Structure—Environment Coherence • 455

is to increase the possibilities for cooperative action in networks by reducing search, monitoring, and control costs without significantly increasing the transactions costs of interactions.

Why Are There So Few Structural Types?

The preceding discussion suggests a relatively small number of structural types from which a would-be organization designer might choose. At first glance, this might seem strange. Given the size and complexity of large firms, one might suspect that there could be a large number of design variables and many ways to organize. Yet this does not seem to happen. While there are a huge variety of firms and numerous organizing criteria, it also seems that, once industry particularities are taken into account, the range of organizational types employed in real firms is fairly constrained. One reason for the limited range of types might be imitation. As large, successful, and visible firms reach solutions regarding how they should organize, other firms may see what they do and what results they achieve and then imitate them. Indeed, as we have already mentioned, it was not uncommon after World War II to hear about the “General Motors” model of multidivisional organization as a norm to which other firms aspired. This explanation may have some value, but it is still hard to see how a small number of firms are so widely imitated across industries as to justify a small set of general types. Nor is it clear how pathbreaking firms identify these types in the first place.

John Roberts suggests that the reasons for the limited range of structural types may be more fundamental.23 Although a large number of design variables are available, choices from among these variables are not independent. Structural variables and related change activities are interdependent and thus may share complementarities, such that choosing more of one design variable increases the returns to choosing others. For example, organizing to pursue a product differentiation strategy based on product quality will be associated with lower volume performance, specialized staffing, frequent product redesign, and higher prices. Design variables may also be interdependent as substitutes, such that doing more of some activities reduces the value from doing more of others.

The ways in which design variables are interdependent determines the structural types that emerge. Complementarities among choice variables will result in a set of variables having a greater overall value for a firm than could be obtained from changing particular variables in isolation. It could even mean that changes in any one design variable might fail to contribute to firm performance and could even detract from it. This suggests that organization design choices will tend to cluster together and that a “mix and match” approach will prove infeasible. If this is so, it suggests that intermediate positions between structural types will not promote effectiveness and may actually reduce it, relative to a choice of one clear type or another.

STRUCTURE—ENVIRONMENT COHERENCE

Discussing organizational structure in the context of a firm’s strategy implies that some sort of optimal structure is identifiable and appropriate for the firm. That is indeed a presumption of much of what consultants and academics have contributed to this topic. But what does it mean for a structure to be optimal for a firm? Is it that an optimal structure will help a firm to make good decisions and to make them efficiently, with the least amount of time and effort relative to the quality of the decisions?

456 • Chapter 13 • Strategy and Structure

Making quality decisions efficiently does seem like a good idea. But what is a “quality” decision when discussing firms and their strategies? In this context, “quality” involves the requirement that decisions be effective in terms of the business and market conditions that the firm is facing. It would be hard to call a structure optimal if it led the firm to efficiently make decisions that did not fit well with the demands of competitors, consumers, or regulators.

Knowing that a structure is coherent is a first step toward solving the organizational design problem. It is also necessary for managers to determine whether a given structure is appropriate for them. A coherent structure will destroy rather than create value if it fails to enable the firm to apply its resources and capabilities to opportunities within its business environment. Thus, the optimal organizational structure for a firm depends on the environmental circumstances it faces.

A functional structure may work well for a manufacturer of supercomputers, such as Cray, but would probably work poorly for a large bank, such as JP Morgan Chase or Citigroup. This is because the banks face a much more complex business environment than Cray. Cray’s customers include large technology firms, businesses in sectors requiring extensive use of data, government agencies, and research universities. Cray and its customers know each other, and while the business environment has its own threats and opportunities, these are likely to be fairly predictable and understandable to all the participants in the industry. These are conditions in which a functional structure can work well.

The environment faced by large banks, on the contrary, is much more complex in terms of products, customers, and regulatory burdens than that faced by Cray. They operate on a global scale in multiple national markets. Even the technological needs of Chase or Citi are significant, as both have spent heavily in building their infrastructures. In addition, the volatility of their businesses is often high and, as the 2008 financial crisis demonstrated, also unpredictable. These are conditions in which large banks will face competing and often conflicting demands for organizing that cannot be easily handled within a traditional hierarchy. This is why these firms have frequently employed more complex organization structures, such as a matrix.

Contingency-based research examines how environmental characteristics may be associated with structural characteristics of firms without considering the product market choices that a firm might make. The idea is that any firm working in such an environment must adapt its structure to address environmental demands. This work has focused on two sets of environmental factors that may influence the relative efficiency of different structures: (1) technology and task interdependence and

(2) information processing.

Technology and Task Interdependence

Technology generally refers to the base of scientific knowledge underlying what a firm does, as well as the general state of know-how behind the application of scientific knowledge to specific products and services. While many firms invest in their own R&D to enhance their competitive positioning, most must take their technology as exogenous, at least in the short term. The characteristics of a firm’s knowledge base will influence the structural type that it adopts. A firm employing a well-known and mature technology will organize differently from a firm working with a rapidly evolving and less well-known technology. The former is better off adopting a hierarchical structure conducive to more stable, standardized, and higher volume production. The latter may prefer a decentralized structure that offers flexibility and responsiveness to change.

Structure—Environment Coherence • 457

As the firm’s technology changes, its structure will also need to change to accommodate new coordination needs. For example, suppose that a firm’s technology changes to permit increased production volume and more routine handling of raw materials. This firm might need to create a new division with purchasing responsibilities in order to accommodate the increased volume of activities. This idea is not confined to technology-driven changes in strategy. To take another example, a firm that expands internationally may need to set up a government affairs department that coordinates all regulatory related activities. As an example, Giulio Bottazzi and his colleagues document the interaction of these technological and regulatory factors and their implications for firms in their study of the evolution of the global pharmaceutical industry.24

James Thompson argues that technology determines the degree of task interdependence— the extent to which two or more positions depend on each other to do their own work.25 Thompson defines three modes of task interdependence: reciprocal, sequential, and pooled. Reciprocal interdependence exists when two or more workers or work groups depend on each other to do their work. Apple’s hardware and software development teams display reciprocal interdependence. Sequential interdependence exists between two or more workers, positions, or groups when one depends on the outcomes of the others, but not vice versa. The regulatory affairs and sales divisions of a pharmaceutical company display sequential interdependence. Messaging by sales depends on regulatory oversight on product labeling. Finally, pooled interdependence exists when two or more positions are not directly dependent on each other but are associated through their independent contributions to the success of the firm. For example, the success of Disney’s Pixar animated motion picture division is largely independent of the success of Disney’s ESPN sports network. This distinction suggests that organization design processes successively group individuals to coordinate their activities and utilize their shared resources. Positions and tasks that are reciprocally interdependent should be grouped together first, since they are the most costly to coordinate. Less costly to coordinate are positions that are sequentially dependent, such as positions at different points of the value chain. Finally, positions that lack a direct interdependence and are related solely in terms of a common affiliation with a firm are the least costly to coordinate.

As technology changes, so may the basis for competition in an industry by changing the industry’s core assets.26 This in turn will alter task interdependence within firms and thus suggest possible changes in what an appropriate structure is for affected firms. Advances in computers and telecommunications weaken reciprocal and sequential interdependence among many positions and reduce the costs of coordinating activities among individuals and groups within the firm or with partners in other firms. Sharing data in the Internet “cloud,” engineers and product specialists located on different continents can coordinate the design of a new product even while few managers of such a team ever meet face to face. With such technology, a small investment research firm based in Zurich can provide analysis to subscribers in Tokyo, while the firm’s marketing agent resides in London and its publicist in New York. Neither of these arrangements would have been possible 20 years ago. This reduction in coordination costs reduces the need for team members to be in the same part of the firm’s formal organization or even to be part of the same firm. (We also note that virtual firms appear to suffer from diseconomies of scale if they attempt to grow too large, since relationships among members become complex and costly to coordinate above a small number of individuals.)