“What activities belong together and what activities belong apart?” A searching analysis is needed that groups activities by the kind of contribution they make. There are four major groups of activities, if distinguished by their contribution.
First, result-producing activities—that is, activities that produce measurable results that can be related, directly or indirectly, to the results and performance of the entire enterprise.
Second, support activities that, while needed and even essential, do not by themselves produce results but have results only through the use made of their “output” by other components within the business.
Third, activities that have no direct or indirect relationship to the results of the business, activities that are truly ancillary. They are hygiene and housekeeping activities.
Finally, is the top-management activity. Among the result-producing activities, there are some that directly bring in revenues (or in service institutions, directly produce “patient care” or “learning”). Here belong innovating activities, selling and all the work needed to do a systematic and organized selling job. Here also belongs the treasury function, that is, the supply and management of money in the business.
Key activities should never be subordinated to nonkey activities. Revenue-producing activities should never be subordinated to nonrevenue-producing activities. And support activities should never be mixed with revenue-producing and result-contributory activities.
The main rule is to look upon simulated decentralization as a last resort only.
Whenever a unit can be set up as a business, no design principle can match federal decentralization. We have learned, however, that a great many large companies cannot be divided into genuine businesses. Yet they have clearly outgrown the limits of size and complexity of the functional or of the team structure. These are the companies that are increasingly turning to “simulated decentralization” as the answer to their organization problem. Simulated decentralization forms structural units that are not businesses but which are still set up as if they were businesses, with maximum possible autonomy, with their own management, and with at least a “simulation” of profit-and-loss responsibility. They buy from and sell to each other using “transfer prices” determined internally rather than by an outside market. Or their “profits” are arrived at by internal allocation of costs to which then, often, a “standard fee,” such as 20 percent of costs, is added.
There must be a kind of “supremacy clause” reserving to central management the decisions that affect the business as a whole and its long-range future welfare.
Top management in a decentralized company must think through carefully what decisions it reserves for itself. For there are decisions that have to do with the entire company, its integrity, and its future. These decisions can be made only by somebody who sees the whole and is responsible for the whole. Specifically, there must be three reserved areas if the business is to remain a whole rather than splinter into fragments. Top management, and top management alone, can make the decision on what technologies, markets, and products to go into, what businesses to start and what businesses to abandon, and also what the basic values, beliefs, and principles of the company are. Second, top management must reserve to itself the control of the allocation of the key resource of capital. Both the supply of capital and its investment are top-management responsibilities that cannot be turned over to the autonomous units of a federal organization.
Third, the other key resource is people. The people in a federally organized company, and especially managers and key professionals, are a resource of the entire company rather than of any one unit. The company’s policies with respect to people and decisions on key appointments in the decentralized autonomous businesses are top-management decisions—though of course, autonomous business managers need to take an active part in them.
As a minimum the unit must contribute a profit to the company rather than merely contribute to the profit of the company.
Federal decentralization has stringent requirements. Federal decentralization is applicable only where a company can truly be organized into a number of genuine “business.” This is its basic limitation. As a minimum the unit must contribute a profit to the company. And it must be a genuine profit determined by the objective judgment of the marketplace.
Federal decentralization will work only if the top-management job is clearly defined and thought through. Federalization, if properly applied, makes top management capable of doing its own job precisely because it does not have to worry about operations, but can concentrate on direction, strategy, objectives, and key decisions for the future. The federal principle demands great responsibility from the operating units, the autonomous businesses. They are given the maximum of autonomy; and this requires that they assume the maximum of responsibility. Federal decentralization requires centralized controls and common measurements. Both the managers of the autonomous businesses and top management must know what is expected of each business, what is meant by “performance,” and what developments are important. To be able to give autonomy one must have confidence. And this requires controls that make opinions unnecessary. A federal unit of a company is autonomous, but it is not independent and should not be. Its autonomy is a means toward better performance for the entire company.
The greatest strength of the federal principle is that it alone of all known principles of organization prepares and tests people for top-management responsibility at an early stage.
In “federal decentralization” a company is organized into a number of autonomous businesses. Each unit has responsibility for its own performance, its own results, and its own contribution to the total company. Each unit has its own management which, in effect, runs its own “autonomous business.”
In a federally organized structure, each manager is close enough to business performance and business results to focus on them. The federal principle therefore enables us to divide large and complex organizations into a number of businesses that are small and simple enough that managers know what they are doing and can direct themselves toward the performance of the whole instead of becoming prisoners of their own work, effort, and skill. Because management by objectives and self-control become effective, the number of people or units under one manager is no longer limited by the span of control; it is limited only by the much wider span of managerial responsibility. The greatest strength of the federal principle is, however, with respect to manager development. This by itself makes it the principle to be used in preference to any other.
Federalism relieves top management from operating duties and sets it free to devote itself to its proper functions.
What the enterprise needs is a principle that gives both the center and the parts genuine managerial functions and powers. This principle is federalism, in which the whole of the enterprise is conceived as made up of autonomous units. The federal enterprise and all its units are in the same business. The same economic factors determine the future of the whole as well as of all units; the same basic decisions have to be made for all of them; the same kind and type of executive is needed. Hence the whole requires a unified management in charge of the basic functions: the decision what business the enterprise is in, the organization of the human resources, and the selection, training, and testing of future leaders.
At the same time, each unit is a business by itself. It produces its own products for a distinct market. Each unit must, therefore, have wide autonomy within limits set by the general decisions of the management of the whole. Each unit has to have its own management. The local management will be primarily an operating management; it will be concerned mainly with the present and immediate future rather than with basic policy. But within a limited scope it will have also to discharge real top-management functions.
Organization is a tool. As with any tool, the more specialized its given task, the greater its performance capacity.
Organizations are special-purpose institutions. They are effective because they concentrate on one task. If you were to go to the American Lung Association and say, “Ninety percent of all adult Americans suffer from ingrown toenails; we need your expertise in research, health education, and prevention to stamp out this dreadful scourge,” you’d get the answer: “We are interested only in what lies between the hips and the shoulders.” That explains why the American Lung Association or the American Heart Association or any of the other organizations in the health field get results.
Society, community, family, have to deal with whatever problem arises. To do so in an organization is “diversification.” And in an organization, diversification means splintering. It destroys the performance capacity of any organization—whether business, labor union, school, hospital, community service, or church. Because the organization is composed of specialists, each with his or her own narrow knowledge area, its mission must be crystal clear. The organization must be single-minded, otherwise its members become confused. They will follow their specialty rather than applying it to the common task. They will each define “results” in terms of that specialty, imposing their own values on the organization. Only a clear, focused, and common mission can hold the organization together and enable it to produce results.
One hears a great deal today about “the end of hierarchy.” That is blatant nonsense.
To attack industrial society, as would the sentimental equalitarian, because it is based on subordination instead of on formal equality is a misunderstanding of the nature of both industry and society. Like every other institution that coordinates human efforts to a social end, the corporation must be organized on hierarchical lines. But, also, everybody from the boss to the sweeper must be seen as equally necessary to the success of the common enterprise. At the same time, the large corporation must offer equal opportunities for advancement. This is simply the traditional demand for justice, a consequence of the Christian concept of human dignity.
The demand for equal opportunities is not, as is often mistakenly assumed, a demand for absolute equality of rewards. On the contrary, equal opportunities automatically assume an inequality of rewards. For the very concept of justice implies rewards graduated according to unequal performance and unequal responsibility.
