The gap seems to grow ever wider between the strategies required to compete effectively and the capabilities needed to execute those strategies. Organizations designed in—and for—a stable environment can easily grow into unwieldy barriers to performance improvement when the environment becomes turbulent. One CEO described the result as being like "running a slalom race in cement." Most know the feeling.

As the demand for performance ratchets upward, many top managers have tried to lead their companies toward some more responsive form of organization—only to find the route as problematic as the goal was elusive. Their wish—indeed, their need—to organize for better performance was perfectly reasonable. But the path chosen was not. All too often, the critical process of determining which route to follow was silently, but effectively, thrown off course by one of several common fallacies about how best to get from here to there.

How often have you heard, for example, that

..."We ought to be able to ’leapfrog’ to the kind of organization we need"

Once a company’s strategy has been agreed, there is a strong temptation to move quickly to align the organization

Once a company’s strategy has been agreed, there is a strong temptation to move quickly to align the organizational elements necessary to implement it. After all, as everyone knows, structure follows strategy. Why wait? One very important reason: since the organizational implications of these strategies now often extend far beyond mere tweaks at the margin of a business to rather massive overhauls, there is a need to translate them into enough painstaking, nuts-and-bolts detail for them actually to be implemented.

One new strategy, for example, called for a company to develop a much simpler and more responsive customer interface. On the surface the implied marching orders were relatively straightforward: simplify our processes for interacting with customers and make them more responsive to their shifting needs. But what does this really mean in practice? Does it mean consolidating all relevant activities into one organizational unit or, at the extreme, into the charter of one individual? Or does it mean preserving existing functional specialization, but linking the relevant players more tightly through better information and performance measurement systems?

Either way, which activities need to be performed at the customer interface and which can be managed off-line? Moreover, how can the company responsibly leap to key decisions about organization design before these microlevel questions—and many others like them—get answered? How can it even hope to do so without sorting through its managers’ differing agendas for—and biases about—the way the business "ought" to be run? It can’t.

In this particular case, the Vice President of Sales took the strong position that, since Sales represented the primary customer interface, all activities involving customer contact should be consolidated into his organization. Not surprisingly, the Vice Presidents of Marketing, Manufacturing, and Distribution all opposed this view and argued that sufficient coordination could be achieved through other means. Each marshaled compelling arguments and repeatedly referred back to the broad statement of strategy as justification for his position. The debate grew sufficiently emotional that each began to view any counter-arguments as personal attacks.

Even when such emotion does not come to the surface, it usually roils the waters just below. If agreement comes easily, it is often because the discussion remains at such a high level of generality that multiple interpretations of what has been agreed are not only possible, but virtually certain. And that, in turn, virtually guarantees endless skirmishing at lower levels of the organization, out of sight of the CEO, as implementation proceeds.

Not surprisingly, the organizational outcomes that emerge from such skirmishing rarely correspond to the true performance needs of the strategy. Instead, they reflect the balance of power within the organization, as well as the performance-destroying accommodations and compromises needed to satisfy different constituencies.

Any attempt to "leapfrog" from strategy to organization is likely to fall far short of the performance levels required

This is why any attempt to "leapfrog" from strategy to organization is likely to fall far short of the performance levels required. The only way to avoid this outcome is for senior managers to work in careful detail through an intermediate stage that focuses on determining what is operationally required—the new approaches to activities, for example, or the new information and material flows—to achieve those performance levels. Only then can they productively engage in a discussion about suitable organization design. The personal agendas and biases will still be there, of course, but they will be balanced by a much richer understanding of what the strategy really implies at the operational level.

Back to the example cited above. Examined closely, field interviews strongly supported the need to concentrate responsibilities in a single, named individual to whom each customer could look as his or her "champion" within the organization, particularly in terms of the complex logistics of after-sales service. All the skills required to perform this combined champion role, however, were not present in the field salesforce. Two people, therefore, would have to provide the linkage with each customer.

To make this arrangement work, the joint customer champions would have to be tightly linked with each other through both information flows and the performance measurement system. Only in that way could the company ensure that they would act as a team held collectively accountable for a customer’s satisfaction. Now, the real work of design could begin.

..."We need one organizational model for our entire business"

Determining whether a single organizational approach is appropriate is not something that can be done in the abstract

In a well-intentioned effort to promote simplicity and clarity, senior managers often act on the principle that their business requires only a single organizational model. In fact, this principle is so basic that it is often left unstated. It shouldn’t be. Determining whether a single organizational approach is appropriate is not something that can be done in the abstract. The answer hinges on the kind of operational analysis described above.

Consider, for example, a company at the low, rapidly-commoditizing end of the computer business. To keep its lead in product innovation while providing fast and reliable delivery, it had to deploy very different sets of skills. Innovation required the integration of a broad range of capabilities within high-risk projects with very aggressive deadlines. Cost was less important than ensuring that the right products reached the market as quickly as possible. The logistics process, by contrast, was much more predictable and called on a more limited range of capabilities. The key here was to coordinate activities so as to achieve the lowest possible cost and throughput time while ensuring reliability.

The organizational models appropriate for each of these processes were quite different. For innovation, tightly-linked teams cutting across functional boundaries were necessary to bring together the necessary skills. This, naturally, put a premium on the ability to work effectively as part of a team and on co-location. Functional affiliations mattered little; performance measurement systems had to focus on team performance in terms of time to market rather than on the cost of the process itself; and compensation systems had to reward team, as well as individual, contributions.

For logistics activities, teams were less important since much of the necessary coordination could be provided through integrated flows of information. People could, therefore, maintain a much closer affiliation to their functional homes and could be much more dispersed geographically, especially when this physical separation helped lower operating costs. Furthermore, the need to exploit economies of scale implied greater centralization. As a result, performance measurement systems had to focus only on individual performance against tightly-defined parameters of operating cost, speed, and reliability, and compensation systems had to reward only individual, not team, contributions.

Different organizational approaches added complexity. But a single approach would have severely compromised performance objectives

To be sure, the different organizational approaches implied by these requirements introduced some added complexity. But any attempt to impose a single approach across the entire business would have severely compromised its performance objectives.

..."Our front-line employees should determine the organizational models that work best for them"

The kinds of strategies most companies are now adopting demand significant, often radical performance breakthroughs

In the current age of empowerment, this is a perfectly natural instinct, but it is an instinct that needs to be resisted, at least in the early stages of organization design. The kinds of strategies most companies are now adopting demand significant, often radical performance breakthroughs, which in turn usually require significant departures from traditional organizational approaches. Front-line employees rarely have the integrated perspective or the breadth of vision to understand the practical implications of such requirements. For this reason, organization design must, if only at the outset, be driven top-down.

Another computer company, this one facing a need to reduce its sales and marketing costs while increasing responsiveness and levels of customer satisfaction, drew on what it had learned from implementing a very successful total quality management program: it relied on a taskforce of front-

line employees to develop organizational solutions. They approached the challenge with enthusiasm but were unable to devise an answer that delivered the necessary level of performance. Two obstacles stood in their way. First, they were good at coming up with suggestions for marginal improvements, but not for truly radical approaches—not least because such approaches might threaten their own job security. And second, the cross-functional changes they proposed ran up against stiff middle-management resistance.

What had been the salesforce would become, in effect, a group of consultants who worked with customers

As a result, senior managers had to shift course, take a direct hand in the process, and reconstitute the taskforce to include some of their own number as well as some of the more creative and aggressive front-line members of the original taskforce. Driven top-down, the effort now produced a genuinely radical solution: eliminate the salesforce entirely and provide customers with direct access to information and order entry systems. What had been the salesforce would become, in effect, a group of consultants who worked with customers to design solutions to their business needs.

..."We need to understand how everyone will respond to these organizational changes before rolling them out"

Senior managers are often very cautious about implementing a new organization design before reaching a high level of comfort with their judg-

ment of how people will react. Again, this is perfectly understandable given the potential for any new design to disrupt existing business processes and threaten economic performance. What if the new compensation systems actually motivate different behaviors from the ones intended? What if front-line employees feel they do not have adequate information to make the right decisions? What if they feel they lack the skills to make the right decisions? What if customers don’t like the new approach to customer service?

The vagaries of day-to-day organizational performance are simply too complex to be captured in a blueprint

But no one—no matter how well briefed on a proposed design—can possibly know all this in advance. The vagaries of day-to-day organizational performance are simply too complex to be captured in a blueprint, no matter how detailed. As my colleague Yoshi Yokoyama has put it, life will fill in the spaces.¹ Modifications will have to be made. Personal styles will prompt adjustments. New ideas will surface in the course of implementation. These are not signs of failure. They are an inescapable part of successful learning.

In one recent case, senior managers invested almost two years in developing a new organization design before rolling it out. Only after yet another year had passed did they discover that customers did not value the new skills that had been developed in the salesforce. Valuable time and even more valuable market share had been lost. So had the morale of the now whipsawed salesforce. Trying to microplan in advance all the fine details in order to reduce risk, management had actually increased risk by postponing the day the design could be tested in live operation.

If managers think of the implementation of a new design not as a one-time event, but as a series of "waves," they can reduce the risk of drifting off-course

It doesn’t have to be this way. If, for example, managers think of the implementation of a new design not as a one-time event, but as a series of "waves" or "releases," they can reduce the risk of drifting significantly off-course.² They can also create a reliable foundation—through, say, near-term changes in performance measurement and compensation systems, which rapidly produce significant changes in behavior—on which to build longer-term changes in areas like skill-building.

In much the same way, they can use "pilot" implementations of the key elements of a new design. Indeed, it is often possible to carve out a discrete piece of the business—a geographic area or product category, for example—in which to do such testing. This is analogous to the kind of rapid prototyping that is now revolutionizing best practice in software development.

..."We can’t afford to wait; we need to implement the full design now"

As market pressures mount and performance deteriorates, managers quite reasonably feel they do not have the luxury of a lot of time to test and refine a new design. They need it—and its promise of performance improvement—now. Delay is intolerable.

The problem, of course, is that "crash" programs all too often become "crash and burn" programs. In true turnaround situations, the urgency of the need may well justify the higher risk of breakneck speed: in the absence of immediate action, the doors close. But such situations are rare. In most cases, management has at least twelve to eighteen months to achieve significant performance improvement.

An insurance company seeking to implement a new organizational approach to claims processing opted for a crash program and rapid roll-out. The new design required a single person to take on activities that in the past had been performed by a number of people. No one, however, had anticipated the level of resistance shown by the people who were expected to take on this expanded range of tasks. Nor had anyone really understood the changes required of the information systems that would support them. And even if they had, the crash environment meant there was no time to do anything about it.

The best approach is usually to strive for the middle ground, balancing urgency against risk

Once again, the best approach is usually to strive for the middle ground, balancing the urgency for performance improvement against the importance of managing risk. As described above, structuring the design process as a series of waves or releases and testing it through a series of pilots can pinpoint any supporting action needed to minimize resistance or provide essential enabling tools.

..."We’ll define the new structure and let the staff work out the other details"

When senior managers think about organization design, they tend to focus heavily on issues of structure: who reports to whom? There are many reasons for this—chief among them, structure is relatively easy to describe and to change. Drawing new organization charts is a lot simpler than trying to define, say, the new kinds of information that the front line will require.

Structure, though a powerful lever, is much more of a sledgehammer than a scalpel. Relying on it, alone, can be dangerous

But structure, though a powerful organizational lever, is much more of a sledgehammer than a scalpel. Relying on it, alone, can be dangerous. So can delegating the responsibility for planning the design’s other dimensions. Organizations succeed or fail based on the degree of alignment among these dimensions—style, shared values, systems, and the like—and only senior management possesses the integrated perspective needed to bridge them and make the inevitable tradeoffs.

The top group at an industrial products company approved a radical new design involving tightly-integrated teams reporting to process "owners" rather than functional managers. Confident that this was the right organizational model for the business, they delegated the task of defining and implementing the remaining organizational details to several low-level taskforces. The IT taskforce, however, quickly found that the information required by these new teams could not be delivered without a fundamental redesign of the company’s "legacy" IT systems. Since such a redesign would take three to five years, the taskforce presented senior management with an unpalatable dilemma: establish the teams without the necessary information or delay implementation of the new design for an unacceptable period of time.

The taskforce, though well-intentioned, lacked the breadth of vision required to make the necessary tradeoffs

Ultimately, the CEO intervened, reconvened the top-level group that had developed the original design, and personally led its efforts to find a way around this dilemma. As it turned out, by focusing on the information that was truly essential for performance improvement, the group was able to define a phased approach to systems redesign that could deliver the necessary information to the teams in just twelve months. The IT taskforce, though well-intentioned, lacked the breadth of vision required to make the tradeoff between information availability and acceptable lead-times. The lesson is simple: top-level participation in a new design should continue throughout the implementation process, and senior executives should take explicit responsibility both for defining all its major dimensions and for making critical tradeoff decisions.

..."We won’t know if the new design will work until all the pieces are in place two to three years from now"

The greatest risk in not establishing performance-focused milestones is that the design initiative will lose momentum at an early stage

Although the full performance impact of a new organization design will often not be visible for several years, it is possible—indeed, essential—that senior managers establish a series of performance-focused milestones that can measure progress toward the goal, not just the level of activity along the way. Did the taskforce complete the organization design? Did the pilot occur in a timely fashion? Were new IT needs identified? One can answer "yes" to all these questions without knowing if there has been tangible progress in improving business performance.

Perhaps the greatest risk in not establishing aggressive, performance-focused milestones is that the design initiative will lose momentum at an early stage. A retail company that launched a design effort with long-term performance objectives, but no interim performance milestones, found within twelve months that the taskforce had effectively stopped meeting and its blueprints for change had not advanced beyond where they had been six months before. A post-mortem revealed that the taskforce had found a high level of resistance to change within the organization, as well as a reluctance to wait patiently for performance improvement. As pressures mounted, senior management attention had quickly focused on a variety of short-term initiatives to improve performance. It became more and more difficult for them to justify the resources and time consumed by the design effort. As the CEO observed, "The taskforce just wasn’t relevant any more."

Performance-focused milestones can help. Properly defined, they can ensure that the performance needs of the business are addressed in a timely fashion and that design efforts are structured and sequenced to deliver as much performance as possible as quickly as possible. They can also help to give the effort credibility within the organization: look, the disruption and uncertainty are really worth it because the effort is delivering concrete performance benefits.

Basic guidelines

Making explicit these common fallacies offers some guidelines for CEOs seeking to capture the performance potential of their new strategies:

Move from strategy to organization in stages. Major organizational change is justified only as an enabler or facilitator of operational changes that are, in turn, justified only by the performance requirements of strategy.

Be open to the possibility of multiple organizational solutions. If the key operating processes of a business are different, they may require different organization designs.

Establish, at least at the outset, a top-down organization design process. True breakthroughs in business-unit or corporate performance rarely come from bottom-up, incremental changes in design. Fundamental change requires the integrated perspective that only senior management can provide. At the same time, of course, the process can—and should—involve front-line personnel, from whom alone comes needed insight on how things really get done.

Use phasing and pilots to establish an appropriate balance between urgency and risk. Avoid the twin temptations of waiting until all the details have been mastered before roll-out and of launching a crash program that increases the risk that key organizational issues will not receive adequate, timely attention.

Pay attention to all dimensions of organization design, not just structure. The success of a new design depends on the alignment of all its dimensions. Senior managers cannot afford to delegate the critical choices about any of them.

Establish, at the outset, aggressive performance-focused milestones. Hold the design taskforce accountable for near-term performance improvement as well as long-term performance targets. Without near-term improvements, the longer-term targets will be at risk.

There is, of course, no panacea here, no approach or design that works for all businesses in all circumstances. Nor is there any shortcut to the difficult task of working through, first, the operational and, then, the organizational implications of each company’s strategy. The companies that navigate this road well will achieve performance levels that few, if any, of their competitors can match.