Getting New Things Done

Networks, Brokerage, and the Assembly of Innovative Action

David Obstfeld

• Publisher: Stanford Business Books
• Year: 2017
• ISBN:
  9781503603097
• Edition: 1
• Topic: Betriebswirtschaft
• Subtopic: Organizational Behavior

Our networks—and how we work them—create vital ties that bind. Organizations recognize and reward this fact by leaning ever more heavily on collaboration, particularly when it comes to getting new things done. This book offers a framework that explains how innovators use network processes to broker knowledge and mobilize action.

How well they do so directly influences the outcome of attempts to innovate, especially when a project is not tied to prescribed organizational routines. An entrepreneur launches a business. A company rolls out a new product line. Two firms form a partnership. These instances and many more like them dot today's business landscape. And yet, we understand little about the social dimension of these undertakings. Disentangling brokerage from network structure and building on his theoretical work regarding tertius iungens, David Obstfeld explains how actors with diverse interests, expertise, and skills leverage their personal and intellectual connections to create new ventures and products with extraordinary results.

1

Brokerage in Action

Determining how actors get new things done requires us to consider both social network structure and social process. As described in the Introduction, this kind of collaborative action grows out of brokerage, the process by which actors are strategically assembled and coordinated, and by which—in some cases—participation is expanded.

Concepts of social network structure provide the architecture and accounting of such activity, while concepts regarding social process explain the formation and aggregation of ties among people or groups (or the absence thereof). Both are of fundamental importance to social network theorists. Yet in social network research, the study of social process is often subordinated to structure, because structure is easier to measure and, in many cases, process is inferred from network data. Such a stance diverts attention from theorizing about network process. It is that imbalance that I seek to address. Building on an emerging stream of research, I will identify various network processes that are important in dynamic organizational phenomena, such as technological innovation and entrepreneurship (Obstfeld 2005; Vissa 2012; Long Lingo and O’Mahony 2010; Davis and Eisenhardt 2011; Bizzi and Langley 2012).

Given its central place in organizing, brokerage requires a sound theoretical foundation. The traditional social network perspective, as noted in the Introduction, defines brokerage in terms of open networks, open triads, or structural holes (Burt 1992). The open triad represents the most “extreme” challenge that might confront the broker: coordinating two parties who were heretofore unconnected and who may have come from unconnected social worlds. But as I will discuss, the coordinative challenge found in the open triad also occurs in closed triads, just as the ongoing challenge to establish intersubjective meaning is found in all dyadic communication (Weick 1979; Parsons 1951; Luhmann 1995).¹ Consequently, brokerage phenomena are found in all groups of three or greater—whether in open or closed networks—and in the more complex networks that combine open and closed characteristics that are most often found in the real world.

This chapter addresses social network structure and process to explain how brokerage functions to get new things done. I present this perspective in a sequence of five steps. First, I frame innovative action as often unfolding in triads through brokerage. I am aware that this may strike many network scholars as obvious, but I believe that it is an “obviousness” worth reviewing, as the triad is so often and so easily overlooked. Second, I explain how network structure sets the context for action, by summarizing some of the well-established ways in which network theorists analyze social network structure, emphasizing the distinction between open and closed social network structures. Third, I clearly distinguish between brokerage as action and brokerage as structure. Fourth, I consider in more depth the brokerage process, first by defining it and then by proposing three fundamental brokerage orientations or behaviors: conduit, tertius gaudens, and tertius iungens. Finally, I revisit Fligstein’s idea of social skill, or the ability to induce cooperation (Fligstein and McAdam 2012; Fligstein 1997, 2001), to argue that inducing cooperation to get new things done is achieved by strategically combining the three brokerage orientations toward action.

Innovative Action Unfolds in Invisible Triads

For the network theorist, the role of the broker may seem obvious, and yet the richness, ubiquity, and importance of triadic or coordinative phenomena may be easily overlooked. The coordination of the broker, whether momentary or ongoing, is often forgotten or overlooked altogether by the parties being connected, by observers outside of the coordinative unit of action, and even by the broker himself. We often note the formal, routinized role of brokers in facilitating transactions such as the sale of stock or the transfer of real estate, but we overlook the deceptively complex coordinative work associated with bringing parties together. If we focus our attention on the informal, “in-between” broker roles in which people continually engage, the number of cases to consider increases dramatically. A parent serves as a go-between in the tense relationship between her spouse and her teenage son.² A biology professor brokers between her students and the scientific community. A head of state endeavors to resolve a point of tension between two other heads of state.

The term “invisible work” (Star and Strauss 1999; M. C. Suchman 1995) has been used to denote how certain kinds of coordinative, uncounted, or uncountable work may be ignored, devalued, or dismissed altogether.³ Conversely, work that is more closely associated with established forms of expertise and status is more likely to be registered, counted, and valued. There is a rough analogue here to the attention that social network analysis pays to structure, which is easier to measure as compared to network process. Triadic coordination, too, is more difficult to observe because it is more spread out over time and space as compared to more localized individual or dyadic phenomena. More simply put, triadic social processes are inherently more complex and, therefore, more difficult for either the casual observer or the social scientist to detect and track.

The broker’s formal and informal coordinative work—consequential yet hard to detect—extends to firms, organizations, or larger social entities such as nations. For example, the venture capital firm links an entrepreneurial organization to investors and other sources of talent the new firm may need to engage. Similarly, brokerage relationships connect the supply chain, in which each firm in the chain brokers between its suppliers and downstream buyers by transforming inputs into outputs, and thereby adding value. In this sense, all commerce is inherently triadic, from Wal-Mart to the hot dog vendor. Mergers and acquisitions may appear to be dyadic pairings of the acquiring and acquired, but in fact they do not proceed without the intercession of brokers before the deal (to locate the participating parties and negotiate terms), during the deal (to plan and facilitate the strategic and operational integration), and after it (to knit together information technology departments and reconcile organizational charts and cultures). The network theorist emphasizes the ubiquity and importance of these triadic patterns of coordinative action, in contrast, for example, to Weick’s dyadic interacts.

As the merger and acquisition example above suggests, brokerage also involves mixed cases of individual and organizational actors. In the example, we find that individuals constitute a critical unit of analysis that stands between two or more merging organizations to facilitate the knitting together of the senior officer, operations, information technology, and marketing clusters as well as two organizational cultures and strategic visions. Individuals, either employed by the merging firms or consultants, reconcile, bridge, and knit together various aspects of the merging organizations. In a similar spirit, a diplomat brokers between individuals and organizations, governmental or otherwise, from the host country, on the one hand, and a corresponding set of entities from his or her country of origin, on the other, in countless variations on a basic triadic scheme.

Each of these coordinative feats involves different degrees of effort, attention, local knowledge, experience, and skill. Although all brokerage action involves agency exerted over time, there are distinctions to be made between the kind of brokerage that occurs largely within a routine versus that which occurs within a creative project (Obstfeld 2012), a distinction I will explore in more depth in Chapter 3. Connecting an interested stock buyer and seller is nearly automatic as there exists an active ongoing market mechanism for supporting such transactions, whereas facilitating a productive meeting between two embattled spouses or unfriendly heads of state is far more difficult. Expectations between the parties to the stock sale are commensurable, and sale procedures are highly routinized. The meeting between the heads of state, on the other hand, might confront not only their deeply rooted antipathy and distrust but, quite likely, opposition to their dialogue from many stakeholders that they represent.

In many cases, the brokerage triad (i.e., the broker and two alters) is less visible than an individual’s reputation, talent, credentials, or the more easily grasped dynamics of dyadic exchange. The (seemingly) dyadic exchange of a job interview obscures the preceding introduction that made the interview possible. We move to the foreground the most immediately apparent dyadic pairing, but we do so at the expense of the brokerage activity that made that pairing possible.

Network Structure Sets the Context for Action

A key objective of social network analysis is to get at how larger patterns of ties enable and constrain action (Gulati and Srivastava, 2014). The first independent variable in the BKAP model is the broad category of social network structure—in other words, the formal “plumbing” or pipes (Podolny 2001) that define relationships among social actors. The basic building block of social networks is the relationship, or tie, between actors. From this fundamental unit, patterns of ties can be identified that reveal opportunities for and constraints upon action.

Ties vary by their strength (i.e., strong, weak, or somewhere in between) and their content (e.g., friendship or work collaboration). One has close friends and work colleagues and, less frequently, close friends at work, and all manner of variations on these parameters, including relationships that feature other tie content, such as advice, mentoring, and information exchange. Ties often involve multiple forms of content, or multiplex ties, as in the case of a tie involving both friendship and work collaboration. Various network measures use tie data to evaluate different facets of network structure. Network degree centrality, for example, is a relatively simple measure of the number of ties that a given actor has—an actor with more ties has greater centrality. Betweenness centrality (Freeman 1977) measures the number of times an actor stands on the shortest path between two other actors in the network.

Open versus Closed Network Structures

The open versus closed network is an analytic dimension of particular relevance to how brokers access knowledge and mobilize action to get new things done. The well-established debate around the relative merits of open and closed networks (e.g., Marsden 1982; Burt 1992, 2004; Krackhardt 1999; Ahuja 2000) bears directly on considerations of network structure and process. Open networks, featuring the absence of connections among those in the network, as noted earlier, afford greater access to novel, nonredundant information and greater discretion to act (i.e., less normative control). Earlier, seminal network research associated with the flow of novel information concerned related but not equivalent network structures involving weak instead of strong ties (Granovetter 1973). Closed or dense networks allow for more efficient communication, greater trust, social support, and normative control (Uzzi 1997; Hansen 1999; Coleman 1988). Coleman (1988) illustrates the virtues of such closed networks with the example of a mother who moves from Detroit to Jerusalem because, among other reasons, her unattended children at play will be looked after by adults in the vicinity: an advantage of a cohesive network at a community level. Similarly, Coleman (1990) illustrates the mutual trust characteristic of certain dense networks by citing the willingness of tightly knit London diamond dealers to exchange diamonds for examination without contracts or documentation.⁴

These two social network types pose different opportunities and challenges for two key aspects of combinatorial innovation: new ideas and the coordinated action to implement those ideas. Open networks present both a knowledge advantage and an action problem. By situating people at the confluence of different social domains, open networks create opportunities for accessing new ideas and combining them in new ways. These same networks, however, may create challenges for acting on such ideas, because the dispersed, unconnected people across different social domains are inherently more difficult to mobilize or coordinate, especially around novel ideas.

Closed or dense networks, conversely, reduce the obstacles to initiating coordinated action necessary to implement innovation (an action advantage) but pose greater obstacles to the generation of new ideas (a knowledge problem). Networks that are closed or dense are conducive to mobilized action, because interests and perspectives are prealigned or normatively constrained, and the language and trust necessary to mobilize those interests are more readily available (Granovetter 2005). Dense networks, while presenting optimal conditions for the efficient exchange of the complex but shared knowledge necessary for innovation in complex organizations (Uzzi 1997; Hansen 1999), present a knowledge problem, because of the redundancy of information circulating within the network (Granovetter 1973).

A variation on the simple open-closed distinction involves what I will refer to as hybrid cases. To understand the hybrid case, consider first the strongest example of the cohesive, dense network at a triadic level, consisting of what David Krackhardt (1999) refers to as Simmelian triads: those in which each of the three actors involved have strong reciprocal ties to one another. Now, by contrast, consider variations on the closed triad, in which two of the actors exhibit weaker ties with each other, perhaps reflecting newly formed relationships. Finally, consider a larger network with a focal actor (often referred to in the social network literature as “ego”) who has many more alters than the two that would constitute a triad, and blends weak, strong, and absent ties of diverse content. These blended networks serve as the basis for most network research from which we must extract insight regarding how various social processes drive meaningful outcomes.

An important line of social network research, for instance, explores how a combination of network cohesion (i.e., a measure similar to density) and range (i.e., ties to different knowledge pools) facilitates knowledge transfer and diffusion, and thus represents an important form of social network advantage (Reagans and Zuckerman 2001; Burt 2002; Gargiulo and Rus 2002; Reagans, Zuckerman, and McEvily 2004; Reagans and McEvily 2008; Srivastava, 2015). Fleming, Mingo, and Chen (2007), while finding evidence for the advantage of open networks for supporting the novel combinations that define creativity, also present evidence that closed networks accelerate the sharing of data and feedback, along with better distributed understanding and ownership of ideas. As we press the open-closed distinction to its limits, we eventually bump up against the need to speak to the action that takes place within hybrid networks that display both open and closed properties.

The limits to the open versus closed social network distinction are suggested by what Aral and Van Alstyne (2011) refer to as the diversity-bandwidth tradeoff. While open networks may provide access to more diverse information and serve as a source of more novel information, Aral and Van Alstyne point out that the weak ties often found in open networks may only support the exchange of “simple news.” In contrast, they find that the stronger, higher-bandwidth ties found more frequently in dense networks and associated with greater mutual understanding, more frequent contact, and higher trust, may actually foster greater exchange of richer, more complex, and potentially more sensitive information, and consequently greater learning. This is consistent with Coleman’s (1990) argument that strong and reciprocal ties with others increase access to information. Similarly, Ter Wal et al. (2016) found that open networks combined with shared professional specialties that provided similar knowledge and experience helped interpret diverse information, and alternatively, the limitations of closed networks can be augmented by shared third parties who can correct for overly narrow interpretations (see also Reagans and McEvily 2003; Rodan and Galunic 2004).

These recent empirical studies point out the limitations of si...