How a Flexible Matrix Structure Could Create Ambidexterity at the Macro Level of Large, Complex Organizations Like MNCs.

Date01 Junio 2020
AuthorEgelhoff, William G.

1 Introduction

In a recent book Egelhoff and Wolf (2017) identify and describe the concept of a flexible matrix structure for organizing and managing large, complex organizations like multinational corporations (MNCs). They suggest that such a structure could make MNCs more ambidextrous. The purpose of this conceptual article is to explore and further develop the proposed relationship between a flexible matrix structure and organizational ambidexterity.

Ambidexterity, as it relates to organizations, can be defined as "an organization's ability to be aligned and efficient in its management of today's business demands while simultaneously being adaptive to changes in the environment," (Raisch and Birkenshaw 2008, p. 375). In other words, ambidexterity means winning in both the short run and the long run--in the short run through the efficient implementation of strategy and in the long run through strategic change that adapts to a changing competitive environment. Ambidexterity would be unimportant if competitive environments were stable and never changed. But competitive environments tend to be extremely dynamic, and most initially successful firms go out of business within two decades of their founding. As a result, ambidexterity is a capability that firms need to possess in order to survive over time. It is a firm's ability to regenerate and remake itself--its strategy and organizational design--to fit a changing competitive environment that count in the long run. But the long run is a moot question if a firm cannot first succeed in the short run.

The present article addresses the creation of ambidexterity at the macro level of large, complex organizations like MNCs. It conceptualizes a specific model for developing more useful and more deployable ambidexterity at this level. Large MNCs generally embrace high levels of product, geographic, and functional diversity and knowledge. When they organize this diverse knowledge using a two-dimensional matrix structure, there is the potential to recombine the two bodies of knowledge above the level of the individual product divisions, geographical regions, or functional divisions, and thereby create new macro-level (firm-level) advantages that can subsequently be used by the lower-level strategies of the product divisions and geographical regions (we will refer to these as business-level strategies). Such macro-level advantages are unavailable to firms that lack either the necessary size and complexity or an organization design capable of organizing and appropriately recombining the diverse bodies of knowledge. We will argue that a flexible matrix structure is the best design for this task.

The next section introduces the concept of a flexible matrix structure, as this concept is developed and described in Egelhoff and Wolf (2017). A subsequent section discusses organizational ambidexterity and describes how a flexible matrix structure can facilitate such ambidexterity. The following section attempts to more strongly and insightfully conceptualize the relationship between a flexible matrix structure and ambidexterity. It extends existing theory, develops propositions, and identifies underlying assumptions. A brief concluding section provides guidance for future research and stresses the importance of better understanding and exploiting the potential associated with organizational size and complexity.

2 The Concept of a Flexible Matrix Structure

The matrix structure is a poorly understood and often under-appreciated form of organization design (Galbraith 2009). Egelhoff and Wolf (2017) use an information-processing perspective of organization design to evaluate and better understand matrix structures and their capabilities. This perspective views the organization as an information-processing system. Information processing includes the collecting of data, processing it into higher-level information and knowledge, storing and transferring such information and knowledge around the organization, and using it to make decisions that guide organizational behavior (Egelhoff 1982, 1991; Galbraith 1973; Tushman and Nadler 1978).

2.1 The Flexible Matrix Structure

Matrix structures are an overlaying of two or more elementary structures. The elementary structures available to MNCs include a worldwide functional division structure, an international division structure, a geographical region structure, and a worldwide product division structure. MNCs adopt a matrix structure because they want to compete with a more complex portfolio of strategies and differential advantages than they can develop and implement with a single type of hierarchical dimension (e.g., some advantages and strategies require product-specific knowledge and integration while others require country- and region-specific knowledge and integration). To support this complexity, a matrix structure provides higher levels of differentiation and specialization than an elementary structure. Under one of the above elementary structures, authority and communications tend to flow along a single primary hierarchy or structural dimension. Under the typical two-dimensional matrix structure, a foreign subsidiary simultaneously reports to the parent along two of the elementary dimensions. For example, it might report to a geographical region HQ and to a number of product division HQs. This example is illustrated in Fig. 1. The firm is a chemical company headquartered in the United States. Figure 1 shows the coordination or information-processing flows that occur in a matrix structure that contains a product division dimension and a geographical region dimension. The solid lines represent hierarchical information processing along each of the hierarchies. The dashed lines represent non-hierarchical network-like information processing between members and subunits of the two hierarchies.

The traditional matrix structure used by MNCs and other large firms has generally been conceptualized as a balanced matrix (Davis and Lawrence 1977). That is, both dimensions of the matrix have relatively equal power and influence over decision making in the firm. While the balance of power in a matrix often shifts from one dimension to the other, reflecting changes in strategies and environments, and the personalities of the relevant managers, the underlying assumption is that there is always a meaningful level of shared power and influence between the dimensions of a matrix. In Fig. 1 the solid and dashed lines together represent the information-processing flows that would occur under the classical balanced mode of decision making. To illustrate how decisions are made with this mode of decision making, consider the information processing around important decisions associated with the annual plan of the German plastics business shown in Fig. 1. Such information processing occurs hierarchically along both the product and the geographical dimensions, and important inconsistencies are reconciled by network information processing between the two dimensions. Usually the German plastics manager will be heavily involved in the reconciliation process as well.

In contrast to the balanced mode of decision making, Egelhoff and Wolf (2017) report that a rule-based mode of decision making seems to be emerging in some MNCs with matrix structures. Under this mode of decision making, formal and informal rules pre-assign different types of decisions to a specific dimension of the matrix for unitary decision making. For example, new product development decisions may be assigned to the product divisions, while pricing decisions for existing products may be assigned to the geographical regions. The rule-based mode of decision making operates differently from the balanced mode of decision making. The balanced mode seeks to reconcile the divergent viewpoints of the two hierarchies with overlapping responsibilities through the exchange of information, and when necessary escalation of the issue up the hierarchy. The first stage of the reconciliation process, however, is for the parties to exchange information and arguments in an ad hoc, non-hierarchical manner. This is shown in Fig. 1 with dashed lines between the two hierarchies. They seek a solution that best addresses the joint concerns of both hierarchies, and such a solution will frequently differ from what either hierarchy by itself would decide. In this sense, the joint solution is innovative, and not something either hierarchy acting alone could have produced. Ideally it reflects the concerns and knowledge possessed by both dimensions of the matrix and is therefore a richer outcome than the more limited outcome of a single hierarchy, and more likely to promote the overall interests of the firm. Its cost is the additional length of time and number of man-hours required to reach reconciliation plus the risk that such non-hierarchical reconciliation may not emerge. In contrast, the rule-based mode of decision making relies on formal or informal rules and standard processes to guide decisions to one hierarchy or the other for processing. As a result, there is less overlapping responsibility for decision making.

The two modes of decision making differ in the speed with which they can reach a decision, and also in the nature of their underlying information processing. A balanced mode of decision making with overlapping responsibilities leads to reciprocal information processing between the two dimensions, as they seek to reconcile their differences and converge around a decision. This can be a lengthy, exhaustive process, involving numerous meetings and exchanges of information. The rule-based mode of decision making seeks to make this process much faster and efficient. While both dimensions of the matrix can still input information into the decision process, one dimension has been given primary responsibility for making the decision. This leads to less back-and-forth discussion, and a more sequential...

Um weiterzulesen


VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT