattern recognition and employ this to anticipate the future and learn to recognize the anticipation of seasonal change.

Many systems are complex but not adaptive (i.e., they meet some of the above conditions, but not all). If sand is gradually piled on a table, it will slide off in patterns. If a wave in a stream is disturbed, it will repair itself once the obstruction is removed. But neither of these complex systems anticipates and learns. Only living systems cope with their environment with a predictive model that anticipates and pro-acts. Thus, when the worldwide community of strep bacteria mutates to circumvent the threat of the latest antibiotic (as it does rather reliably within three years), it is reaffirming its membership in the club of complexity.

Work on complexity originated during the mid-1980s at New Mexico’s Santa Fe Institute. A group of distinguished scientists with backgrounds in particle physics, microbiology, archaeology, astrophysics, paleontology, zoology, botany, and economics were drawn together by similar questions.2 A series of symposia, underwritten by the Carnegie Foundation, revealed that all the assembled disciplines shared, at their core, building blocks composed of many agents. These might be molecules, neurons, a species, customers, members of a social system, or networks of corporations. Further, these fundamental systems were continually organizing and reorganizing themselves, all flourishing in a boundary between rigidity and randomness and all occasionally forming larger structures through the clash of natural accommodation and competition. Molecules form cells; neurons cluster into neural networks (or brains); species form ecosystems; individuals form tribes or societies; consumers and corporations form economies. These self-organizing structures give rise to emergent behavior (an example of which is the process whereby pre-biotic chemicals combined to form the extraordinary diversity of life on earth). Complexity science informs us about organization, stability, and change in social and natural systems. “Unlike the earlier advances in hard science,” writes economist Alex Trosiglio, “complexity deals with a world that is far from equilibrium, and is creative and evolving in ways that we cannot hope to predict. It points to fundamental limits to our ability to understand, control, and manage the world, and the need for us to accept unpredictability and change.3

The science of complexity has yielded four bedrock principles relevant to the new strategic work:

1. Complex adaptive systems are at risk when in equilibrium. Equilibrium is a precursor to death.4
2. Complex adaptive systems exhibit the capacity of self-organization and emergent complexity.5 Self-organization arises from intelligence in the remote clusters (or “nodes”) within a network. Emergent complexity is generated by the propensity of simple structures to generate novel patterns, infinite variety, and often, a sum that is greater than the parts. (Again, the escalating complexity of life on earth is an example.)
3. Complex adaptive systems tend to move toward the edge of chaos when provoked by a complex task.6 Bounded instability is more conducive to evolution than either stable equilibrium or explosive instability. (For example, fire has been found to be a critical factor in regenerating healthy forests and prairies.) One important corollary to this principle is that a complex adaptive system, once having reached a temporary “peak” in its fitness landscape (e.g., a company during a golden era), must then “go down to go up” (i.e., moving from one peak to a still higher peak requires it to traverse the valleys of the fitness landscape). In cybernetic terms, the organism must be pulled by competitive pressures far enough out of its usual arrangements before it can create substantially different forms and arrive at a more evolved basin of attraction.
4. One cannot direct a living system, only disturb it.7 Complex adaptive systems are characterized by weak cause-and-effect linkages. Phase transitions occur in the realm where one relatively small and isolated variation can produce huge effects. Alternatively, large changes may have little effect. (This phenomenon is common in the information industry. Massive efforts to promote a superior operating system may come to naught, whereas a series of serendipitous events may establish an inferior operating system —such as MS-DOS — as the industry standard.)

Is complexity just interesting science, or does it represent something of great importance in thinking about strategic work? As these illustrations suggest, treating organizations as complex adaptive systems provides useful insight into the nature of strategic work. In the following pages, I will (1) briefly describe how the four bedrock principles of complexity occur in nature, and (2) demonstrate how they can be applied in a managerial context. In particular, I use the efforts underway at Royal Dutch/Shell to describe an extensive and pragmatic test of these ideas.

The successes at Shell and other companies described here might be achieved with a more traditional mind-set (in much the same way as Newton’s laws can be used to explain the mechanics of matter on earth with sufficient accuracy so as to not require the General Theory of Relativity). But the contribution of scientific insight is much more than descriptions of increasing accuracy. Deep theories reveal previously unsuspected aspects of reality that we don’t see (the curvature of space-time in the case of relativity theory) and thereby alter the fabric of reality. This is the context for an article on complexity science and strategy. Complexity makes the strategic challenge more understandable and the task of strategic renewal more accessible. In short, this is not a polemic against the traditional strategic approach, but an argument for broadening it.

Stable Equilibrium Equals Death

An obscure but important law of cybernetics, the law of requisite variety, states: For any system to survive, it must cultivate variety in its internal controls. If it fails to do so internally, it will fail to cope with variety successfully when it comes from an external source.8 Here, in the mundane prose of a cybernetic axiom, is the rationale for bounded instability.

A perverse example of this axiom in action was driven home by the devastating fires that wiped out 25 percent of Yellowstone National Park in 1992. For decades, the National Park Service had imposed equilibrium on the forest by extinguishing fires whenever they appeared. Gradually, the forest floor became littered with a thick layer of debris. When a lightening strike and ill-timed winds created a conflagration that could not be contained, this carpet of dry material burned longer and hotter than normal. By suppressing natural fires for close to 100 years, the park service had prevented the forest floor from being cleansed in a natural rhythm. Now a century’s accumulation of deadfall generated extreme temperatures. The fire incinerated large trees and the living components of top soil that would otherwise have survived. This is the price of enforced equilibrium.

The seductive pull of equilibrium poses a constant danger to successful established companies. Jim Cannavino, a former IBM senior executive, provides an anecdote that speaks to the hazards of resisting change. In 1993, Cannavino was asked by IBM’s new CEO, Lou Gerstner, to take a hard look at the strategic planning process. Why had IBM so badly missed the mark? Cannavino dutifully examined the work product — library shelves filled with blue binders containing twenty years of forecasts, trends, and strategic analysis. “It all could be distilled down to one sentence,” he recounts. “‘We saw it coming’ — PC open architecture, networking intelligence in microprocessors, higher margins in software and services than hardware; it was all there. So I looked at the operating plans. How did they reflect the shifts the strategists had projected? These blue volumes (three times as voluminous as the strategic plans) could also be summarized in one sentence: ‘Nothing changed.’ And the final dose of arsenic to this diet of cyanide was the year-end financial reconciliation process. When we rolled up the sector submissions into totals for the corporation, the growth opportunities never quite covered the erosion of market share. This shortfall, of course, was the tip of an iceberg that would one day upend our strategy and our primary product — the IBM 360 mainframe. But facing these fundamental trends would have precipitated a great deal of turmoil and instability. Instead, year after year, a few of our most senior leaders went behind closed doors and raised prices.”9

While equilibrium endangers living systems, it often wears the disguise of an attribute. Equilibrium is concealed inside strong values or a coherent, close-knit social system, or within a company’s well-synchronized operating system (often referred to as “organizational fit”). Vision, values, and organizational fit are double-edged swords.

Species are inherently drawn toward the seeming oasis of stability and equilibrium — and the further they drift toward this destination, the less likely they are to adapt successfully when change is necessary. So why don’t all species drift into the thrall of equilibrium and die off? Two forces thwart equilibrium and promote instability: (1) the threat of death, and (2) the promise of sex.

The Darwinian process, called “selection pressures” by natural scientists, imposes harsh consequences on species entrapped in equilibrium. Most species, when challenged to adapt too far from their origins, are unable to do so and gradually disappear. But from the vantage point of the larger ecological community, selection pressures enforce an ecological upgrade, insofar as mutations that survive offer a better fit with the new environment. Natural selection exerts itself most aggressively during periods of radical change. Few readers will have difficulty identifying these forces at work in industry today. There are no safe havens. From toothpaste to camcorders, pharmaceuticals to office supplies, bookstores to booster rockets for space payloads, soap to software, it’s a Darwinian jungle out there, and it’s not getting easier.

As a rule, a species becomes more vulnerable as it becomes more genetically homogeneous. Nature hedges against this condition through the reproductive process. Of the several means of reproduction that have evolved on the planet, sex is best. It is decisively superior to parthenogenesis (the process by which most plants, worms, and a few mammals conceive offspring through self-induced combination of identical genetic material).

Sexual reproduction maximizes diversity. Chromosome combinations are randomly matched in variant pairings, thereby generating more permutations and variety in offspring. Oxford’s evolutionary theorist, William Hamilton, explains why this benefits a species. Enemies (i.e., harmful diseases and parasites) find it harder to adapt to the diverse attributes of a population generated by sexual reproduction than to the comparative uniformity of one produced by parthenogenesis.10

How does this relate to organizations? In organizations, people are the chromosomes, the genetic material that can create variety. When management thinker Gary Hamel was asked if he thought IBM had a chance of leading the next stage of the information revolution, he replied: “I’d need to know how many of IBM’s top 100 executives had grown up on the west coast of America where the future of the computer industry is being created and how many were under forty years of age. If a quarter or a third of the senior group were both under forty and possessed a west coast perspective, IBM has a chance.”11

Here’s the rub: The “exchanges of DNA” attempted within social systems are not nearly as reliable as those driven by the mechanics of reproductive chemistry. True, organizations can hire from the outside, bring seniors into frequent contact with iconoclasts from the ranks, or confront engineers and designers with disgruntled customers. But the enemy of these methods is, of course, the existing social order, which, like the body’s immune defense system, seeks to neutralize, isolate, or destroy foreign invaders. “Antibodies” in the form of social norms, corporate values, and orthodox beliefs nullify the advantages of diversity. An executive team may include divergent interests, only to engage in stereotyped listening (e.g., “There goes Techie again”) or freeze iconoclasts out of important informal discussions. If authentic diversity is sought, all executives, in particular the seniors, must be more seeker than guru.

Disturbing Equilibrium at Shell

In 1996, Steve Miller, age fifty-one, became a member of Shell’s committee of managing directors — the five senior leaders who develop objectives and long-term plans for Royal Dutch/Shell.12 The group found itself captive to its hundred-year-old history. The numbing effects of tradition — a staggering $130 billion in annual revenues, 105,000 predominantly long-tenured employees, and global operations — left Shell vulnerable. While profits continued to flow, fissures were forming beneath the surface.

Miller was appointed group managing director of Shell’s worldwide oil products business (known as “Downstream”), which accounts for $40 billion of revenues within the Shell Group. During the previous two years, the company had been engaged in a program to “transform” the organization. Yet the regimen of massive reorganization, traumatic downsizing, and senior management workshops accomplished little. Shell’s earnings, while solid, were disappointing to financial analysts who expected more from the industry’s largest competitor. Employees registered widespread resignation and cynicism. And the operating units at the “coal face” (Shell’s term for its front-line activities within the 130 countries where Downstream does business) saw little more than business as usual.

For Steve Miller, Shell’s impenetrable culture was worrisome. The Downstream business accounted for 37 percent of Shell’s assets. Among the businesses in the Shell Group’s portfolio, Downstream faced the gravest competitive threats. From 1992 to 1995, a full 50 percent of Shell’s retail revenues in France fell victim to the onslaught of the European hypermarkets; a similar pattern was emerging in the United Kingdom. Elsewhere in the world, new competitors, global customers, and more savvy national oil companies were demanding a radically different approach to the marketplace. Having observed Shell’s previous transformation efforts, Miller was convinced that it was essential to reach around the resistant bureaucracy and involve the front lines of the organization, a formidable task given the sheer size of the operation. In addition to Downstream’s 61,000 full-time employees, Shell’s 47,000 filling stations employed hundreds of thousands, mostly part-time attendants and catered to more than 10 million customers every day. In the language of complexity, Miller believed it necessary to tap the emergent properties of Shell’s enormous distribution system and shift the locus of strategic initiative to the front lines. He saw this system as a fertile organism that needed encouragement to, in his words, “send green shoots forth.”

In an effort to gain the organization’s attention (i.e., disturb equilibrium), beginning in mid-1996, Miller reallocated more than 50 percent of his calendar to work directly with front-line personnel. Miller states:

“Our Downstream business transformation program had bogged down largely because of the impasse between headquarters and the operating companies Shell’s term for its highly independent country operations. The balance of power between headquarters and field, honed during a period of relative equilibrium, had ground to a stalemate. But the forces for continuing in the old way were enormous and extended throughout the organization. We were overseeing the most decentralized operation in the world, with country chief executives that had, since the 1950s, enjoyed enormous autonomy. This had been part of our success formula. Yet we were encountering a set of daunting competitive threats that transcended national boundaries. Global customers —like British Airways or Daimler Benz — wanted to deal with one Shell contact, not with a different Shell representative in every country in which they operate. We had huge overcapacity in refining, but each country CEO (motivated to maximize his own P&L) resisted the consolidation of refining capacity. These problems begged for a new strategic approach in which the task at the top was to provide the framework and then unleash the regional and local levels to find a path that was best for their market and the corporation as a whole.”

Shell had tried to rationalize its assets through a well-engineered strategic response: directives were issued by the top and driven through the organization. But country heads successfully thwarted consolidation under the banner of host-country objections to the threatened closing of their dedicated refining capacity. Miller continues: “We were equally unsuccessful at igniting a more imaginative approach toward the marketplace. It was like the old game of telephone that we used to play when we were kids: you’d whisper a message to the person next to you, and it goes around the circle. By the time you get to the last person, it bears almost no resemblance to the message you started with. Apply that to the 61,000 people in the Downstream business across the globe, and I knew our strategic aspirations could never penetrate through to the marketplace. The linkages between directives given and actions taken are too problematic.” What made sense to Miller was to fundamentally alter the conversation and unleash the emergent possibilities. Midway through the process, Miller became acquainted with core principles of living systems and adopted them as a framework to provide his organization with a context for renewal.

Miller’s reports in the operating companies were saying, “Centralization will only bog us down.” “They were partly right,” he acknowledges. “These are big companies. Some earn several hundreds of millions a year in net income. But the alternative wasn’t centralization — it was a radical change in the responsiveness of the Downstream business to the dynamics of the marketplace — from top to bottom such that we could come together in appropriate groups, solve problems, and operate in a manner which transcended the old headquarters versus field schism. What initially seemed like a huge conflict has gradually melted away, I believe, because we stopped treating the Downstream business like a machine to be driven and began to regard it as a living system that needed to evolve.”

Miller’s solution was to cut through the organization’s layers and barriers, put senior management in direct contact with the people at the grassroots level, foster strategic initiatives, create a new sense of urgency, and overwhelm the old order. The first wave of initiatives spawned other initiatives. In Malaysia, for example, Miller’s pilot efforts with four initiative teams (called “action labs”) have proliferated to forty. “It worked,” he states, “because the people at the coal face usually know what’s going on. They see the competitive threats and our inadequate response every day. Once you give them the context, they can do a better job of spotting opportunities and stepping up to decisions. In less than two years, we’ve seen astonishing progress in our retail business in some twenty-five countries. This represents around 85 percent of our retail sales volume, and we have now begun to use this approach in our service organizations and lubricant business. Results? By the end of 1997, Shell’s operations in France had regained initiative and achieved double-digit growth and double-digit return on capital. Market share was increasing after years of decline.” Austria went from a probable exit candidate to a highly profitable operation. Overall, Shell gained in brand-share preference throughout Europe and ranked first in share among other major oil companies. By the close of 1998, approximately 10,000 Downstream employees have been involved in this effort with audited results (directly attributed to the program) exceeding a $300 million contribution to Shell’s bottom line.

Self-Organization and Emergent Complexity

Santa Fe Institute’s Stuart Kauffman is a geneticist. His lifetime fascination has been with the ordered process by which a fertilized egg unfolds into a newborn infant and later into an adult. Earlier Nobel Prize-winning work on genetic circuits had shown that every cell contains a number of “regulatory” genes that act as switches to turn one another on and off. Modern computers use sequential instructions, whereas the genetic system exercises most of its instructions simultaneously. For decades, scientists have sought to discover the governing mechanism that causes this simultaneous, nonlinear system to settle down and replicate a species.13

Kauffman built a simple simulation of a genetic system. His array of 100 light bulbs looked like a Las Vegas marquee. Since regulatory genes cause the cells (like bulbs) to turn on or off, Kauffman arranged for his bulbs to do just that, each independently of the other. His hypothesis was that no governing mechanism existed; rather, random and independent behavior would settle into patterns — a view that was far from self-evident. The possible combinations in Kauffman’s arrangement of blinking lights was two (i.e., on and off), multiplied by itself 100 times (i.e., almost one million, trillion, trillion possibilities!).

When Kauffman switched the system on, the result was astonishing. Instead of patterns of infinite variety, the system always settled down within a few minutes to a few more or less orderly states. The implications of Kauffman’s work are far-reaching. Theorists had been searching for the sequence of primordial events that could have produced the first DNA — the building block of life. Kauffman asked instead, “What if life was not just standing around and waiting until DNA happened? What if all those amino acids and sugars and gasses and solar energy were each just doing their thing like the billboard of lights?” If the conditions in primordial soup were right, it wouldn’t take a miracle (like a million decks of cards falling from a balcony and all coming up aces) for DNA to randomly turn up. Rather, the compounds in the soup could have formed a coherent, self-reinforcing web of reactions and these, in turn, generated the more complex patterns of DNA.14

Emergent complexity is driven by a few simple patterns that combine to generate infinite variety. For example, simulations have shown that a three-pronged “crow’s foot” pattern, if combined in various ways, perfectly replicates the foliage patterns of every fern on earth. Similar phenomena hold true in business. John Kao, a specialist in creativity, has observed how one simple creative breakthrough can evoke a cascade of increasing complexity.15 “Simple” inventions such as the wheel, printing press, or transistor lead to “complex” offshoots such as automobiles, cellular phones, electronic publishing, and computing.

The phenomenon of emergence arises from the way simple patterns combine. Mathematics has coined the term “fractals” to describe a set of simple equations that combine to form endless diversity.16 Fractal mathematics has given us valuable insight into how nature creates the shapes we observe. Mountains, rivers, coastline vegetation, lungs, and circulatory systems are fractal, replicating a dominant pattern at several smaller levels of scale. Fractals, in effect, act like genetic algorithms enabling a species to efficiently replicate essential functions.

One consequence of emerging complexity is that you cannot see the end from the beginning. While many can readily acknowledge nature’s propensity to self-organize and generate more complex levels, it is less comforting to put oneself at the mercy of this process with the foreknowledge that we cannot predict the shape that the future will take. Emerging complexity creates not one future but many.

Self-Organization and Emergence at Shell

Building on (1) the principles of complexity, (2) the fractal-like properties of a business model developed by Columbia University’s Larry Seldon,17 and (3) a second fractal-like process, the action labs, Steve Miller and his colleagues at Shell tapped into the intelligence in the trenches and channeled it into a tailored marketplace response.18

Miller states: “We needed a vehicle to give us an energy transfusion and remind us that we could play at a far more competitive level. The properties of self-organization and emergence make intuitive sense to me. The question was how to release them. Seldon’s model gave us a sharp-edged tool to identify customer needs and markets and to develop our value proposition. This, in effect, gave our troops the ‘ammunition’ to shoot with — analytical distinctions to make the business case. Shell has always been a wholesaler. Yet the forecourt of every service station is an artery for commerce that any retailer would envy. Our task was to tap the potential of that real estate, and we needed both the insight and the initiatives of our front-line troops to pull it off. For a company as large as Shell, leadership can’t drive these answers down from the top. We needed to tap into ideas that were out there in the ranks — latent but ready to bear fruit if given encouragement.”

At first glance, Shell’s methods look pedestrian. Miller began bringing six- to eight-person teams from a half-dozen operating companies from around the world into “retailing boot camps.” The first five-day workshop introduced tools for identifying and exploiting market opportunities. It also included a dose of the leadership skills necessary to enroll others back home. Participants returned ready to apply the tools to achieve breakthroughs such as doubling net income in filling stations on the major north-south highways of Malaysia or tripling market share of bottled gas in South Africa. As part of the discipline of the model, every intention (e.g., “to lower fuel delivery costs”) was translated into “key business activities” (or KBAs). As the first group went home, six more teams would rotate in. During the next sixty days, the first group of teams used the analytical tools to sample customers, identify segments, and develop a value proposition. The group would then return to the workshop for a “peer challenge” —tough give-and-take exchanges with other teams. Then it would go back again for another sixty days to perfect a business plan. At the close of the third workshop, each action lab spent three hours in the “fishbowl” with Miller and several of his direct reports, reviewing business plans, while the other teams observed the proceedings. At the close of each session, plans were approved, rejected, amended. Financial commitments were made in exchange for promised results. (The latter were incorporated in the country’s operating goals for the year.) Then the teams went back to the field for another sixty days to put their ideas into action and returned for a follow-up session.

“Week after week, team after team,” continues Miller, “my six direct reports and I and our internal coaches reached out and worked directly with a diverse cross-section of customers, dealers, shop stewards, and young and mid-level professionals. And it worked. Operating company CEOs, historically leery of any ‘help’ from headquarters, saw their people return energized and armed with solid plans to beat the competition. The grassroots employees who participated in the program got to touch and feel the new Shell — a far more informal, give-and-take culture. The conversation down in the ranks of the organization began to change. Guerrilla leaders, historically resigned to Shell’s conventional way of doing things, stepped forward to champion ingenious marketplace innovations (such as the Coca-Cola Challenge in Malaysia — a free Coke to any service-station customer who is not offered the full menu of forecourt services. It sounds trivial, but it increased volume by 15 percent). Many, if not most, of the ideas come from the lower ranks of our company who are in direct contact with the customer. Best of all, we learned together. I can’t overstate how infectious the optimism and energy of these committed employees was for the many managers above them. In a curious way, these front-line employees taught us to believe in ourselves again.”

As executives move up in organizations, they become removed from the work that goes on in the fields. Directives from the top become increasingly abstract as executives tend to rely on mechanical cause-and-effect linkages to drive the business: strategic guidelines, head-count controls, operational expense targets, pay-for-performance incentives, and so forth. These are the tie rods and pistons of “social engineering” — the old model of change. Complexity theory does not discard these useful devices but it starts from a different place. The living-systems approach begins with a focus on the intelligence in the nodes. It seeks to ferret out what this network sees, what stresses it is undergoing, and what is needed to unleash its potential. Other support elements (e.g., controls and rewards) are orchestrated to draw on this potential rather than to drive down solutions from above.

Miller was pioneering a very different model from what had always prevailed at Shell. His “design for emergence” generated hundreds of informal connections between headquarters and the field, resembling the parallel networks of the nervous system to the brain. It contrasted with the historical model of mechanical linkages analogous to those that transfer the energy from the engine in a car through a drive train to the tires that perform the “work.”

Edge of Chaos

Nothing novel can emerge from systems with high degrees of order and stability — for example, crystals, incestuous communities, or regulated industries. On the other hand, complete chaotic systems, such as stampedes, riots, rage, or the early years of the French Revolution, are too formless to coalesce. Generative complexity takes place in the boundary between rigidity and randomness.

Historically,19 science viewed “change” as moving from one equilibrium state (water) to another (ice). Newtonian understandings could not cope with the random, near-chaotic messiness of the actual transition itself. Ecologists and economists similarly favored equilibrium conditions because neither observation nor modeling techniques could handle transition states. The relatively inexpensive computational power of modern computers has changed all that. Nonequilibrium and nonlinear simulations are now possible. These developments, along with the study of complexity, have enabled us to better understand the dynamics of “messiness.”

Phase transitions occur in the realm near chaos where a relatively small and isolated variation can produce huge effects. Consider the example of lasers: while only a complex system and not an adaptive one, the infusion of energy into plasma excites a jumble of photons. The more the energy, the more jumbled they become. Still more and the seething mass is transformed into the coherent light of a laser beam. What drives this transition, and how can we orchestrate it? Two determinants — (1) a precise tension between amplifying and damping feedback, and (2) (unique to mankind) the application of mindfulness and intention — are akin to rudder and sail when surfing near the edge of chaos.

Two factors determine the level of excitation in a system. In cybernetics, they are known as amplifying (positive) and damping (negative) feedback.20 Damping feedback operates like a thermostat, which keeps temperatures within boundaries with a thermocouple that continually says “too hot, too cold.” Amplifying feedback happens when a microphone gets too close to a loudspeaker. The signal is amplified until it oscillates to a piercing shriek. Living systems thrive when these mechanisms are in tension.

Getting the tension right is the hard part. Business obituaries abound with examples of one or the other of these feedback systems gone amok. IT&T under Harold Geneen or Sunbeam under “Chainsaw” Al Dunlap thrive briefly under stringent damping controls, then fade away owing to the loss of imagination and creative energy. At the opposite end, Value Jet thrives in an amplifying phase, adds more planes, departures, and staff without corresponding attention to the damping loop (operational controls, safety, reliability, and service standards).

Psychologists tell us that pain can cause us to change, and this is most likely to occur when we recontextualize pain as the means by which significant learning occurs. When the great Austro-American economist Joseph Schumpeter described the essence of free-market economies as “creative destruction,” it could be interpreted as a characterization of the hazards near the edge of chaos. Enduring competitive advantage entails disrupting what has been done in the past and creating a new future.

Hewlett-Packard’s printer business was one of the most successful in its portfolio. Observing a downward spiral of margins as many “me too” printers entered the market, HP reinvented its offering. Today, HP’s printers are the “free razor blade” — the loss leader in a very different strategy. To maintain scale, HP abandoned its high-cost distribution system with a dedicated sales force, opting instead for mass channels, partnering, and outsourcing to lower manufacturing costs. To protect margins, it targeted its forty biggest corporate customers and formed a partnership to deliver global business printing solutions —whether through low-cost, on-premise equipment, or networked technology. States Tim Mannon, president of HP’s printer division: “The biggest single threat to our business today is staying with a previously successful business model one year too long.”21

Shaping the Edge of Chaos at Shell

Shell moved to the edge of chaos with a multi-pronged design that intensified stress on all members of the Shell system.22 First, as noted, Miller and his top team performed major surgery on their calendars and reallocated approximately half their time to teaching and coaching wave after wave of country teams. When the lowest levels of an organization were being trained, coached, and evaluated by those at the very top, it both inspired — and stressed —everyone in the system (including mid-level bosses who were not present). Second, the design, as we have seen, sent teams back to collect real data for three periods of sixty days (interspersed with additional workshop sessions). Pressure to succeed and long hours both during the workshops and back in the country (where these individuals continued to carry their regular duties along with project work) achieved the cultural “unfreezing” effects. Participants were resocialized into a more direct, informal, and less hierarchical way of working.

Miller states: “One of the most important innovations in changing all of us was the fishbowl. The name describes what it is: I and a number of my management team sit in the middle of a room with one action lab in the center with us. The other team members listen from the outer circle. Everyone is watching as the group in the hot seat talks about what they’re going to do and what they need from me and my colleagues to be able to do it. That may not sound revolutionary — but in our culture, it was very unusual for anyone lower in the organization to talk this directly to a managing director and his reports.

“In the fishbowl, the pressure is on to measure up. The truth is, the pressure is on me and my colleagues. The first time we’re not consistent, we’re dead meat. If a team brings in a plan that’s really a bunch of crap, we’ve got to be able to call it a bunch of crap. If we cover for people or praise everyone, what do we say when someone brings in an excellent plan? That kind of straight talk is another big culture change for Shell.

“The whole process creates complete transparency between the people at the coal face and me and my top management team. At the end, these folks go back home and say, ‘I just cut a deal with the managing director and his team to do these things.’ It creates a personal connection, and it changes how we talk with each other and how we work with each other. After that, I can call up those folks anywhere in the world and talk in a very direct way because of this personal connectedness. It has completely changed the dynamics of our operations.”

Disturbing a Living System

An important and distinct property of living systems is the tenuous connection between cause and effect. As most seasoned managers know, the best-laid plans are often perverted through self-interest, misinterpretation, or lack of necessary skills to reach the intended goal.

Consider the war of attrition waged by ranchers and the U.S. Fish and Wildlife Service to “control” the coyote. A cumulative total of $3 billion (in 1997 dollars) has been spent during the past 100 years to underwrite bounty hunters, field a sophisticated array of traps, introduce novel morsels of poisoned bait, and interject genetic technology (to limit fertility of females) — all with the aim of protecting sheep and cattle ranchers from these wily predators. Result? When white men first appeared in significant numbers west of the Mississippi in the early 1800s, coyotes were found in twelve western states and never seen east of the Mississippi. However, as a direct result of the aggressive programs to eliminate the coyote, the modern day coyote is 20 percent larger and significantly smarter than his predecessor. The coyote is now found in forty-nine of the fifty states —including suburbs of New York City and Los Angeles. How could this occur? Human intervention so threatened the coyote’s survival that a significant number fled into Canada where they bred with the larger Canadian wolf. Still later, these visitors migrated south (and further north to Alaska) and, over the decades, bred with (and increased the size of) the U.S. population. The same threats to survival that had driven some coyotes into Canada drove others to adapt to climates as varied as Florida and New Hampshire. Finally, the persistent efforts to trap or hunt or poison the coyote heightened selection pressures. The survivors were extremely streetwise and wary of human contact. Once alerted by a few fatalities among their brethern, coyotes are usually able to sniff out man’s latest stratagem to do them harm.

As the tale of the coyote suggests, living systems are difficult to direct because of these weak cause-and-effect linkages. The best laid efforts by man to intervene in a system, to do it harm, or even to replicate it artificially almost always miss the mark. The strategic intentions of governments in Japan, Taiwan, and Germany to replicate Silicon Valley provide one example. The cause-and-effect formula seemed simple: (1) identify a region with major universities with strong departments in such fields as microelectronics, genetics, and nuclear medicine and having a geography with climate and amenities suitable to attract professionals, and (2) invest to stimulate a self-reinforcing community of interests. But these and many similar efforts have never quite reached a critical mass. The cause-and-effect relationships proved unclear.23 A lot depends on chance. One is wiser to acknowledge the broad possibilities that flow from weak cause-and-effect linkages and the need to consider the second- and third-order effects of any bold intervention one is about to undertake.

Disturbing a Complex System at Shell

In today’s fast-changing environment, Shell’s Steve Miller dismisses the company’s old traditional approach as mechanistic. “Top-down strategies don’t win ballgames,” he states. “Experimentation, rapid learning, and seizing the momentum of success is the better approach.”24

Miller observes: “We need a different definition of strategy and a different approach to generating it. In the past, strategy was the exclusive domain of top management. Today, if you’re going to have a successful company, you have to recognize that the top can’t possibly have all the answers. The leaders provide the vision and are the context setters. But the actual solutions about how best to meet the challenges of the moment, those thousands of strategic challenges encountered every day, have to be made by the people closest to the action — the people at the coal face.

“Change your approach to strategy, and you change the way a company runs. The leader becomes a context setter, the designer of a learning experience —not an authority figure with solutions. Once the folks at the grassroots realize they own the problem, they also discover that they can help create and own the answers, and they get after it very quickly, very aggressively, and very creatively, with a lot more ideas than the old-style strategic direction could ever have prescribed from headquarters.

“A program like this is a high-risk proposition, because it goes counter to the way most senior executives spend their time. I spend 50 percent to 60 percent of my time at this, and there is no direct guarantee that what I’m doing is going make something happen down the line. It’s like becoming the helmsman of a big ship when you’ve grown up behind the steering wheel of a car. This approach isn’t about me. It’s about rigorous, well-taught marketing concepts, combined with a strong process design, that enable front-line employees to think like businesspeople. Top executives and front-line employees learn to work together in partnership.

“People want to evaluate this against the old way, which gives you the illusion of ‘making things happen.’ I encountered lots of thinly veiled skepticism: ‘Did your net income change from last quarter because of this change process?’ These challenges create anxiety. The temptation, of course, is to reimpose your directives and controls even though we had an abundance of proof that this would not work. Instead, top executives and lower-level employees learn to work together in partnership. The grassroots approach to strategy development and implementation doesn’t happen overnight. But it does happen. People always want results yesterday. But the process and behavior that drive authentic strategic change aren’t like that.

“There’s another kind of risk to the leaders of a strategic inquiry of this kind — the risk of exposure. You’re working very closely and intensely with all levels of staff, and they get to assess and evaluate you directly. Before, you were remote from them; now, you’re very accessible. If that evaluation comes up negative, you’ve got a big-time problem.

“Finally, the scariest part is letting go. You don’t have the same kind of control that traditional leadership is used to. What you don’t realize until you do it is that you may, in fact, have more controls but in a different fashion. You get more feedback than before, you learn more than before, you know more through your own people about what’s going on in the marketplace and with customers than before. But you still have to let go of the old sense of control.”

Miller’s words testify to his reconciliation with the weak cause-and-effect linkages that exist in a living system. When strategic work is accomplished through a “design for emergence,” it never assumes that a particular input will produce a particular output. It is more akin to the study of subatomic particles in a bubble chamber. The experimenter’s design creates probabilistic occurrences that take place within the domain of focus. Period. Greater precision is neither sought nor possible.