Virtual Organization

Keywords virtual organization; virtual constructs; metamanagement; switching; commoditization of information; standardization; abstractification of wealth

INTRODUCTION

Several constructs characterized by the adjective `virtual’ have achieved prominence in the fields of computing and telecommunications. Among the most important are virtual memory, virtual circuit and virtual reality. These ideas have proved of great theoretical and practical importance in the design and implementation of computer systems, networks and human-computer interaction, respectively. Related social and organizational constructs such as virtual office (Giuliano, 1982), virtual classroom (Hiltz, 1986), virtual corporation (Davidow and Malone, 1992), virtual team (Hammer and Champy, 1993), and virtual community (Rheingold, 1993) have also been introduced. Since the early 1980’s the author has been developing a theory of virtual organization that subsumes and generalizes these specific constructs. The most extensive formulation of the theory to date is presented in Mowshowitz (1994). The current article is intended as a further contribution to the theory of virtual organization. Thus, the focus is on the nature and inner workings of this organization innovation, rather than on its consequences for individuals and communities. Using the framework of the earlier article, the definition, applicability, and socio-technical foundations of virtual organization are elaborated. See Mowshowitz (1994, 1992b) and Harrington (1991) for a discussion of virtual organization in relation to theories of organization in general.

WHAT IS VIRTUAL ORGANIZATION?

Virtual organization applies to goal-oriented activity. This includes much (but not all) of the activity undertaken by biological, social, and artificial systems. The philosophical foundation of virtual organization is a categorical distinction between needs and the means for satisfying them. This distinction makes it possible to manage activities in a way that insures systemic efforts to find the `best’ match between requirements and satisfiers at all times. An analogous distinction plays a critical role in the organization of virtual memory computer systems. Here. logical (or virtual) memory constitutes the storage requirements referenced by the programmer, while the physical memory of the machine comprises the means for satisfying these requirements. The categorical distinction in this case allows the operating system to execute a scheme for optimizing the use of the computer’s limited storage capacity.

The author first conceived the idea of virtual organization as an analogy between virtual memory and the practices of multinational firms. The analogy combines elements drawn from three disparate lines of inquiry. These are (1) the structure of virtual memory, (2) the distinction between an object language and a metalanguage that is made in the foundations of mathematics, and (3) the practices of complex organizations. A virtual memory computer is able to overcome the limitation of a fixed storage capacity, because only a portion of the contents of the unbounded logical space (or virtual memory) needs to be resident in the finite physical memory at any given moment. The distinction between an object language and a metalanguage in metamathematics is comparable in the sense that it facilitates logical analysis that would otherwise be impossible–that is, certain questions about a mathematical system (object language) can be formulated in a metalanguage which is not subject to the limitations of the object language. Multinational firms that switch between production units or between suppliers located in different parts of the world constitute the third source of the idea of virtual organization. Here again, leverage is achieved by means of distinguishing between means (e.g. plants or suppliers) and requirements (e.g. production components).

These foregoing remarks are intended to motivate the following three-part definition of virtual organization adapted from Mowshowitz (1994). Part 1 captures the structure of virtual organization as `virtually organized tasks’; part 2 describes the function of virtual organization as `metamanagement’; and part 3 combines the two elements to complete the definition.

PART 1

A virtually organized task is a goal-oriented activity that is implemented by an appropriate assignment of concrete satisfiers to the abstract requirements of a task.

Abstract requirements refer to the logically defined needs of a task. Making a product, for example, requires raw materials, tools, and labor. Each of these requirements may be viewed as an abstract need, in the sense that it can be met in a variety of ways. The particular ways–specific raw materials from a given supplier, designated tools in a particular building, named individual workers–in which a requirement can be met constitute concrete satisfiers. This separation of abstract requirements from concrete satisfiers corresponds to the distinction between logical and physical storage in a virtual memory computer, which, in turn, corresponds to the distinction between object language and metalanguage in metamathematics.

Implementation of the activity of a virtually organized task takes the form of an assignment of satisfiers to requirements. This assignment may vary over time and must be tracked by management. Such tracking corresponds to the function of the operating system of a virtual memory computer in keeping tabs on the whereabouts of things in the various levels of memory in the computer.

PART 2

The management of a virtually organized task is called metamanagement and consists of four basic activities:

(1) analyzing abstract requirements;

(2) tracking the possibilities for satisfying requirements;

(3) developing and maintaining the procedure for assigning (or allocating) satisfiers to requirements; and

(4) adjusting the optimality (or `satisficing’) criteria of the allocation procedure.

This definition emphasizes the shared conceptual features of virtual organization and metamathematics. Since virtual organization is based on a categorical distinction between means and ends, the activities involved in analyzing requirements, on the one hand, and investigating possible satisfiers, on the other hand, are independent of each other. But it is important to note that each of these activities is closely related to the other. Requirements analysis corresponds to the use of a metalanguage, while satisfiers investigation is analogous to the function of an object language in metamathematics. The former must take account of the universe of concrete means, and the latter must incorporate the abstract specification of needs. This linkage between requirements analysis and satisfiers investigation is mirrored in metamathematics. Metalanguages are used to investigate object languages that usually `serve as models for parts of informal mathematics and logic’ (Kleene, 1952, p. 64).

Requirements and satisfiers can be distinguished in any goal-oriented activity, and the ability to draw such a distinction is the only restriction on the applicability of metamanagement. This approach can thus be used to manage informal as well as formal systems.

Managing under the structural separation inherent in virtually organized tasks focuses attention on the procedure for assigning satisfiers to requirements. Developing and maintaining the procedure constitutes an independent metamanagement function. New information, changes in knowledge (e.g., about the environment or about methods for selecting alternatives), altered strategies or goals, etc., may dictate changes in the allocation procedure itself. The advantage of making the allocation procedure explicit in the function of management is evident from the success of operations research. Not all procedures can be as formal as the techniques of operations research, but it is clear that improvements in efficiency and performance can be achieved by improving the way in which satisfiers are assigned to requirements.

The last of the four functions of metamanagement–examining criteria and goals–explicitly injects self-reflection in organizational life. This function makes it clear that metamanagement is no mere mechanical or reductionist system of management. On the contrary, by identifying criteria and goals explicitly and making their analysis into an independent activity, metamanagement introduces a systematic discipline which calls upon managers to justify strategic choices and commitments. The separation of satisfiers from requirements, and the need for an explicit allocation procedure, force management to examine and possibly revise goals and criteria continually. These conditions cultivate the habit of self-reflection as a regular feature of management behavior. This has never been present in such a systematic way before. The word `systematic’ is stressed because isolated instances of self-reflection can certainly be found in conventional organizations, but in metamanagement it is standard operating procedure.

This new approach does not diminish the capacity of managers to set goals and establish values, nor does not it reject intuition and creativity in decision-making and problem-solving. By making goals and values explicit, and establishing standards for judging action, metamanagement neither eliminates nor abridges, tools–it sharpens them.

Reflection on organizational goals and consequent examination of the criteria to be met in satisfying abstract requirements are essential for realizing the potential advantages of metamanagement. This reflective activity makes it possible to anticipate necessary or desirable adaptations, or to effect changes in the definition of the organization itself. To survive and prosper, an organization must not only use effective means to reach its goals, but it must have appropriate goals; i.e., it must be in a position to adjust goals and strategies as internal and external conditions warrant.

PART 3

A virtual organization is a goal-oriented enterprise operating under metamanagement.

The individual components in the definition of virtual organization have been around for some time. What is new is their integration into a coherent and complete organizational paradigm. Neither the conception nor the practice is likely to have arisen without the technological innovations that make virtual organization practicable. Advanced information technology, especially networking, has made it possible to exploit this way of thinking about and organizing activities.

An important element of virtual organization, namely the allocation of satisfiers to requirements, draws on the concept of resource allocation in operations research. Whereas operations research employs this concept to solve narrowly defined optimization problems, the new paradigm incorporates it into a systematic approach to the management of tasks.

Virtual organization is a way of structuring, managing and operating. In separating the abstract requirements from concrete satisfiers, management is forced to dedicate independent activities–in an ongoing systematic way–to examining requirements, and, in parallel, to looking at the possibilities for satisfying those requirements, both of which may change over time. This is very different from the conventional way of managing, which is to mix the two together with no clear distinction between a requirement and a way of satisfying it. In the conventional approach these two activities are `hard-wired’ together, as occurs in a company with traditional suppliers whose relationship to the company is never seriously examined.

As suggested earlier, one indication of virtual organization in practice comes from the operations of multinational enterprises. In particular, the coupling between multinationals and their suppliers has become more fluid. This is characteristic of a marketplace in which `buyers’ search for the cheapest way to satisfy their requirements, and `sellers’ seek the highest price for the satisfiers they offer. The marketplace model could be applied advantageously to a variety of functions within an organization. Turoff (1985) argues for such an approach to information management within organizations. The attraction of the fluid arrangements of the marketplace derives from the possibility of gaining some economic advantage by being able to shift from one supplier to another. A shift may be warranted by cost, reliability, or other advantages, that may ultimately translate into a competitive edge for the firm. Of course, the idea of replacing a supplier is nothing new, but structuring operations in such a way as to make it feasible to shift suppliers on demand is something new. Virtual organization allows for systematic shifting on demand.

This ability to shift–called here `switching–‘ is an example of a basic principle of management, on a par with the innovations that Adam Smith observed in the factory system that gave rise to such enormous increases in productivity. Like those earlier innovations, switching is based on a paradigm shift rather than a material change. The rational factory system was based on division of labor, the new production systems are based on virtual organization.

SWITCHING: THE UNIQUE CONTRIBUTION OF VIRTUAL ORGANIZATION

Switching depends on the categorical separation of requirements from satisfiers. At any given moment there is an allocation of satisfiers to requirements, but the particular allocation can change over time. This dynamically changing allocation is governed by considerations of cost and other factors relevant to competitive advantage. The use of switching in business practice is only just beginning. Although the requisites for its use are in place, the idea will not gain widespread acceptance overnight. It will take some time for switching to become a standard instrument in the manager’s toolkit.

The history of the factory tells us what to expect. Factory methods of production were in use long before Adam Smith provided an analytic demonstration of their advantages over craft methods. Indeed, as Mumford (1934) observed, factory production existed in nascent form in late medieval times. However, this revolutionary form of production organization was not exploited in a systematic way until much later. It was not until the late eighteenth century that the potentialities of the new methods were fully appreciated. The steam engine came along when the basic social innovations of the factory were clearly understood. In a word, the technological seed provided by the steam engine as a reliable source of power fell on fertile social ground.

Like earlier innovations, information technology has emerged in a receptive social environment. The ground has been prepared for the exploitation of this new technology. However, considerable experimentation in organizational practices will be needed to hone the switching techniques of virtual organization into an effective management instrument. Here and there one can observe the new paradigm in action. Eventually, however, metamanagement will become standard operating procedure–not just an ad hoc practice occurring in isolated instances.

As noted earlier, the theory of virtual organization can be used to model a wide variety of systems, ranging from virtual circuits in computer networks to virtual teams in corporations. Examples of current practices in the business world that reflect virtual organization are given below.

Automobile Assembly

Automobiles are complex machines made up of many different components, produced by a variety of suppliers. The abstract requirements in this case may be seen as, the components required to put together an automobile. Concrete satisfiers are the suppliers of the various parts. Since components may be available from a number of different suppliers, it is possible for management to switch from one to another to take advantage of dynamically changing opportunities in the marketplace–for example, a new player may come along offering lower costs, etc. It is important to bear in mind that this kind of switching is not a one-time effort, but an ongoing management activity, triggered by events in the marketplace. This dynamic activity, made possible by treating requirements and satisfiers independently, is the essence of virtual organization. The optimality criteria in this example have to do with cost and related parameters of the assembly process. As in the supply domain, the objectives of the organization require continual re-examination in relation to the changing conditions in the marketplace. For example, the willingness of prospective automobile buyers to pay a premium for quality can change over time, requiring management to adjust its production criteria.

Tax Management

The management of a firm’s tax obligation is another facet of manufacturing in which virtual organization can be used to advantage. This has to do with the financial rather than the production side of the business. In this case, the abstract requirements can be interpreted as the general provisions of corporate income tax regulations in different jurisdictions; the concrete satisfiers can be interpreted as the specific measures that could be taken to satisfy, those provisions of the tax regulations.

Transfer pricing, for example, can be used to exploit differences in rates of taxation between jurisdictions. Suppose Cosmic Multicorp with plants X and Y located in countries A and B, respectively, manufactures a product at those plants that requires a certain component T that is also produced at each of those plants. Suppose further that country A has a higher tax rate than country B. Corporate income tax can be reduced by having plant X buy some of its quota of T’s from Y at a higher price than Y normally gets from other customers. Assuming the real costs of producing component T are roughly the same at X and Y, and disregarding possible legal complications, this ploy would increase overall corporate profits because the taxes paid by X would be lowered more than the taxes of Y would be raised (by increased profits on the sale of T).

The concrete satisfiers in this example are the different levels of expenditure on component T. Corporate taxes are typically levied on profits, and profit levels are determined in part by the costs of production inputs. Since the pricing of component T is an intra-company matter, the differential pricing scheme is strictly a bookkeeping operation. Switching, in this case, means the changes in the production cost schedules of different plants, and consequent shifts in interplant shipments of components. Clearly, the optimality criterion is the minimization of overall corporate taxes.

Investment Management

The way in which financial institutions invest funds for clients can also be modeled as metamanagement. In this case, the abstract requirements include expected rates of return, asset allocation strategies, and risk profiles, specified by the clients; the concrete satisfiers are the available investment vehicles (e.g., stocks, mutual funds, commodity futures, options, bonds, treasury securities, money market funds). Switching is the movement of funds between different investment vehicles according to changes in the possible satisfiers, in relation to the investment criteria. These criteria (the optimality criteria)–which must be reviewed and adjusted continually in light of market conditions–are the risk/return ratios specified by the portfolio manager.

Arbitrage operations in securities trading is a related example in which the abstract requirements are defined by the investment specifications of the traders, and the concrete satisfiers are the possible mixes of securities and index futures in a portfolio. Switching means shifting the balance between stock index futures and positions in the stocks making up the index. The aim of such switching is to take advantage of imbalances in the relative valuations of securities by stock and index traders. As in the previous example, the optimality criterion is a risk/return ratio.

The idea of metamanagement (and switching, in particular) was suggested to the author by the striking similarities between the practices of some multinational companies and the management of memory in a virtual computer. Company actions to relocate production facilities from one country to another strongly suggest operations similar to those of virtual memory. Shifts in the locus of production correspond to changes in the correspondence table linking virtual and actual storage locations in a virtual memory computer. Relocation of facilities suggests that management treats production as an abstract requirement capable of being satisfied by a plant located anywhere.

Closer inspection of the workings of virtual memory reinforces the analogy with the activities of multinational management. The operating system of a virtual memory computer keeps a program in primary memory until it detects an opportunity to make `better’ use of the storage space. For example, the program may be executed until an input-output instruction occurs. Since such instructions take much longer than `ordinary’ (e.g., arithmetic or logical) ones, the operating system may relocate the program to secondary memory and bring in another program while the input-output operation is performed. By taking advantage of differences in the time required for different computational tasks, the operating system can reduce the total time required to execute a collection of programs. Systematic pursuit of this strategy results in improved utilization of the computer’s limited memory resources. The same is true of the management of multinational companies. By taking advantage systematically of differences in production costs in different countries, management can improve the utilization of a company’s limited capital resources.

SOMETHING FOR NOTHING?

Metamanagement, like the operating system of a virtual computer, achieves better performance than conventional management by exploiting an organizational trick that permits the systematic use of switching. At first glance, this stratagem may appear to be yielding something for nothing. The benefits of virtual organization are very real, but they are not free. Before elaborating on the possible costs, we will examine some related cases of apparent `free lunches’ with a view to clarifying the organizational trick that makes switching possible.

Maxwell’s Demon (see Brillouin, 1962)

The famous physicist Maxwell constructed a `thought experiment’ to show how the second law of thermodynamics might be violated; i.e., how the entropy of a closed system could be decreased without a corresponding increase somewhere else. Maxwell proposed to `organize’ the molecules of a gas in a closed container by inserting a trap door dividing the container into upper and lower chambers. The trap door was to be operated by a `demon’ capable of distinguishing between high and low-velocity molecules. Upon detecting a high-velocity molecule, the demon would open the trap door permitting it to move into (or remain) in the upper chamber, consigning all the low velocity molecules to the lower chamber.

So it would appear that with this demon you could in fact get something for nothing because eventually all the fast-moving particles would be in one part of the container and the slow moving ones in the other. This would involve a decrease in the entropy of the system and, according to the second law of thermodynamics, that is not supposed to happen. We will see later that the decrease in entropy is not really free.

Task Simplification

Another putative example of getting something for nothing is given by the organizational innovations associated with factory production. The demon in this case is the task simplifier. By breaking down complex production tasks into simpler subtasks requiring less skill than the original, it is possible to achieve dramatic improvements in productivity. Improvements result from savings in labor input, decreased waste, and a lower wage bill. At the dawn of the industrial revolution, Adam Smith illustrated the effectiveness of these organizational `tricks’ by comparing the productivity of a late eighteenth-century pin factory with that of a handicraft operation in the same period. But these gains too are not free.

No Free Lunch

As noted above, the trick in virtual organization is the categorical separation of abstract requirements from concrete satisfiers, which supports switching as a systematic management procedure. Neither this capability nor the capacity of Maxwell’s demon nor the task simplifier is free.

In the case of Maxwell’s demon, Leo Szilard showed that the entropy decrease of the closed system is equal to the information required by the demon to distinguish fast from slow-moving molecules (see Brillouin, 1962). The demon requires information to operate the trap door effectively, and the amount required–calculated according to Shannon’s uncertainty measure–equals the decrease in entropy of the physical system. So there really is no free lunch. A gain in one area is offset by an expenditure elsewhere. Nevertheless, within certain limits, the value added to the system by the demon may exceed the information costs.

Similarly, in Adam Smith’s pin factory, the organization of production into controlled sequences of relatively simple subtasks, introduces new overhead costs. In handicraft, coordination and management are integrated in the production itself. Factory methods, by differentiating tasks and deploying specialized labor, enlarge the roles of coordination and management and establish them as independent functions. Again, within certain limits, the relatively high costs of coordination and management in the factory are more than offset by decreases in production costs.

Virtual organization also incurs new overhead costs. These derive from new management activities and from the transactions associated with switching. New management activities are needed to organize activity virtually, i.e., to analyze abstract requirements (e.g., the components of a complex product like an automobile), and to track concrete satisfiers (e.g., potential suppliers of components). Moreover, each time an abstract requirement is reassigned to a new satisfier a transaction cost is incurred. This occurs, for example, when new suppliers are chosen for given components. Switching suppliers requires a number of administrative and logistic changes that take time and resources to complete.

As in the Maxwell demon and task simplification examples, the overhead costs incurred by virtual organization are small compared to the gains, so long as certain limits are observed. These limits are characteristic of a given production system, i.e., an organization in a particular market environment. Switching must be exercised within strict system boundaries in order to avoid excessive costs. These system boundaries are easy to understand in the case of a virtual memory computer system. Users of time-sharing computers, whose operating systems work with virtual memory, routinely experience these limits directly. When the number of users simultaneously logged on the system exceeds some critical value, the capacity of the time-sharing computer becomes completely absorbed by its operating system in switching information between primary and secondary storage. This behavior is called thrashing.

An organizational analog of thrashing behavior occurs when satisfiers (e.g., suppliers) are changed too frequently. Businesses try continually to take advantage of lower prices for required goods and services, but if suppliers are switched too often, the transaction costs in making these changes can nullify the advantages of the cost savings. Thus, to reap the benefits of virtual organization it is necessary to operate within certain boundaries.

SOCIO-TECHNICAL FOUNDATIONS OF VIRTUAL ORGANIZATION

Armed with a definition and some examples, we will now examine the foundations of virtual organization. This new organizational construct is supported by three socio-technical innovations: commoditization of information, standardization, and abstractification of wealth. These socio-technical innovations are essential to the realization and exploitation of virtual organization and thus we will examine each one in some detail.

Commoditization of Information

Computer-based information technology has opened the door to a new world of products and services distinguished by their capacity to furnish information. Although products and services of this type (e.g., books and consultants) have been around for a long time, the new technology has made it possible to extend, refine and multiply them to an extraordinary degree. We call these products and services (computer-based or not) information commodities. Generally speaking, computer-based information commodities are of two different types, namely, software and databases. Examples abound in practically every sphere of activity. Common, computer-based information commodities include: financial software, stock market quotation systems, securities trading software, on-line reservation systems for hotels and airlines, online databases, transaction processing systems, expert systems for insurance underwriting, expert system shells, etc. The list grows longer every day.

To give a precise definition of information commodity, the notions of `information’ and `commodity’ must first be specified. The definitions presented briefly below are from Mowshowitz (1992a).

Information is taken to be the ability of a goal-seeking system to decide or to control. Note that both deciding and controlling can be reduced to the more basic concept of choosing. To decide and to control mean simply to make a choice among a set of possibilities. The definition of information as ability rather than as a material thing or stuff provides a sound foundation for analyzing the economic value of information.

A commodity is an economic good that has a determinate value, and can be owned.

Combining these two concepts, we define an information commodity as a commodity that furnishes information (i.e., the ability to decide or control).

To avoid paradoxes in the economic analysis of information, it is essential to distinguish between information and an information commodity. They are two very different ideas, and only the latter is appropriate for economic analysis. Loosely speaking, one might say that an information commodity is information in a container, or a package with information. Although it is somewhat misleading to speak of information–a non-material ability–in a container, this locution helps to reinforce the distinction between information and information commodities.

The reason for this distinction is revealed in the commercial relations between buyers and sellers of so-called information. Consider a commodities trader, who deals in orange juice futures, hogbellies, soybeans or some other agricultural product. Suppose the trader has an informant in the Department of Agriculture, and that this informant provides inside information in advance of official announcements of crop forecasts. For concreteness, let us consider forecasts of the orange crop. If the trader knows in advance what the announcement will be, he can exploit that information to make a substantial profit. If he knows that the Department of Agriculture is going to announce that the current estimate is much higher than the previous estimate, then he knows he has to sell short or to purchase put options. If on the contrary, the crop estimate is to be lower than the previous one, he will know to buy futures contracts or purchase call options in anticipation of rising prices.

This vignette of the trader and the spy suggests a critical question, namely, how does the commodities trader figure out how much to pay his informant? On the one hand, there is no way for the trader to determine a fair price for the spy’s information before receiving the information itself because without it there is no way to estimate the potential profit to be made. On the other hand, once the information has been communicated to the trader, it loses its value to the spy; and, what is more, both the trader and the spy have it so technically speaking nothing has been taken from the spy. These two conclusions are mutually contradictory, and thus we have the makings of a paradox.

Since spying is an old profession, and agreements on payment are readily concluded in practice, the paradox suggests that something is amiss with the interpretation of the transaction. One must reconsider what is being exchanged. Is the trader considering payment for information or for a product (or service) that furnishes information? With the latter there is no paradox at all. If the commodities trader pays his informant not for specific information to be furnished, but on the basis of the informant’s qualities as a source of information, then it is possible for the trader to determine a fair price before getting the information. The determination would be made on essentially the same grounds as those used to estimate the value of any service, i.e., the record of the service provider’s past performance. Knowledge of the spy’s past performance would enable the trader to assess the economic value of his service. Thus, what the commodities trader is paying for is not the information per se but an information commodity, namely, the capacity of an informant to furnish information reliably on a regular basis.

This example demonstrates the importance of distinguishing between information and information commodities. The attempt to assign economic value to information gives rise to a paradox which is easily banished by heeding the distinction.

Commoditization of information is essential for implementing virtual organization because it facilitates human-independent uses of knowledge and skill. The process of commoditization typically occurs in three stages. First, the knowledge or skill needed by a human being to perform some task is analyzed and modeled; then, the knowledge or skill is incorporated in an artifact such as a computer program; and, finally, the artifact is turned into a commercial product or service. This process has given rise to an extensive and rapidly growing information market.

The dependence of virtual organization on the information market is analogous to that of industrial (factory) organization on the labor market. Industry must have a pool of readily available workers to meet its various skill requirements. The labor market provides this pool efficiently by offering choices for most skills; i.e., a number of different workers with essentially equivalent capabilities can be hired to satisfy a given skill requirement. Similarly, virtual organization must have a pool of commercially available knowledge artifacts to enable it to reap the advantages of switching. The mapping of abstract requirements to concrete satisfiers can be changed without incurring excessive transaction costs only if production processes can be coupled and uncoupled easily and quickly. Knowledge and skill provided by computer-based information commodities make this possible. The coupling and uncoupling of production processes dependent on human labor are too cumbersome and costly to allow for the systematic use of switching.

Standardization

The meaning of standardization and its importance to industrial development are well known. Like commoditization of information, and for many of the same reasons, it is necessary for the realization of virtual organization. Standardization facilitates switching by making it possible to couple and uncouple production processes.

Virtual organization is a new arena for standardization. The establishment of standards for tools, products and processes is just as important in the new arena as in the older industrial environment. But there is more in virtual organization, namely, standardization in the social domain. This applies to organizations and to individual human beings.

Metamanagement requires standardized organizational structure and behavior to achieve interchangeability and compatibility. Interchangeability facilitates the replacement of one organizational unit by another with essentially the same functionality: compatibility enables two different units, with a minimum of effort, to interact with each other in the performance of a common task.

Both of these aspects of organizational standardization are essential to switching. As noted before, unless abstract requirements can be reassigned to concrete satisfiers smoothly and easily, the transaction costs of switching could nullify the benefits. Interchangeability makes it possible to select any one of several equivalent units, services or suppliers to meet some organizational requirement. Compatibility allows a given unit to switch easily from one cooperating partner to another in performing its function. That is to say, a standardized, organizational interface minimizes the problems of disengaging from one partner and reconnecting to another. The use of electronic data interchange (EDI) for computer-to-computer exchange of structured business forms creates de facto organizational compatibility standards, or, as expressed by Kalakota and Whinston (1996, p. 334), `boundaryless relationships’. First, standards for electronic communication are introduced, and then organizational standards designed to exploit the cost-saving possibilities of the new form of communication come into being. For example, some corporations (e.g., General Motors) and government agencies (e.g., United States Department of the Treasury) have turned to EDI to stem the rising costs of paper-based transactions. General Motors was one of the first large corporations to recognize the possibility of achieving enormous costs savings by using electronic rather than paper business forms in transactions (e.g., ordering, invoicing, making payments) with its many suppliers. Similarly, the United States Treasury recognized the opportunity of using EDI to reduce the paperwork costs incurred in the Federal Government’s interaction with its many contractors. But the deployment of EDI implies much more than simply installing hardware, software and netware to facilitate the electronic exchange of structured information. Organizational structures and practices–in all the cooperating organizations–must adapt to the needs of the new EDI systems, and these accommodating structures and practices define de facto compatibility standards.

These standards facilitate switching and reduce its transaction costs by simplifying the organizational processes of unit substitution and coupling-decoupling. Organizations and firms can achieve competitive advantages by making use of the standards–therein lies the incentive for adoption. The tendency in large corporations to permit various units to function as relatively independent cost or profit centers sets the stage for adoption of standards and encourages experimentation with switching. As standardization advances, such switching will be practiced more systematically.

Virtual organization also calls for standardization of individual behavior. Desirability aside, such standardization is certainly not new in human history. Language, shared culture, rituals, social conventions, and codes of conduct are an different forms of standardization which are essential to many types of social interaction. The new paradigm requires of individuals the capacity to move freely from one organizational unit or setting to another. Movement within large organizations is a time-honored practice, especially in the careers of managers who may be expected to become familiar with many different aspects of a business. Unique to virtual organization is the need for behavioral standards to facilitate switching. When one unit is substituted for another, or decoupled from one and recoupled to another, some individuals must adapt to new environments. If the affected individuals have to internalize entirely new behavioral codes and procedures, the costs incurred in such adaptations could be considerable. Behavioral standards limit these costs.

Generally speaking, standardization can be achieved in two ways. One approach is to deploy a universal standard by common consent, and make it compulsory for all parties who would communicate and interact with each other. Another approach, made feasible by advances in distributed systems and networks, 4 to permit the use of multiple standards or protocols and to rely on intermediaries (e.g., computer programs) to translate from one protocol to another. EDI provides an example of the latter approach. Different industries have developed different standards for the same business forms, and protocol conversion programs are used to facilitate the exchange of business forms between companies with different EDI standards. The two approaches are not incompatible. In some cases it makes more sense to adopt multiple standards and in others to insist on a uniform standard.

Abstractification of Wealth

The third socio-technical innovation underpinning virtual organization consists of the various forms of abstract wealth that make it possible to mobilize, manipulate and own distributed resources. In addition to the commoditization of knowledge, the switching principle of virtual organization calls for the capacity to shift wealth or assets freely from one form to another and from one place to another. These shifts are occasioned by the ever-changing economic scene–fluctuations in supply and demand for products and services in different markets, everchanging exchange rates and borrowing costs, etc. The need to manage and exploit such fluctuations has fathered the invention of a host of new forms of abstract wealth designed to meet cash flow requirements, leverage capital, or hedge risk.

Abstract wealth can only be understood in relation to concrete wealth. The latter consists of land, buildings, natural resources, machinery and so forth. Familiar examples of abstract wealth are money, stocks, leases, futures contracts, options, etc. A key feature of abstract wealth is that new forms can be created at will. One important way in which this is done is through the process of securitization, i.e., the transformation of an asset or a bundle of assets into a security.

This process can be defined with the help of a `securitization operator’ denoted here by S. If x represents wealth (e.g., the assets of a company), then S(x) can be interpreted as the collection of shares defining the ownership of that company. Theoretically, securitization can be carried to any level of abstraction. This is shown by the following recursive definition.

Concrete wealth is wealth.

If w is wealth, then so is S(w).

Applying the securitization operator S to concrete wealth c gives S(c), a form of abstract wealth. Applying S to this result yields S(S(c)), a yet more abstract form of wealth. Clearly, there is no limit to the recursive application of the securitization operator. For example, a mortgage on real property is an abstract form of wealth–i.e., a financial instrument entitling its owner to certain revenues–based on land. Consider a collection of mortgages on various properties. Such a collection can serve as an asset–abstract wealth in this case–which can be securitized to yield a new form of abstract wealth. This is not a hypothetical example since mortgage-backed securities have been traded in the marketplace for some years now. The same holds for securitized collections of leases. Now, these securities could themselves be bundled together to serve as assets to be securitized, and so on.

Abstract wealth, like commoditization and standardization, is essential to virtual organization. In particular, abstract wealth supports the transfer of resource ownership on a global basis, and thus facilitates switching.

We pointed out earlier that switching has limitations analogous to the thrashing behavior of a virtual memory computer system. The ability to abstractify wealth implies related limitations. One source of these limitations is the need for formal record keeping.

Ownership is a social and legal concept. In a simpler period of human history people lived in relatively small and stable communities. In such communities, ownership could be determined on the basis of personal knowledge; i.e., person x could be recognized and acknowledged directly as the owner of some concrete wealth. As communities grew in size, differentiated into subcommunities, and generally increased in complexity, direct means of identifying and acknowledging ownership came to be supplemented by formal records. The need for such formal records has increased dramatically with the elaboration of abstract forms of wealth. A bank deposit with First Cosmic Bank, for example, exists only as a notation in First Cosmic’s record system. There is no segregated pile of money anywhere with the owner’s name attached, and it is quite possible in this age of automated teller machines that no one in the bank would be able to identify the owner.

Dependence on formal records to determine ownership means additional costs and vulnerabilities in the conduct of business. Costs are incurred in the data gathering and management of record systems: vulnerabilities arise from the possibility of error and deliberate fraud. Thus, these systems act as a brake on virtual organization by reducing the potential gains from switching.

CONCLUSION

The foregoing remarks sketch a theory of virtual organization and point to the key sociotechnical innovations that make it possible to realize the idea in practice. Further research is needed to demonstrate (quantitatively as well as qualitatively) the advantages and limitations of virtual organization.

(1) This article is an outgrowth of a plenary lecture prepared for delivery at the Annual Meeting of the ISSS in 1996 at the Vrije Universiteit, Amsterdam, and of a lecture presented at the Symposium on the Automated Society at Universiteit Twente, The Netherlands in 1993.

REFERENCES

Brillouin, L. (1962). Science and Information Theory, Academic Press, New York.

Davidow, W. H., and Malone, M. S. (1992). The Virtual Corporation, HarperCollins, New York.

Giuliano, V. E. (1982). The mechanization of office work. Scientific American 247, 149-164.

Hammer, M., and Champy, J. (1993). Reengineering the Corporation, HarperCollins, New York.

Harrington, J. (1991). Organizational Structure and Information Technology, Prentice-Hall International, Hertfordshire, UK.

Hiltz, S. R. (1986). The virtual classroom: using computer-mediated communication for university teaching. Journal of Communication 36, 95-104.

Kalakota, R., and Whinston, A. B. (1996). Frontiers of Electronic Commerce, Addison-Wesley, Reading, MA.

Kleene, S. C. (1952). Introduction to Metamathematics, Van Nostrand, Princeton.

Mowshowitz, A. (1992a). On the market value of information commodities. I. The nature of information and information commodities. Journal of the American Society for Information Science 43, 225-232.

Mowshowitz, A. (1992b) Virtual feudalism: a vision of political organization in the information age. Information and the Public Sector 2, 213-231.

Mowshowitz, A. (1994). Virtual organization: a vision of management in the information age. The Information Society 10, 267-288.

Mumford, L. (1934). Technics and Civilization, Harcourt Brace Jovanovich, New York.

Rheingold, H. (1993). The Virtual Community, Addison-Wesley, Reading, MA.

Turoff, M. (1985). Information, value, and the internal marketplace. Technological Forecasting and Social Change 27, 357-373.

Abbe Mowshowitz Department of Computer Science, City College of New York, Convent Avenue at 138th Street, New York, New York 10031, USA

Correspondence to: Abbe Mowshowitz, 348 Wagner Avenue, Mamaroneck, NY 10543, USA.