Featured Guest Speaker
Stephan H. Haeckel
Enterprise Strategy - Past and
Editor's Note: Steve Heackel's book, Adaptive Enterprise: Creating and Leading Sense-And-Respond Organizations, is a must read for anyone who wants to understand enterprise agility. I'm highly biased in my recommendation -- though I've yet to meet Steve personally, reading his book was like reading my own beliefs on agile enterprise. Low key but powerful in its message and conclusions, he states his case and gets quickly to the point. Heackel is sensitive to the realities of large complex organizations operating under uncertainty, and walks a solid path where others have told half a story or appealed to metaphor and idealistic theory. Chapter three, reprinted below, offers valuable insights on the roles of trust, empowerment, and emergence in corporate strategy, and a powerful conclusion that the future of strategy is not at all like we've known it in the past, but rather "a design for an adaptive structure."
Paul Gaddis, a professor of corporate strategy at the University of Texas at Dallas, has described the history of strategy as an oscillation between "proactive purposefulness" and "reactive powerlessness." The make-and-sell era can be characterized as an age of proactive purposefulness. Corporate leaders prescribed broad courses of action in advance of events, confident that they could control their companies destinies. These leaders often believed that they could not only predict the future but create it as well, by influencing their customers behaviors. A comment frequently made by T. Vincent Learson, president of IBM in the late 1960s and chairman in the early 1970s, encapsulates this view: "I dont worry very much about making the right decision. I worry a lot about making the decision right." IBMs success during Learsons executive career appeared to justify this approach.
Learsons leadership exemplifies a familiar, traditional view of strategy: A group of corporate "generals" survey the territory they want to capture or protect (their customers), study the capabilities of the enemy (their competitors), develop a battle plan to achieve their objectives, and, finally, direct the troops who carry out the plan. The origins of the word strategy support this image. Strate-gos refers to the responsibility of an army commander to decide what the troops should do and then make sure they do it. This image of top-down leadership and unquestioning obedience closely mirrors the hierarchical leadership model of classic make-and-sell firms and the kind of strategic planning they used. Bruce Harreld, IBMs senior vice president for strategy, calls this the "we know, you do" theory of management.
To succeed, traditional strategic plans must meet two basic conditions. First, critical opportunities and threats must be sufficiently predictable. Every strategic plan includes a list of the assumptions on which the strategys effectiveness depends. Westpac executives, for example, were unwilling to invest in the assumptions underlying 1985 product forecasts, but they were willing, ultimately, to invest $100 million in an assumption of generic unpredictability about product demand. For most other large companies, too, rapid technological innovation and new sources of competition eroded or eliminated the kind of control that Learson describes. The sheer complexity of managing such large enterprises further undermined that control.
Second, successful strategic planning requires an effective mechanism for translating plans into action. In large make-and-sell firms, the planning process typically began with senior management setting the companys financial goals. These goals and the plans developed for achieving them were built on assumptions about future customer demand and competitors offerings, along with expectations about economic growth, technological change, and geopolitics. In many firms, a large central planning staff was responsible for the management system, whose purpose was to ensure machine-like efficiency in forecasting, developing, making, selling, and distributing the kinds and quantities of products called for by the plan. Most staffs did this by overseeing internal dependenciesthe actions, product components, or information owed by each unit to the others. The staff tracked these dependencies, following up on the fulfillment of each units commitments to carry out their part of the plan. Again, success depended on a sufficiently predictable business environment. When unanticipated events made it impossible for a unit to meet one or more of its commitments, the planning staff managed an escalation process to reallocate resources, adjust schedules or revise objectives to resolve the problem with the least possible disruption to the annual plan. Although managers talked about "the constants of change" as early as the 1960s, the central staffs of many large companies successfully managed the relatively few unwelcome surprises that came their way in those years.
Beginning in the early 1970s, however, many firms experienced a rate and discontinuity of change great enough to overwhelm the ability of their ever-growing staffs to coordinate the many interdependent adjustments needed to modify the plan as its assumptions were proved wrong. In 1973, for example, IBMs planning bureaucracy had grown to three thousand people and its "annual" planning process approached an eighteen-month cycle time. At IBM, as elsewhere, the centralized staff that had once served company needs so well, was now creating more delay and problems than it did value. The Economist, reporting on a 1997 finding that business executives expect strategy to be the most pressing issue they would face over the next five years, connected the demise of the central planning staff with the increasing speed of change.
One after another, IBM, Kodak, GM, DuPont, GE, Xerox, and other industrial giants dismantled their central staffs. Yet, for all their problems, these staffs had provided the essential control mechanism in the command-and-control management system. Their elimination led to a coordination vacuum. By the mid-1980s, consequently, a number of large firms, including American Express, announced that they were abandoning attempts to achieve synergy, the additional value derived from having two or more units work cooperatively as parts of a system. The need for coordination has not gone away, however, and later we will discuss a new coordinating mechanism appropriate to a sense-and-respond world. But most policy executives didnt focus on this issue when they abandoned central planning.
Managements willingness to eliminate the planning staff signaled, above all, that the very idea of a strategic plan was losing viability. In truth, the track record of strategic plans over the past thirty years has been dismal. Russell Ackoff estimates that no more than 2 percent are ever successfully executed.
In short, radical uncertainty about the future washed away the foundation on which traditional strategic planning rested. Pervasive uncertainty about customer preferences calls into question the usefulness of both strategic plans and the business structures designed to carry them out. If you cannot know what your customers will want or your competitors will offer next yearor even who your customers or competitors will beyou cannot develop an effective plan for achieving targeted levels of sales and profits.
Some companies have found scenario thinking useful in recent years, because it does not depend on predictions about the future. Scenario development helps identify the most relevant areas of uncertainty and the future events that might signal an unfolding in reality of a given scenario or combination of scenarios. Such exercises help organizations prepare themselves for a wider range of possibilities, encouraging them to become more effective at sensing "whats going on out there." Arie de Geus, discussing scenario planning at Royal Dutch/Shell, has said that "scenarios act as a signal-to-noise filter," sensitizing observers to what may turn out to be significant environmental signals. Scenarios can help organizations learn what environmental signals to track and how better to interpret them.
De Geus points out that the scenario thinking process involves preparing multiple scenarios, rather than relying on a single "most likely" description of the future. Executives think through in advance actions that would make sense in each scenario and identify strategies that would be appropriate in all of them. With this mental preparation, management can sometimes reduce its reaction time to events it did not predict.
From Control to Empowerment
When the central planning era ended, management began decentralizing decision-making: The trend toward empowerment had begun. Leaders hoped that giving more autonomy to those closest to the customer and to the companys daily operations would improve responsiveness. Rather than draw up a detailed battle plan, corporate generals authorized lieutenants to take the actions they deemed appropriate to their situations. The idea of high-performance teams took root. But what were these teams to do? How were their efforts to be coordinated? Lacking an overarching plan, and with only mission/vision/value statements to navigate by, the newly empowered people were left to decide for themselves what to do. As a result, decision-making became more local and less consistent. Management would intervene occasionally, settling conflicts and establishing priorities for the various units. But with every change in the business climate, new conflicts arose and old priorities became obsolete, leading to yet more sporadic interventions. Situational problem-solving dominated. But without a consistent context, these piecemeal solutions failed to add up to coherent behavior at the enterprise level.
Strategy all but disappeared from the landscape. Financial plans and communications about priorities replaced strategic plans. Even if a strategic plan were created, at most companies no mechanism would exist for turning it into action. A bemused observation by Robert Hendry, CEO at Saab, summarizes very well the state of strategy: "These days, strategic thinking is all about tactics." Hendry finds this state of affairs uncomfortable, but for others, it affirms that top-down planning never was a very good idea. Given the right chance, they think, strategy will emerge.
Can strategy arise from the spontaneously autonomous decisions and actions of empowered groups within an organization? Some theorists would answer "yes," but distinguish emergent strategies from the "reactive powerlessness" described by Gaddis. They maintain that an enterprise has a latent capacity to behave as a coherent systemeven when its leaders cant determine in advance what actions are needed to achieve that coherency. The concept of emergent strategy implies reactive rather than proactive behavior. Its supporters believe that complexity and uncertainty render organizational leadership powerless to chart an effective course of action.
A few "emergent behaviorists" believe that scientific theory about complex adaptive systems can be directly applied to human organizations. They argue that coherent behavior can evolve without any intentional shaping or even direction-setting, as hurricanes sometimes emerge from tropical depressions formed by the autonomous actions of sea currents and shifting winds. Gaddis associates this view with the concept of a "super-organization that can continuously develop, increment by increment, its own strategic direction to a prosperous (undefined) future." Long term direction is not needed from the top of the organizationonly day-to-day support of the on-going efforts of the superior organization.
Less doctrinaire believers in emergent strategy hold that it requires conscious direction, but only in the limited form of statements about common goals or values. In this view, once leaders formulate and articulate the signposts, the autonomous behaviors of groups within the organization will create collective behavior that is both coherent and responsive to the changing environment. Still other theorists maintain that direction is not enough: small numbers of decision rules must exist that apply throughout the organization.
These different viewpoints about emergent strategy stem from the same basic premise: Complex human systems, like some complex physical systems, naturally tend toward order rather than randomness. Efforts to direct their behaviorbeyond, perhaps, establishing a few basic goals and rulesis not only unnecessary but probably destructive, because anything more would impose an unnatural rigidity on an organic system that "wants" to thrive in its changing environment. In other words, a system can instinctively know how to act purposefully and strategically.
If the theory of emergent strategy actually works, it will go a long way toward solving the problem of strategy in discontinuous environments. Emergent strategies will predominate in turbulent times and planned strategies in periods of relative stability. If the actions of empowered groups, free to respond to the environmental changes they sense, naturally cohere into useful, purposeful behavior, a benignly neglectful strategy is the best strategy.
The logic supporting this hope, unfortunately, is fundamentally flawed. Human organizations are not like flocks of birds or hurricanes or wolf packs. Human organizations are social systems; that is, not only the systems but the elements of the system can make decisions about the systems purpose and rules. This is precisely why coherency is an issue. People can and often do make conscious choices that are inconsistent with the systems purpose. Humans can think outside the system; they can change the rules and even the function of the system of which they are part. They can also decide whether or not to cooperate with others in the system. At a minimum, research needs to be done to understand the behavior and principles of complex, adaptive human systems whose elements are capable of conscious decision-making. Pending a theory, we hear more about poor execution, lack of synergy, and people working at cross-purposes than we do about successful emergent strategies.
Nevertheless, successful emergent strategic behavior has been demonstrated by a few large organizations. The following brief examples provide insight into what emergent strategy means and how adaptive systems work. The organizations described, however atypical they may be, can contribute to a better understanding of sense-and-respond organizations. They also suggest why most corporate leaders would do well to remain cautious about the promise of applying complexity theory to achieve emergent purpose and strategy in human organizations.
One very impressive case of large-scale emergent strategy is most definitely not an Information Age exemplar. For seven hundred years, the textile merchants of Prato, a city about twenty-five miles northwest of Florence, have been operating as a virtual corporation, in which a large number of small groups, acting autonomously, cohere into a purposeful, productive, and very adaptive system.
In Prato, approximately 8,500 small firms, half employing fewer than ten people, form a dynamic network producing custom fabrics for the middle and upper end of the fashion industry. The single largest agglomeration of textile manufacturing facilities in Europe, with annual revenues of more than $5 billion, Prato merchants work together to respond to specific customer requests. Most of these go first to independent master brokers, called impannatore, who interact with customers and dispatch work to individual firms. The impannatore has no authority and exercises no control over these firms, and the whole system works without formal contracts. Kuldeep Kumar, a professor at Erasmus University, describes this coordination as "achieved by horizontal communication between adjacent parts [of] the [dynamic value] chain. It is very common for the impannatore to communicate only with the first and last actor in the chain." Because there are only about half a dozen types of specialty manufacturing operators, sequencing the steps for a given order is not a complex task. Individual business units need not be connected to a very large number of other units to respond with the product variety and production capacity any given order requires. In theory, every order could give rise to a different chain of firms, each chain ceasing to exist once the order is filled. Surprisingly, but aptly, Kumar likens Prato to a slime mold, a collection of single-cell organisms that aggregate to form a mold that can move to search for food. For slime molds, mobility is the emergent strategy and finding food the emergent purpose. When the mold locates food, it disaggregates again.
Prato collectively exhibits successful organizational performance in the face of unpredictable demand. No part of this virtual enterprise attempts to forecast aggregate demand or even predict the next customer request. Like the adaptive systems of complexity theory, its relatively simple local behaviors create a systematic ability to respond to complex, changing demands from the environment. Simple, tacit, and unambiguous rules govern behavior within the network, rules very analogous to those that govern the behavior of birds in flocks. Units in the Prato network must deal only with other units in the network, do what they say they will, stick to what they know how to do, and incrementally improve the way they do it. The penalty for not adhering to the rules is shunning, resulting in economicand probably socialdeath for the offending unit. These rules do not constitute a strategy in any usual sense of the word, but taken together, the behavior of Pratos textile merchants is evidently both strategic and adaptive: It has successfully met changing demands for fabric for seven hundred years. So, it seems, emergent strategy can work, and work on a large scale.
An element clearly essential to Pratos success is a high level of trust, trust that has been centuries in the making. Because of this trust, the Prato system defies the economic theory that networks have lower production costs but higher transaction costs than do vertically integrated organizations. The merchants of Prato enjoy both low production and low transactions costs. Like information, trust reduces uncertainty. In fact, trust eliminates the need for certain kinds of informationcontracts and due diligence reports, for example. The member firms in the Prato network trust each others competence, honesty, and mutual interest in preserving the network. Severe penalties follow rule violations, because the system survives on mutual confidence that other members will observe the established conventions.
The importance of trust is one of the major elements we should take from this example. The overall effectiveness of any system depends on the interaction of its elements, and mutual trust makes possible the collaboration of humans to fulfill common purposes, without bogging down over complex, rigid contracts or in lengthy negotiations. The Prato system also illustrates the modularity on which sense-and-respond organizations depend.
The strategic behavior of taxi fleetsfor instance, clustering at hotels and airports in the morning and eveningresults not from a master plan scheduling particular drivers to be at specific places at appointed times. It emerges from the individual behaviors of fairly autonomous individual drivers. Some writers assert that this coordination can be produced by having all taxi drivers follow two simple rules: Do not get into any line that has more than five cabs in it, and bid on every call the dispatcher broadcasts (with the closest cab getting the fare). Personally, I have not met a cab driver who admits abiding by these rules or even being aware of them. Nevertheless, distribution of taxi cabs over the course of a day does tend to be predictable and strategic, in that cabs usually show up where they are most needed, without being told where to go by any planning staff. The cumulative result of thousands of decisions made by empowered individual cab driversmost of them in competition with each otheris a purposeful system, even though no one in the system has consciously designed or managed the interactions between taxis.
Taxi fleets thus provide another example of emergent strategy. Individual cab drivers apparently (and perhaps unconsciously) follow emergent rules that trigger behaviors from which a successful strategy emerges. As in the case of Prato, the system works much better to accomplish its ends than could a centrally organized one. No top-down plan could produce a system of remotely comparable efficiency and flexibility for a city the size of Boston, for example. Further, a centrally controlled system would almost certainly fail to be as robust. In Boston, more than ten percent of the taxi fleet can be out of service without noticeably degrading the systems ability to handle a normal days demand.
The number of large organizations intentionally designed to foster emergent strategic behavior remains quite small. Probably the most famous example of one is Visa. Dee Hock, founder and first CEO of Visa International, claims that "Like the body, the brain, the biosphere, [Visa] is largely self organizing." Hock describes Visa as a system that works effectively as a whole but leaves members free to make their own decisions and to develop their individual strategic initiatives.
When Hock left Visa in 1984, the venture was clearly successful, although he estimated that he had achieved no more than 25 percent of what he had wanted to do. By consciously establishing a system-level purpose and a prescribed mechanism for coordinating certain behaviors of otherwise autonomous units, Hoch created an important variation on the Prato and taxi-fleet models of emergent behavior. The crucial issue of how system-level purpose and coordination can co-exist and connect with lower-level autonomy will be discussed later in the chapters on organizational governance.
The Limitations of Emergent Strategy
The examples above demonstrate that emergent strategic behavior does exist in the real world. Visa is clearly a success story. Both taxi fleets and the Prato merchants show how individual behaviors can cohere into a system that effectively handles the demands of unpredictability about what the next customer will want. It therefore appears that, in some situations, emergent behavior governed by a clear goal and few rules can succeed better in dealing with uncertainty than can centrally-planned systems. It is difficult to even imagine a plan that specifies all the movements of all the taxicabs in Boston.
All three of these systems, however, are inherently simple. They are all flat, slime-mold organizations, to use Kumars term, or swarms, in the vernacular of complexity theory. In each system, the only vertical communications required are those with the dispatcher or coordinating mechanism. None of these networks consists of multiple layers of subsystems, thus limiting coordination problems. Also, the constituents of the system perform a very limited number of different functions: only one, in the case of the taxi fleet, and only a few in the case of Prato (such as cutting, dyeing, and pattern design) and Visa (for example, credit checks, issuing cards, clearance).
Not surprisingly, most large corporations remain skeptical about the applicability of these examples to their own organizations. General Electric, AT&T, Royal Dutch Shell, Westpac, and General Motors have structures much more complicated than do Prato, taxi fleets, and Visa. They consist of multiple layers of many subsystems performing a wide range of tasks to generate a host of products and services. Logic suggests that, given the size and complexity of these organizations and the multiplicity of behaviors, goals, and conditions that characterize them, the likelihood that coherent strategic behavior will spontaneously emerge from their millions of individual decisions must be near zero. In this particular instance, logic and reality coincide. Synergy no longer seems an attainable goal in most large organizations. The less lofty aim of simply doing appropriate work consistently well has also come to seem beyond the reach of well-intentioned employees, as well as of executives who feel frustrated by poor operational execution and confusion over priorities.
Henry Mintzberg, an influential authority on business strategy, argues that strategy emerges in the sense that it can only be described after the fact by inspecting the decision chain that led to a particular result. But hindsight does not explain how the thousands of dynamic decision chains that characterize large organizations managed to align themselves with one another, especially if each chain was comprised of a separate, autonomous unit. Even if it were possible to analyze each chain to discover its inherent strategy, how could those many micro-strategies possibly add up to an effective organizational strategy?
The Future of Strategy
We have looked at two closely related themes. First, we examined the idea that strategy, whether planned or emergent, is evidenced by behavior that meets environmental challenges to achieve an organizational purpose. Second, we discussed the importance, especially in large organizations, of coordinating behaviors to ensure that the parts of a system work together to produce coherent behavior.
Strategy-as-plan has ceased to be a viable option for most large organizations because creating an overall plan of organizational behavior requires the ability to predict or control the future, an ability that, according to our premise, no longer exists. Coordinating behaviors from the top down is also no longer possible: There is no mechanism for doing so when the required responses are not predictable. Strategy-as-emergent-behavior does not appear to be a promising alternative for complex organizations that require unplanned, yet coordinated, behavior by many groups to produce a wide range of potential outcomes.
If strategy is neither a purposeful plan to make and sell offers nor a naturally emerging phenomenon, what is it? Westpacs strategy of the mid-1980s answers the question for large firms operating in highly unpredictable environments: Strategy is a design for an adaptive structure. I argue that this is the only strategy that makes sense under such conditions. The future of strategy, therefore, lies in expressing it as a design. The next chapter, describing the sense-and-respond alternative to both make-and-sell and emergent strategies, explores the idea of strategy as structure. It also explains how a strategic organizational structure can solve part of the coordination problem, though only part. Later chapters on governance deal more fully with the coordination issue.
Steve Haeckel, Haeckel@us.ibm.com,
|Would you like to offer some thoughts or add to the dialog? Responses of general interest may be posted below. Send your comment to . IMPORTANT: Make sure the subject line of your message contains: Comment on Guest Speaker 11/99.|
|========= Reply =========================
From: email@example.com (Paul T. Kidd), Date: Fri, 12 Nov 1999
Enterprise Strategy - Past and Future touches upon an important issue that lies at the heart of agility, although most people seem to have missed the point completely or were not brave enough to voice it. Modern design theory tells us that there are a number of ways one can design (a strategy, a product or even an enterprise). The first is the stage-wise model. It is dominant in our culture and can be found in software engineering, enterprise design as well as strategy design. But there are also other models - incremental, adaptive and spiral. All four models can be applied to strategy development. Which model to use depends on the situation - characterised by the level of uncertainty, design complexity, constraints, and risks. The trick is to know which approach to apply and when to shift models. Design theory also tells us that problem finding and problem solving are intertwined activities. A grave weakness in all design activities is the separation of the two - it can be seen in strategy development, software engineering and IT systems implementation. These kinds of issues I wrote about in my book on Agile Manufacturing (Agile Manufacturing: Forging New Frontiers). I am glad to see that at last these matters are being addressed in other books.
Five years on from the publication of my book, my conclusions about the strategy development process required in modern business environments have not changed. However, it is now clear that to make different models work several factors must be addressed:
1. The structure and processes of the enterprise need to be realigned to support these strategy developments models - this in itself may require a strategy (to become adaptive or agile) as well as novel high level business processes.
2. Strategy needs to become everyone's business. Operationalising this is a key concept in achieving a sense and respond enterprise - it's a different form of empowerment to that which we currently see discussed and implemented.
3. Above all else, these are matters of culture and paradigm. Many of the issues surrounding the area are soft and intangible. As a result many will not fully realise the capability. Here therefore, lies a source of sustainable competitive advantage for those few who can.
From what I have see so far your book looks like a must-read. I will be ordering a copy without delay. I look forward to reading it. Paul T. Kidd
========= Reply =========================
2. Strategy needs to become everyone's business. Operationalising this is a key concept in achieving a sense and respond enterprise.
3. "Above all else, these are matters of culture and paradigm. Many of the issues surrounding the area are soft and intangible." [Paul T. Kidd]. Operationalising of the "soft and intangible" models, strategies and processes require effective communication supported by relevant decomposition and notational technology. Strategies and processes are the partitions | views of the enterprise context. Enterprise is a partition | view of the environment context. A map of associations represents context. Association (a.k.a. elementary context) is a non-atomic construct relating objects, messages, operations and stages. There are many types of associations including transitions. The strategies are clusters of transitions. For non-trivial strategies and processes, a workstation is needed to provide the environment for the associations' development, processing, browsing, searching and presenting
Strategy worth cannot be assessed without the context. The Merchants of Prato case illustrates the importance of context decomposition and integration of transitions into local and 'global' strategies. The strategic role of transient context communication is demonstrated by The Emergent Strategy of Taxi Fleets. Expressing strategy as an adaptive design will allow one to adjust it to a highly unpredictable context. Non-trivial design will require use of relevant decomposition and notational technology.
Thanks for the chapter with strategic knowledge and insightful comments.
========= Reply =========================
By "Interactive marketing" I mean "An exchange between a firm and a customer (or prospective customer) that changes the state of at least one member of the exchange." That's a generic definition. The exchange of most interest, to me, is one where the customer provides information about some value being sought, and the firm provides information about its ability to deliver that value. Naturally, I include tacit as well as explicit expressions in the exchange.
Thank for your interest. Steve
Send mail to
with questions or comments about this web site.