The ECCO System
Cybernetic Principles for Effective Control in Complex Organizations
Click here for footnotes, Ch. 3.
American Management Legacy: The Weber/Taylor Bureaucracy
Roots of the Weber/Taylor System
The literature shows almost unanimous agreement that the Weber/Taylor organizational structure is typical of U.S. companies, and that it is an impediment to the adoption of TQM, other alternative business management strategies1,2,3 and technology-based systems solutions.4 It appears intuitively obvious that this should be the case, but it will be instructive to examine the genesis and implications of the system in more detail, to see exactly what about the system does not work. This way, we will see not only what needs to be eliminated, but what functions need to be replaced by more appropriate controls.
Max Weber was a sociologist in Europe around the turn of the century, at the dawn of the industrial revolution and the end of the late feudal period. Feudal organizations were personal, governed by favoritism, cults of personality, and were rife with nepotism and other forms of institutionalized inequity. The industrial revolution and emerging democracy in Europe raised the problem of how to construct organizations able to deal with masses of people services which would be both efficient and fair. The feudal form of organization was not capable of dealing with large administrative tasks, such as tax collection, mass delivery of services, or census. Weber observed, chronicled, and gave a theoretical basis to the emerging form of organizational structure developed to fill this need. He described it as being fair by being impersonal, governed by rules rather than by personality and favoritism, the earmarks of feudal organizations.
In administrative terms, one person was no longer enough to control such a complicated operation as mass production single-handedly. For one thing, there were fiscal aspects, such as the amassing and management of large amounts of capital which enabled investment in mass production equipment, machinery, and manpower. Handling sales of massive quantities of goods, and the accounting required to track sales income, expenditures on raw materials, subcontracting of services, and payroll were likewise detailed and complicated operations which required specialized training and familiarity. The sheer number of man/hours required to supervise the number of employees necessary to perform the operations of the firm clearly outstripped the capabilities of a single person, or handful of persons, to control. Finally, in an organization with so many facets, and so many employees, it was imperative that control be organized and reliable--reproducible as it were--so that the central plan could be carried out without relying on (or being impeded by) particular personalities inside the organization.
Organizations at the turn of the century were also on the verge of having to deal with large problems and co-ordinate detailed operations. Frederick Taylor's scientific management techniques were designed to break down complicated tasks, like the assembly of an automobile, or the steps in a complicated accounting operation, into a series of elementary tasks which could be routinized to take advantage of the cheap labor from an uneducated workforce. Once the tasks were divided and specified, each task could be taught to a worker who could perform it by rote, and the application of analytic techniques, such as time studies, could further optimize the task performance.
Thus the Weber model of efficient bureaucracy provided the form. The Taylor system of scientific management provided the content. In a Weber/Taylor structure5 it is tacitly assumed that the taxonomic divisions of the organization into divisions, departments, sections is appropriate for the tasks they are to perform--that is that the tasks are fairly discrete by nature from other operations in the organization. It is also assumed that once a Taylor analysis has been performed on the operations of each department that it has optimally distributed and sequenced the tasks. Finally, it is assumed that resources can be distributed appropriately to each part of the organization, based on their projected operational costs. If all of these conditions hold, then further increases in efficiency can be obtained by optimizing the operations of individual departments.
One of the major contributions of the Weber/Taylor structure was its ability to expand central control while distributing both authority and responsibility. In feudal organizations, control was granted based on political favors or familial obligation. A ruling authority would hand over rights to control of a region of land in return for another's fealty, or a portion of income or goods from the land. The administration of the region would be largely at the whim of the local ruler, and out of control of the central authority. Clearly, this was not acceptable in a manufacturing organization which had detailed output demands and extensive, though linear, co-ordination requirements.
The Weber/Taylor system provided a method for the accountable distribution of authority commensurate with responsibility. This distribution devolved in an orderly manner, from the top down, with each layer of organization unfolding as a set of further specified tasks, components which made up the mission of the next higher level, controlling the next lower level of detailed tasks distributed among immediate subordinates. The Weber/Taylor system could track operations by formal reports and other forms of documentation. It demanded that specific scientifically designed procedures be adhered to, and that work output be provided in particular forms, and meet specified standards.
The sine qua non of the control method of the Weber-defined bureaucracy was the notion of limited allocation of personal responsibility along with commensurate authority for specified tasks or aspects of operation. Thus, the major form of control was limitation of choice on the part of employees and managers. To increase control, the system dictated that a decrease in the number of choices allowed at various levels of management was the way to do so. In addition, the individual manager would assume personal responsibility for the actions of his subordinates.
All of these practices enabled supervision and correction to proceed on empirical, impartial grounds. So long as markets remained reasonably stable, advances in technology progressed slowly, and competition was moderate, the Weber/Taylor organization proved adequate to the task of controlling the complication inherent in large manufacturing organizations.
Inadequacies of the Weber/Taylor System
Why is the Weber/Taylor system inappropriate for the control of modern manufacturing organizations? In truth, challenges to the system were mounted as far back as 1930s with the work of such management theorists as MacGregor, Mayo, Barnard and Selznick. Their views challenged the old Theory X view of workers as mechanisms, and proposed a new school of thought based on co-operative, creative human behavior.6 Though others followed, and new approaches to management as people-co-ordination were developed, no significant work was advanced to change the essential structure and control mechanisms of the organization beyond taking steps to humanize them and give workers a voice, individually or collectively.
Most U.S. organizations, according to Ouchi, remain highly bureaucratic in the Weber/Taylor mold. Because of the structural rigidity of the "functional" departmental divisions, Weber/Taylor organizations demanded and developed enormous specialization. This was both a strength and a weakness. Ouchi notes:
The great strength in the American approach lies in the ability to organize its specialized workers into a co-ordinated workforce. . . .
[However,] [t]he weakness of the American form is that workers can never be intimately integrated with one another. . . . [T]hey can co-ordinate so long as there is no call for understanding of anything beyond their own specialty. . . . When the underlying production process demands integration of effort at an intimate and custom-crafted level, the employees must be asked to give up their local objectives and narrow professional point of view to learn new skills, to take on new goals.7
We can understand why this must be the case by looking at the effects of the bureaucratic rules, communication channels and organizational structure in terms of hierarchical models of control. A subsystem within a system is defined as a subset of elements being highly dynamically interconnected among themselves. Boundaries can be formally, if relativistically, drawn when few I/O structures feed into and out of this highly interconnected set of elements. The Weber/Taylor system tends to define the subsystems formally based on historic task distribution. Communications are primarily vertical: thus, intradepartmental communications far exceed interdepartmental communications as well as inter-level communication. The boundaries are drawn first, which defines departments and their specialized subunits. Therefore, departmental specialization in the bureaucratic organization formally constrains the department-subsystem's relationships with each other and the outside world. These units also tend to operate as subsystems since most of the elements (people) within the subsystem interact around common outputs (goals) and within common rule requirements, etc. with others in the subsystem, far more than they do with elements (individuals) outside the subsystem.
Compare this description with Koestler's description of "holons", a general characteristic of certain kinds of subsystems in hierarchies. The common interactions in addition to the formal boundary and control functions defined on a subsystem tend to create holons in fact as well as in principle. Holons are defined by Koestler:
Every holon has a dual tendency to preserve and assert its individuality as a quasi autonomous whole; and to function as an integrated part of (an existing or evolving) larger whole. This polarity between the self-assertive and integrative tendencies is inherent in the concept of hierarchical order; and a universal characteristic of life. The self-assertive tendencies are the dynamic expression of holon wholeness, the integrative tendencies of its partness.8
Thus, subsystems tend to develop subcultures and internal perspectives of themselves.
Ziemke and Spann cite the results of this characteristic structural specialization as a major impediment to concurrent engineering efforts.
This division of work into departments or functions is intended to increase productivity and efficiency, but it also creates problems. Members of subunits develop values consistent with the subunit. Marketing department personnel may value unique product design features whereas accounting department personnel may value cost control. Departments can also differ in their time orientation. Production departments may have a very short time frame whereas R&D departments may have a much longer one. Based on [these differences], departments develop departmental goals, which may or may not be consistent with overall organizational goals.
As organizational structures grow taller, response times slow down, and bottlenecks occur. Interdepartmental communications become more difficult. Interdepartmental barriers arise. Departments build and protect their own turf. Rivalry between departments develops and becomes especially bitter when organizational resources are scarce. Organizational goals get subverted by department goals.9
This theme is echoed by Melan in relation to product development.
People working within an enterprise are organized onto traditional, vertical structures by skill, function, or product specialty. However, work continues to flow horizontally, crossing the functional boundaries of organizations. Because of the value and reward systems inherent in these classical structures, suboptimization of work flow tends to occur. This occurs when the global values of the entire system take second place to local or functional considerations.10
The Taylor approach to division of labor and measurement of individual task productivity is equally damaging, even within departments. There is an implicit assumption, given the Taylor breakdown of tasks and task areas, that the maximization (optimization) of subsystem performance is the goal of the subsystem. Again, all of its rules and reward structures stress local optimization over organizational goals. The fundamental nature of this error can be observed by looking at the nature of dynamic processes such as those found on a production line. Ouchi relates the following example of a production line at a toilet paper factory:
The factory consists mainly of huge machines that put raw material through a series of steps, ending up with rolls of toilet paper. An error at one step of the process may not show up until three steps later, at which point a whole batch will be ruined: loss of productivity. Traditionally, these machine operators were supervised, as they are in virtually all of United States industry, by a foreman who directed their efforts. . . . Each machine operator had his or her own individual measure of performance - time away from the work station, scrap loss and so on. If one worker suspected that a small adjustment at his station could produce a better flow two steps down, he had no incentive to go talk to the other worker concerned. Indeed, his work report would reflect time away from his station, and he would be punished.
Roles of Managers
It would seem obvious that managers observing this kind of behavior would be able to compensate and change operations within their purview to avoid this kind of behavior. And even with appropriate organizational change going on around them, the manager's job remains a difficult one. Rosabeth Moss Kanter looks at the impact that demands of flexible organizations have on managers:
With little precedent to guide them, [managers] are watching hierarchy fade away and even the clear distinctions of title, task, department, even corporation, blur. . . .
But so far, theorists have given scant attention to the dramatically altered realities of managerial work in these transformed corporations. We don't even have good words to describe the new relationships. 'Superiors' and 'subordinates' hardly seem accurate, and even 'bosses' and 'their people' imply more control and ownership than managers today actually possess.
In non-transformed organizations, managers might "know" their job (albeit knowledge of a dysfunctional role); and the fact is that managers function within and are subject to the bureaucratic control structures as well. In the bureaucracy, the flow of control is explicit and top-down. As Allen and Starr observe about the nature of hierarchical systems in general:
"Ordered systems are so, not because of what the components do, but rather what they are not allowed to do. . . . Hierarchies can be profitably viewed as systems of constraint. Any holon higher in the hierarchy exerts some constraint on lower holons with which it communicates. . . . Higher holons provide, to various extents, the environment in which lower holons operate. . . . We define sister holons as having no direct constraint over each other."
Each successive level of management governs (constrains) the subsystems immediately below them. Feedback is bottom up, and comes in two forms: First, reports of selected output, and second, documentation that certain prescribed procedures were followed. This communication is commonly collected, attenuated and fed back to successively higher levels in the hierarchy. There is an implicit historical assumption that managers are at least as knowledgeable about operations as their operational employees whom they supervise. This is often not the case in modern manufacturing organizations.
Resources are typically allocated and distributed top-down through the chain of control, based on bottom-up requests and supporting information (documentation and projections) in response to task allocation. Thus, some sorts of negotiations take place vertically through the hierarchy, but rarely if ever horizontally. Centralized control is maintained by the orderly distribution of authority and concomitant responsibility at each level, and monitored by feedback on output, and on standardized procedures and measures (such as productivity man/hours). There are no direct information links to other subsystems (laterally or diagonally). Most communications, negotiations, etc., are mediated by mutually superior nodes in the hierarchy.
In order to control the human variability in behavior in these systems, individual managers must be held accountable for the performance of those subsystems under them, both for output of the subsystem, and generally also for following specified procedures. The manager is responsible for negotiating the allocation of resources for his subsystems, and each manager of each subsystem is rewarded for optimizing the performance of his own subsystem. Since managers of subsystems must compete against other subsystems for allocation of scarce resources, and each is responsible for the maximization of his subsystem's performance, there is a natural incentive for competition rather than co-operation among subsystems.
Thus managers are caught in multiple traps. The procedures and measurements which drive their departments are largely dictated by higher authority; the basis on which they are judged for retention, promotion and raises require competition against other departments rather than co-operation for the company's greater good; and there is an absence of alternative theory as well as role models to support them in efforts to change, even under the best of conditions.
Advantages of a Weber/Taylor System
There are a number of problems which are well-solved by the Weber/Taylor control structure. First, the structure allows for the localization of errors or problems in operational subsystems. It accomplishes the orderly distribution of tasks which are too numerous, too detailed, or too technical for a single individual or small group of individuals, and ensures accountability for the tasks performance. This structure enables a system to be impersonal, that is, to be independent of particular personalities in order to accomplish a task, particularly at Taylor-analyzed task levels. It is ideal for organizing large operations which are stable (unchanging) for long periods of time, which have many linear dependencies, and parallel operations. It is designed so that a number of unchanging constraints may be incorporated at the command level, assessed and distributed among the subsystems.
A strict hierarchy is vital for operations such as standard military campaigns, particularly where diverse forces (air, sea, and land) must be centrally co-ordinated for attacks and defense. It is virtually essential for operations where participants are not familiar with each other, nor with the tasks at hand. Its effectiveness may be highly dependent on a strong intelligence force and timely and excellent quality communications. It also depends on its officers ("managers") knowing their sub-ordinates' jobs and capabilities and having increasing experience at ascending levels of command. It is not as necessary for guerrilla operations, however, where battle lines are unclear (if they exist at all) and the necessary command and control communication links are often not available.
Disadvantages of the Weber/Taylor System
Because of the vertical control structure and the incentive for competition among subsystems, this approach is not well-designed for problems which are complex and non-linear. There are no structural provisions for constraints between subsystems nor the negotiations thereof. It is not designed so that complex, detailed information can be transmitted to the top level of the bureaucracy for consideration--nor does it assume that the lower levels of the hierarchy contain much useful information or knowledge.
Structurally, one of the most potentially damaging features of the Weber/Taylor hierarchy is the incentive to optimize subsystem performance. In any dynamic system, optimization of all subsystems almost always results in the sub-optimization of the overall system goals. The only conditions under which this may not be the case is if 1) all subsystems goals are strictly supportive of overall system goals; 2) each subsystem has been allocated precisely the appropriate share of resources; and 3) the subsystem configuration is flawless with respect to achieving overall system goals and 4) these conditions do not change. Far more often, the optimization of subsystem goals results in serious sub-optimization of overall system goals. Judicious sub-optimization of subsystem goals, and the ability to trade-off and dynamically reconfigure subsystem interactions as necessary will generally provide a superior system outcome.
Though the Weber/Taylor form of organization was a significant step forward into the industrial revolution, current conditions have radically changed the demands placed on modern manufacturing organizations. These changes include shortening of product lifecycles, volatility of design requirements, shortened time to market, a market which is much more sensitive to increased quality and reliability in both design and manufacture, and diminishing profit margins. These add up to a need for close co-operation and co-ordination of a variety of departments and functions. In addition, much critical and specific technical knowledge rests with operational personnel, particularly in the engineering and manufacturing sector, rather than with upper management. Enabling technologies exist, but these, too, are a source of organizational complexity. The old cornerstone of the Weber/Taylor organization, the need-design-make-sell model of industry may be gone forever. As Nevins and Whitney observe, "It has too much linearity. . . . It is too compartmentalized, assuming that design is the domain of the designer, manufacturing the domain of the manufacturing engineer, purchasing the domain of the purchasing manager, and so on."
The next step is to explore what mechanisms might co-ordinate and control manufacturing and collateral operations in the current environment. The foundations of such a theory might be found in the field of cybernetics, which concerns itself with the study of communication and control in complex systems. In the next chapter, these foundations will be examined and specific principles for such a control system will be proposed.
Back totable of contents On to next chapter Back to home page