Reaching and Acquiring Valuable Resources: The Entrepreneur's Use of Brokerage, Cohesion, and Embeddedness

Authors


  • We thank our editor Andreas Rauch and two anonymous reviewers for their invaluable guidance during the review process. We also thank Nicole Coviello, Waverly Ding, Benson Honig, Byungchae Jin, Moren Levesque, Matts Lingsblad, Louise Mors, J. Peter Murmann, and Tyge Payne for their helpful comments on the earlier versions of this manuscript. An earlier version of this paper won the Strategic Management stream's Best Paper Award in the 2011 ANZAM Conference. This paper has been supported in part by an internal faculty research grant at Australian School of Business.

Abstract

Entrepreneurs have two key aims in managing their ego-networks: extending reach to valuable resources and facilitating resource acquisition. This study provides a synthesis of the brokerage, cohesion, and embeddedness literatures to develop and present a multilevel theoretical framework and analytical model that treat both aims jointly. It makes three contributions. First, it highlights a trade-off that entrepreneurs face in allocating their available networking time and energy while pursuing these two aims. Second, it explores the central role of two types of embeddedness—relational and structural—in resolving this trade-off. Third, it helps entrepreneurs decide when to embed a particular dyadic connection relationally or structurally. We show that entrepreneurs can better balance their dual aim by structurally embedding some ties rather than trying to relationally embed all. The resultant network is one that meshes characteristics of brokerage and cohesive ego-network structures.

Introduction

How are some entrepreneurs able to obtain the resources they need for their ventures, while others struggle only to come up with little to show for their efforts? To answer this question, researchers have employed social capital theory and network analysis. They have examined how characteristics of entrepreneurs’ social relations and the structure of their social networks may help entrepreneurs either enhance their reach to or facilitate the acquisition of valuable resources (see reviews by Gedajlovic, Honig, Moore, Payne, & Wright, 2013; Hoang & Antoncic, 2003). Two distinct streams have gained prominence, offering guidance to entrepreneurs. However, the suggestions of one stream are sometimes at odds with the other. In this study, we integrate extant research into a coherent theoretical framework and analytical model. Our approach allows us to offer an interpretation of extant findings and show how entrepreneurs can make networking decisions to help them achieve their dual aim of reaching and acquiring valuable resources.

At the core of one prominent stream of research is Granovetter's (1985) strength of weak ties argument and Burt's (1992, 2005) structural holes theory. This stream advocates the benefits of brokerage ego-networks in extending the entrepreneur's reach to valuable resources (e.g., Sing, Hills, Hybels, & Lumpkin, 1999). The basis of the second prominent stream is Coleman's (1988) approach to social capital. This stream suggests that entrepreneurs should form cohesive ego-networks to improve the likelihood of acquiring more of the resources in their ego-networks (Elfring & Hulsink, 2007). To date, however, there is little consensus about the overarching benefits of either approach to the entrepreneurial process (Martinez & Aldrich, 2011; Stam, Arzlanian, & Elfring, 2014). This is largely because each focuses on only half of the problem, while essentially assuming away the other half. That is, studies that recommend brokerage ego-networks are concerned with enhancing reach to resources that are likely to be more valuable and take for granted their acquisition (see Hansen, 1999; Uzzi, 1996, for early discussions). Similarly, those that favor cohesive ego-networks primarily highlight their advantage in facilitating resource acquisition, while assuming that the acquired resources are valuable (see Burt, 2005, for a critique).

Consequently, these two streams offer confusing guidance on how entrepreneurs should develop and manage their ego-networks (Martinez & Aldrich, 2011). Recent reviews report that scholars seek to resolve this confusion by promoting a contingency approach that has resulted in a growing set of contingencies. For instance, whether a brokerage or cohesive ego-network is more appropriate is argued to depend upon: the life stage of the venture (Martinez & Aldrich); its age; type of industry or country in which it is located (e.g., high-tech or low-tech industry and developed or developing country); or even its expected outcomes (Stam et al., 2014).

In this study, we offer a solution that minimizes the need to revert to contingencies. We highlight the trade-offs encountered in the pursuit of the entrepreneur's dual aims of increasing reach to and ensuring acquisition of valuable resources. More importantly, we provide a way for the entrepreneur to achieve these dual aims at the ego-network level by managing the trade-offs on a relationship-by-relationship basis at the dyadic level. The solution is derived from a theoretical framework and an analytical model that integrate three literatures. As already mentioned, two are concerned with the ego-network level and advocate either brokerage or cohesive entrepreneurial ego-networks. The third one is the embeddedness literature at the dyad level (Krippner & Alvarez, 2007; Moran, 2005; Uzzi, 1996). We focus on two means of embedding a single connection—relationally and/or structurally. Relational embeddedness refers to the closeness of the two parties, while structural embeddedness is the number of third parties who oversee the relationship between the two. We discuss how these forms of embeddedness impact the entrepreneur's reach to and acquisition of valuable resources from the alter.

Combining dyad and overall network levels in a multilevel approach not only lays out the reach vs. acquisition trade-off but also guides the decisions involved in shaping the entrepreneur's ego-network (see Payne, Moore, Griffis, & Autry, 2011, for a recent call for such multilevel studies). This approach also allows us to contribute to the literature by clarifying steps in the entrepreneur's decision-making process. We elucidate how they can think about the trade-offs involved with alternative ways of both forming new connections and maintaining existing ones (also see Vissa, 2012).

We formalize our theoretical framework with an analytical model. In this literature where confusion exists due to “contrasting claims” and “ambiguity in interpreting research results” (Martinez & Aldrich, 2011, p. 8), analytical models would be especially invaluable, given their ability to account for and integrate, with some precision, diverse theories and findings. Analytical models also help scholars consider alternative elaborations while taking into account various factors that may influence how entrepreneurs shape their social networks.

Our analytical model serves four distinct purposes. First, it enables us to demonstrate how pursuing each of these dual aims one at a time rather than simultaneously can result in network structures that stymie the ability of entrepreneurs to achieve both aims. Second, it takes into account the limited time and energy entrepreneurs have for networking purposes (Greve & Salaff, 2003) and allows us to accommodate this “activity capacity constraint.” Accounting for this constraint addresses a recent call for researchers to consider not only benefits but also costs of relationships (Gedajlovic et al., 2013, p. 464). Third, given entrepreneurs' desire to both enhance reach to and ensure acquisition of valuable resources subject to an activity capacity constraint, our model helps us explore how entrepreneurs can simultaneously balance the pursuit of both aims. Finally, by considering both aspects of the reach vs. acquisition problem jointly, the model provides a foundation for future research. In particular, it offers guidance for following the suggestions of Martinez and Aldrich (2011) and Stam et al. (2014) to investigate how entrepreneurs can shape their personal networks in different contexts. Our findings further help scholars develop a more nuanced understanding of how entrepreneurial social networks may evolve and hint at the potential agentic role entrepreneurs play in this process.

In the next section, we review the importance of social networks to entrepreneurs and discuss how their social networks’ structure affects their reach to valuable resources. We then elaborate the mechanisms that facilitate acquisition of resources from their connections. Specifically, we consider how entrepreneurs can use relational and structural embeddedness as alternative means. Next, we present our theoretical synthesis and analytical model that combine the dual aims of enhancing reach and facilitating acquisition. In the third section, we present the results of the analytical model, discuss implications of our findings for entrepreneurship research and practice, and reinterpret a number of findings in the extant literature. The last section concludes and highlights potential venues for future research.

Social Networks, Resources, and Entrepreneurship

Entrepreneurs organize efforts needed to create and/or discover and exploit new opportunities. To do this, they need to marshal a wide range of resources. However, they rarely own all they may need (Garnsey, 1998). Locating and obtaining resources is a particular challenge for entrepreneurs (Baker & Nelson, 2005; Elfring & Hulsink, 2007). Many are priced above their meager means and some may even be unavailable via market transactions (Starr & MacMillan, 1990; Witt, 2004). Given their limited means and confronted with organizational and market failures (Floyd & Wooldridge, 1999; Kuratko, Ireland, Covin, & Hornsby, 2005), entrepreneurs depend more than most other economic actors upon their network of social relations as a crucial source of valuable resources (Anderson, Park, & Jack, 2007; Bates, 1997; Kim & Aldrich, 2005).

As is common in network analyses, the social network of the entrepreneur (i.e., his/her ego-network) refers to the web of connections that link the entrepreneur and those others to whom he or she is directly connected (the alters). These others may include suppliers (of goods and materials), professional service providers (including consultants, lawyers, and accountants), investors, customers, friends, and family. Some alters may also be entrepreneurs themselves. Resources are all tangible and intangible things entrepreneurs could potentially obtain from their connections. These include, but are not limited to, financial capital; manufacturing facilities; marketing capabilities; status and prestige; labor; support, advice, and ideas.

Before obtaining a particular resource from an alter, the entrepreneur must ensure a number of conditions are met (Cross & Parker, 2004; Zhong & Ozdemir, 2010). First, the resource must be something the entrepreneur demands; that is the resource should be valuable to him or her. Second, the resource should be within reach. The alter must possess or control access to the resource. Third, the alter must be willing and able to transfer the resource to, or share it with, the entrepreneur (Aldrich & Martinez, 2001). Unless these conditions are met, the entrepreneur is unlikely to derive much benefit from a connection with the alter. In other words, not all valuable resources within their ego-network are easily acquirable by entrepreneurs (Hansen, 1999). Nor are all acquired resources valuable to them (Burt, 2005). Of all the resources within their reach, only a subset is both valuable and acquirable.

An ego network structure that is appropriate for achieving one aim is often unsuitable for the other (Martinez & Aldrich, 2011). Simply put, the two aims of increasing reach to valuable resources and ensuring acquisition of resources require entrepreneurs to conduct networking activities that are at times conflicting. As such, investments in forming an ego-network suitable for one aim usually come at an opportunity cost of not achieving the other. Entrepreneurs regularly face this trade-off when evaluating alternative ego-network structures or thinking about how to design or change their personal networks. To accomplish both aims simultaneously, entrepreneurs need to carefully expend the limited time and energy they have for networking and relationship-building activities (Greve & Salaff, 2003).

Increasing Reach to Valuable Resources: Ego-Network Level Debate

Since their social networks are such crucial sources of resources, entrepreneurs strive to have an ego-network that allows them to extend their reach to valuable resources. A resource's value is indicated by its relevance, usefulness, accuracy and reliability, exclusivity, and non-redundancy (Bruderl & Preisendorfer, 1998; Casson & Della Giusta, 2007). The literature suggests two alternative ego-network structures for increasing reach to valuable resources: (1) cohesive and (2) brokerage. In cohesive ego-networks the entrepreneur's alters are directly connected to one another, forming a closed group. In brokerage ego-networks, these alters are not directly connected to one another; they may even be unaware of each other.

Greater alter interconnectedness in a cohesive ego-network brings forth both benefits and problems for the entrepreneur. Higher reliability and support are two potential benefits, whereas redundant access to similar resources and potential lack of autonomy in deploying them are two common drawbacks. Interconnectedness among alters gives the entrepreneur multiple ways to reach the same resource. This enables the entrepreneur to “verify” the resource and uncover potential problems. For example, Casson and Della Giusta (2007) suggest that cohesive ego-networks are useful for learning about a prospective employee or a supplier. The received information can be confirmed or disputed multiple times by the mutual contacts who know both the entrepreneur and the alter (Granovetter, 1985). Entrepreneurs with cohesive ego-networks can also leverage the fact that their connections know each other. They can gather them and benefit from their joint efforts to solve problems they may be facing. Anderson et al. (2007) recount the case of Paul, an entrepreneur, who reported that when he needs business advice, he discusses it in a meeting with three to four people he trusts. The support, useful feedback, and joint problem solving that cohesion enable are all valuable resources for entrepreneurs (Elfring & Hulsink, 2007; Krackhardt, 1992; Lechner & Dowling, 2003; Uzzi, 1996).

At the same time, interconnectedness makes alters interchangeable; their resources are likely similar or redundant. This higher likelihood of redundancy may imply a misallocation of the limited time and energy entrepreneurs devote to networking (Burt, 1992; Greve & Salaff, 2003). Interconnectedness may also inhibit entrepreneurs’ ability to freely use the resources they may acquire. Entrepreneurs with cohesive ego-networks have to organize their actions according to the group's norms of behavior or risk facing sanctions. This limits their ability to use resources in novel or unconventional ways (Coleman, 1988; Jack & Anderson, 2002). Anderson et al. (2007, p. 261) describe how interconnected clients limited behavioral options of one entrepreneur. As Conor, the entrepreneur, tells it: “two of the clients came from that first client. They all work together … if you cock-up … everybody knows it.” The cohesive ego-network formed by these interconnected alters puts Conor under a spotlight. If he deviates from the group's expectations or does something wrong, all three alters would learn of it and could easily punish him. As a result, lack of autonomy reduces one's flexibility in using resources once (or if) they are acquired and thereby reduces the value of alters’ resources to the entrepreneur.

In contrast, entrepreneurs with brokerage ego-networks can tap into different segments of the social fabric more efficiently (Burt, 2005; Ibarra, 1995). As a result, they have higher potential to obtain resources that are likely to be different, unique, and diverse, and hence more valuable to them. For example, in a 12-month study comparing entrepreneurs with brokerage vs. cohesive ego-networks, Sing et al. (1999) observed that the former identified more opportunities than the latter. Entrepreneurs with brokerage ego-networks also have greater freedom from normative constraints. Since the alters do not know each other, informing one another of the entrepreneur's actions is difficult. The absence of normative sanctioning gives entrepreneurs more latitude to use resources innovatively if those resources are acquired. This makes the diverse resources that they can reach through a brokerage ego-network even more valuable.

In sum, both cohesive and brokerage ego-networks have their own advantages and disadvantages (Anderson et al., 2007; Davidsson & Honig, 2003; Elfring & Hulsink, 2003). Entrepreneurs can reach and potentially obtain higher quality and verifiable, but more limited, resources through a cohesive ego-network. With a brokerage ego-network, they can extend their reach to a wider variety of non-redundant resources, try to obtain them earlier, and once (or if) acquired, use them with fewer constraints (Burt, 2005). However, when focusing only on the aim of extending reach to valuable resources, the literature predominantly sees brokerage ego-networks as the better alternative for entrepreneurs (Stam et al., 2014).

While the brokerage literature stresses the underlying potential value that comes from having access to various non-redundant resources, it underplays the conditions under which such resources are likely to be acquired. In essence, the resource acquisition problem is assumed away in the brokerage literature. As Krippner and Alvarez (2007) reiterate an observation by Uzzi (1997), “A network rich in structural holes is virtually all that is needed to induce information and resources to flow through the network like electric current through a circuit board” (see also Hansen, 1999; Podolny, 2001). Although we adopt the brokerage literature's arguments as our anchor for how entrepreneurs can enhance their reach to valuable resources, we consider how they can facilitate acquisition of those resources as well.

Ensuring Acquisition of Resources: Dyad-Level Mechanisms

The fact that an entrepreneur's connection possesses some valuable resource is not enough to guarantee that the resource will flow to the entrepreneur (Lin, 2001, 2010). For example, Steier and Greenwood (2000, p. 172) describe the trouble that entrepreneur Don Clark encountered in securing funding. Clark explains, “I was looking everywhere. I talked to maybe forty-fifty people. There were no takers.” In other words, despite Clark's ties to potential investors, which placed funding—a valuable resource—within his reach, he was unable to acquire this resource. In fact, acquiring the valuable resources their ego-network puts within reach is a common and crucial problem for entrepreneurs—a problem that can simply determine their venture's success.

To explore how entrepreneurs acquire resources from their alters, we draw on the embeddedness literature. Our starting point is the insight that alters with embedded ties to the entrepreneur have greater motivation and are typically more readily able to help (Granovetter, 1985). We focus on two forms of embeddedness: relational and structural (Gulati, 1998; Moran, 2005). In relationally embedded ties, the dyadic relationship between two parties is strong (Marsden & Campbell, 1984; Uzzi, 1997). Strength of the tie is posited to be a function of the level, frequency, and emotional intensity of interactions between them (Granovetter, 1973). Consequently, such ties require significant expenditure of time and energy (Bourdieu, 2002). Structurally embedded ties, on the other hand, can be less demanding. Structural embeddedness is extent to which the entrepreneur and the alter share ties to the same others (Feld, 1997; Granovetter, 1992; Moody & White, 2003). The structural embeddedness of their dyadic connection increases with the number of their mutual acquaintances.

Figure 1 illustrates how an entrepreneur can utilize relational or structural embeddedness to make an alter more willing and able to provide resources. Both forms of embeddedness work through two primary mechanisms: (1) by developing a mutual understanding between the alter and the entrepreneur and (2) by helping the alter develop the trust needed to be willing to aid the entrepreneur. Mutual understanding accrues either from an interpersonal understanding (when a tie is relationally embedded) or from a shared group identity (when the tie is structurally embedded). Trust also emerges in one of two types. In the case of relational embeddedness, the alter comes to see the entrepreneur as trustworthy. When their tie is structurally embedded, the alter expects the entrepreneur's behavior to be appropriately governed by the web of connections in which their tie is embedded. We discuss these two mechanisms in turn.

Figure 1.

Mechanisms Through Which Embeddedness Enhances Resource Acquisition

Mutual Understanding

The entrepreneur may be unable to acquire resources, unless the alter understands what the entrepreneur wants and why he/she needs it. One way to foster such a mutual understanding is to relationally embed his/her connection with the alter through frequent or close interactions. As Ian, an entrepreneur interviewed in Jack and Anderson (2002, p. 478, emphasis added) explains: as one's contacts “get to know you as an individual and on a personal level, you develop a rapport and you develop a business association.” The rapport that relational embeddedness engenders cultivates a shared context through which the entrepreneur better comprehends the nature of the alter's resources and can more easily explain what he/she wants (Krackhardt, 1992). Similarly, it helps the alter to understand the entrepreneur's situation and what he/she is trying to achieve (Jack, 2005). Relational embeddedness also results in common relationship-specific heuristics (Uzzi, 1997) that simplify communication, increase transmission efficiency, and promote mutual understanding.

Alternatively, the entrepreneur can develop a mutual understanding with the alter by structurally embedding their dyadic connection with ties to mutual acquaintances. Shared acquaintances give rise to familiarity and a group identity among all parties. Familiarity and group identity in turn make it easier for both the entrepreneur and the alter to adopt group norms and to situate knowledge in a common context (Hite & Hesterly, 2001; Ibarra, Kilduff, & Tsai, 2005). Mental maps, frameworks, and understanding shared by all within the group help both the alter and entrepreneur lessen the cognitive distance between them (Reagans & McEvily, 2008; Rowley, 1997). In sum, group identity and shared mental maps promote mutual understanding, which facilitates the flow of resources to the entrepreneur (Cook, Emerson, Gillmore, & Yamagishi, 1983; Hansen, 1999).

Trust

The alter often needs sufficient assurance that he will benefit or at least not be harmed by allowing his resources to flow to the entrepreneur. Cognizant of this, the entrepreneur seeks to convince the alter of her trustworthiness (see review by Rousseau, Sitkin, Burt, & Camerer, 1998). She does this by embedding their connection either relationally or structurally.

Granovetter (1992) highlights that all economic actors prefer to deal with those they have dealt with before. When a connection is relationally embedded, both parties are more familiar with each other. Relational embeddedness presents more opportunities for the alter to learn about the entrepreneur and makes it easier to assess her trustworthiness (Gabarro, 1978; Krackhardt, 1992). Such an informed perception of the trustworthiness of the entrepreneur (an interpersonal trust), in turn, increases the alter's willingness to share or transfer his resources. For example, John recalls the early days of his start-up as follows, “suppliers provided me with stock and allowed me to return what I didn't want. They knew me and knew I wasn't going to run off with their stuff” (Jack, 2005, p. 1245). In other words, John's suppliers decided to transfer their stock because they were convinced that John would not misuse the resource or misappropriate its value (also see Krackhardt, 1990; McEvily, Perrone, & Zaheer, 2003). This confidence came from the relational embeddedness of each supplier's individual connection with John.

The entrepreneur can also structurally embed her relationship with an alter to make it easier for the alter to assure himself that the entrepreneur will act appropriately (Moran, 2005). Structural embeddedness provides such assurance in three primary ways. First are the multiple references regarding the entrepreneur's trustworthiness. Second are the established group norms that direct all parties’ behavior. Third is the oversight that would alert all alters of any behavior violations, and the potential collective sanctions and reputation effects that would surely follow such misbehavior (Raub & Weesie, 1990). We discuss each briefly in turn.

Mutual acquaintances can vouch for the trustworthiness of the entrepreneur (Batjargal, 2007; Levin & Cross, 2004; Wong & Boh, 2010). Knowing this, the entrepreneurs can leverage existing ties between any mutual acquaintances and the alter to allow the alter to learn about her reputation and credibility. Thus, the entrepreneur can establish her trustworthiness without spending the considerable time and energy required to relationally embed her ties with the alter. Even when the alter remains doubtful, mutual acquaintances provide a low-cost way to monitor the behavior of the entrepreneur and alert relevant parties (Gulati & Gargiulo, 1999). The more public one's behavior, the greater the incentive to behave as expected to uphold his/her reputation (Greif, 1989; Tullock, 1985). This enforced incentive to “act socially” and subordinate individual desires (Leana & Van Buren, 1999) helps allay most of an alter's concerns regarding the entrepreneur. The alter is more likely to be assured that resources he may provide would not be used in any way that can potentially endanger him. In other words, the oversight of mutual acquaintances assures the alter that the entrepreneur will honor the norms of the relationship, even when he may not know the entrepreneur very well. This impersonal form of trust (i.e., trust in the conduct of the relationship rather than in the individual) nevertheless increases the willingness of the alter to help the entrepreneur.

Don Clark, founder of Z.I. Probes, describes his structurally embedded tie with a banker: “It was a good thing that we were dealing with this one banker. He was a friend of Mr. B's… . This guy, although I didn't realize it at the time, gave us a good deal of slack. I'd come in and sign a pile of $10,000 demand notes and he'd just leave them in his drawer. No questions. Nothing. I am sure he wouldn't have done it, if he didn't know Mr. B” (Steier & Greenwood, 2000, p. 174, emphasis in original). In this case, Mr. B, who was Clark's main investor and mentor, as well as the banker's friend, served as a reference for the entrepreneur and as a monitor of the entrepreneur's future actions. Both informed the banker's assessment that Clark would behave as expected. As a result, he loaned money to the entrepreneur and did so at good terms.

Entrepreneur's Activity Capacity: Actor-Level Constraints

Entrepreneurs engage in many different activities during the entrepreneurial process. For example, the Panel Study of Entrepreneurial Dynamics (PSED) points to 25 different activities that entrepreneurs may carry out (Reynolds, Carter, Gartner, & Greene, 2004). Besides obtaining resources, these include engaging in discussions about ideas; developing business concepts; applying for copyrights, patents, or trademarks; marketing and selling; developing and managing human capital; and scanning the environment for recent developments such as changes in customer preferences (see Carter, Gartner, & Reynolds, 1996, for the list).

Many of these activities must be performed simultaneously (Lichtenstein, Carter, Dooley, & Gartner, 2007) and this may impose scheduling difficulties for entrepreneurs. Some can be delegated to employees they may have, but others must be performed by the entrepreneur either alone or with others. Such activities, including those that necessitate networking, compete for the entrepreneurs’ time, energy, and attention. We call the limited time and energy allocated for networking, their “activity capacity.” It is crucial for entrepreneurs to efficiently utilize this capacity.

Relational and Structural Embeddedness: Compared and Contrasted

We now turn to presenting evidence that relational and structural embeddedness are alternative means for acquiring resources. We also explain how entrepreneurs’ choice of one form of embeddedness over the other is constrained by the time and energy available to embed a connection. A study of reciprocation likelihood between two venture capital firms illustrates the substitutability of structural and relational embeddedness for ensuring resource acquisition. Bothner, Meadow, and Ozdemir (2006) find two conditions that make reciprocation (a form of resource acquisition) more likely. One is when the two firms have previously completed many exchanges, i.e., when their tie is relationally embedded. The other is when their tie is structurally embedded by ties to common third parties who can oversee the connection between the two firms. The authors observed high reciprocation in structurally embedded ties even when there were few previously completed exchanges, i.e., even when the tie's relational embeddedness was low. In other words, resource acquisition is more likely through a connection that is embedded either relationally or structurally and the use of one makes the need for the other unnecessary.

In terms of time and energy demands, structurally embedding a tie is the less costly alternative. Its cost advantage stems from how the two forms of embeddedness differ in the mechanisms through which the alter's willingness and ability to transfer resources is developed. Relational embeddedness requires significant investments of time and energy by both parties so each can gain mutual understanding of and trust towards the other. Structural embeddedness, however, relies on the existence of shared ties with third parties and the mutual understanding and expectations in behavior that such ties engender. It facilitates third-party monitoring and their ability to sanction behavior, neither of which requires substantial time and energy investments by either the entrepreneur or the alter. Feld (1997) provides evidence that in order to ensure similar resource flows through a tie, structural embedding requires, on average, lower investments than relational embedding. His data, collected at two different time periods, reveal that as the number of ties to common third parties shared by an ego and an alter increased in period 1, they spent fewer hours together in period 2. Nevertheless, they still had similar levels of liking towards each other (a dependent variable). Bothner et al.'s (2006) findings further confirm that structural embeddedness can have its effect without a strong tie between the two. Similarly, Levin and Cross (2004) find that once trust is controlled for, resources also flow through weaker ties (in their case: knowledge). They also find, consistent with our argument, that trust mediates the relationship between strong ties and receipt of more useful knowledge. However, they do not explain how those connected by weak ties can come to be trusted. We contend that some form of embeddedness, other than relational, is likely at work. Structural embedding is one means for this trust to be built (in addition to developing a mutual understanding). At the same time, structural embeddedness allows the entrepreneur to keep the strength of the tie, and hence time and energy expended on it, low.

In sum, structurally embedded ties can be weak but still be effective in facilitating resource acquisition. Moreover, the time and energy cost advantage of weaker ties has long been widely accepted throughout the network literature (Boorman, 1975; Burt, 1992; Granovetter, 1973; Hansen, 1999). Hence, structural embeddedness can be an effective substitute for the more costly relational embeddedness. However, a structurally embedded tie is also likely to reduce the entrepreneur's reach to valuable resources. Reach is reduced because each direct tie linking any two alters forms a closed triad among the entrepreneur and the two alters and inevitably results in a more cohesive ego-network at the overall network level. As a result, although structurally embedded ties need less time and energy to ensure acquisition, they also decrease reach to valuable resources. The relative advantage and disadvantage of structurally embedded ties forms the kernel of our theoretical framework and analytical model that we develop next.

Theoretical Framework and Analytical Model

Figure 2 presents the theoretical framework. Our starting point is the following: The entrepreneur cannot benefit from an alter's valuable resource unless she is able to acquire it. Similarly, there is little benefit from acquiring low-value resources. Therefore, a particular connection's usefulness is the extent to which the entrepreneur can reach a valuable resource and whether she can facilitate the acquisition of this valuable resource by virtue of this connection.

Figure 2.

Theoretical Framework

The entrepreneur can use either relational or structural embeddedness or both to increase reach to valuable resources and/or ensure acquisition of resources from a particular alter. Either form of embeddedness increases the likelihood that the entrepreneur would acquire resources from the alter. However, the entrepreneur needs to expend more time and energy to ensure similar levels of resource flow when they relationally, rather than structurally, embed the tie. While structural embeddedness is less costly, it also reduces the potential value of the alter's resources, because those resources may also be available to the entrepreneur from one or more of their mutual acquaintances.

Formalizing this theoretical framework, the analytical model helps us answer our fundamental question: How should entrepreneurs manage their social relationships to efficiently acquire many valuable resources? Based on Figure 2, we start by modeling resource value and flow at the entrepreneur–alter dyad level. We then aggregate the dyad-level model to the ego-network level by summing the dyad-level equations over all alters. Next, we introduce activity capacity as a constraint into the full model. These two, the model and the constraint function, set up the decision problem the entrepreneur is trying to solve, which we solve using standard techniques. Finally, we discuss the results and implications in detail.

Analytical Model at the Entrepreneur–Alter Dyad Level

Following the notation and conventions of Wasserman and Faust (1994), we denote the strength of the tie between the entrepreneur E and the alter i with vEi ≥ 0, and the number of mutual acquaintances they share as zEi ∈ {0, 1, … , N − 2}, where N is the number of individuals in the ego-network (including the entrepreneur). We model the incremental benefit derived from a particular connection as a product of two likelihoods: the likelihood that the alter has a Resource of Value to the entrepreneur multiplied by the likelihood that the entrepreneur can Acquire the Resource from that alter. That is, ΔSCEi = val(zEi) × acq(vEi, zEi). Since the literature portrays both of these likelihoods as curvilinear rather than linear (e.g., Semrau & Werner, 2013), we adopt inline imageand acqEi(vEi, zEi) = 1 − exp(− τ1vEi − τ2vEizEi) as two tractable functional forms. These two functions mimic the arguments made in the literature and result in the following dyad-level analytical model: inline image.

The functional form for the likelihood that alter's resource is valuable to the entrepreneur (valEi) has its maximum (= 1) when they share no mutual acquaintances (i.e., zEi = 0). This likelihood decreases (to a minimum of 0) as the number of their mutual acquaintances (structural embeddedness) increases when the parameter β is positive. The functional form for the likelihood of acquiring that resource from the alter (acqEi) has its minimum (= 0) when the relationship does not exist (i.e., vEi = 0). This likelihood increases with the number of their mutual acquaintances (structural embeddedness) and/or the strength of the direct connection between the entrepreneur and the alter (relational embeddedness) to a maximum of 1.

Aggregating the Dyad-Level Model to the Entrepreneur's Ego-Network Level

The important issue for entrepreneurs, however, is not only the net contribution made by each of their contacts but also the total benefit contributed by all alters in their ego-network. Entrepreneurs aim to have an efficient ego-network: one that ensures an appropriate mix of enhanced reach to valuable resources and their successful acquisition without exceeding their activity capacity constraint. They achieve this by carefully choosing relational, structural, or both forms of embeddedness for each of their ties at the dyad level. We mirror this behavior and model the contribution of the entire ego-network by aggregating the dyad level. For this purpose, we follow the approach presented in Burt (1992). Here we also introduce the constraint into the model. Equation ((1)) presents this full analytical model at the ego-network level.

display math((1))

Clarifying the Parameters β, τ1, τ2, and CE

The parameter β allows us to specify how strongly structural embeddedness affects the entrepreneur's likelihood to reach a valuable resource from an alter. If β = 0, mutual contacts have no effect on the value of the alter's resource. In this case the entrepreneur would only be concerned with facilitating resource flow. For positive β values, however, mutual contacts decrease the likelihood that resources available from the alter are valuable. When β = 1, as is implicit in the brokerage literature, the entrepreneur cannot reach any additional valuable resource via her connection to the alter, beyond what she can already reach from their mutual contacts. Figure 3 illustrates the extent to which mutual acquaintances affect the likelihood of reaching a valuable resource from a particular alter for a range of β values.

Figure 3.

Likelihood of Alter's Resource Being Valuable to the Entrepreneur

The parameters τ1 and τ2 respectively specify the influence of tie strength (relational embeddedness) and the number of mutual contacts (structural embeddedness) on the likelihood that the entrepreneur will acquire resources from an alter. Figure 4 illustrates this effect for a set of τ1 and τ2 parameter combinations. At one extreme where τ1 ↑ ∞, τ2 ≥ 0, resource acquisition occurs automatically through any tie, regardless of its embeddedness. This approach of seeing connections as “pipes” and assuming that resources flow freely within a network is implicit in the brokerage literature. At another extreme where τ1 = 0, τ2 > 0, embeddedness matters but only in its structural form. The entrepreneur can acquire alter's resource only if they have mutual contacts (i.e., zEi > 0). The strength of the direct tie between the entrepreneur and the alter can be ignored.

Figure 4.

Likelihood of the Entrepreneur Acquiring a Resource From the Alter.

The parameter CE enables us to account for the entrepreneur's activity capacity. This is the total time and energy he/she is able to allocate to building and maintaining relationships. Empirical studies provide evidence not only for the existence of such a constraint but also for variation in the time and energy that entrepreneurs have available to commit to networking during the entrepreneurial process (Greve & Salaff, 2003; Hite, 2005; Semrau & Werner, 2013; Zhao & Aram, 1995). Although this constraint varies among entrepreneurs, all must decide how to efficiently allocate their activity capacity (CE) to reach valuable resources and acquire them.

Specification of Parameters and Accommodating Contextual Variations

The model is quite robust in accommodating a wide variety of contexts by adjusting its parameters. These contexts can include a diverse range of sought-after resources, external environments, network types, entrepreneurs, and alters. We illustrate this as we discuss each in turn.

Types of Resources

The following two examples of different resource types illustrate the range of values the β parameter can take. If an entrepreneur is seeking input on her business plan, the comments and feedback she will get from two alters is more likely to be similar when all three are interconnected. Hence, the entrepreneur would be less likely to reach additional valuable insights by connecting to both. In situations like this, the β parameter would be specified as close to 1. If, however, the entrepreneur is looking for financial support for her venture idea, the support (i.e., money) she would receive from two interconnected alters, although similar, would complement each other. The fact that one alter provides financial capital will not decrease the value of a similar investment from the other alter. For resources like money, fidelity, or status, where more of the same resource does not necessarily reduce its value, the β parameter would be closer to 0 (also see Podolny & Baron, 1997).

Different levels of τ1 and τ2 can also be specified for networks designed to increase reach to and ensure acquisition of certain kinds of resources (Borgatti, 2005; Lawrence, Morse, & Fowler, 2005). For example, a stranger can give someone directions even though the two do not know each other and have no common contacts. In other words, for such a resource, its flow is in essence automatic and hence the τ1 parameter is close to infinity (∞). However, resources such as experience, capabilities, best practices, or tacit knowledge are likely to be more difficult to transfer than financial capital, data, advice, or simple information (Hansen, 1999). If so, greater embeddedness is needed to ensure their acquisition by the entrepreneur.

Even resources that are relatively easy to transfer (like money) often require assurances before one is willing to provide them. The example of Don Clark (Steier & Greenwood, 2000) illustrates that different levels of embeddedness may be needed to facilitate the flow of different types of resources. Mr. B was willing to invest his money in Clark's new venture because of the strength of their initial connection (i.e., its relational embeddedness). In time, Mr. B also became a mentor to Clark and began to share his knowledge of the industry (another type of resource). However, he was willing to do this only after Clark had “proved” himself further—not so much as trustworthy in preserving and growing Mr. B's investment but also as able to benefit from Mr. B's particular experience and advice. While the relational embeddedness of their tie stimulated the flow of both financial and mentoring resources, one (i.e., the mentorship) required further relational embedding. For such resources that require extra assurance, τ1 or τ2 would be lower depending on whether more relational or structural embeddedness is needed.

Types of Entrepreneurs and Alters

Characteristics of the entrepreneur or the alter may also influence the alter's willingness to cooperate and as a result dictate the extent of embedding required to ensure acquisition of the resource. These factors include the entrepreneur's and the alter's personalities, status, their risk aversion, and whether they are part of a sociocultural or ethnic group. For example, research in experimental economics offers evidence of the effects of gender and age on trusting behavior and perceptions of trustworthiness (Croson & Buchan, 1999; Sutter & Kocher, 2007). Similarly, Stuart and Sorenson (2007) argue that having high status would make it easier for entrepreneurs to obtain resources, reducing the need for embedding their relationships. Also entrepreneurs’ stock of resources, including their capabilities and experience, e.g., whether they are serial entrepreneurs or novices, may affect the level of embeddedness that alters are likely to “demand” before they are willing to part with their resources. Entrepreneurs who are more likely to be perceived as trustworthy or capable may not need to embed some (or even any) of their relationships as much as would otherwise be needed. In such cases, either spending less time and energy on the connection or using fewer mutual contacts would be enough. In our analytical model, for these entrepreneurs who can ensure resource acquisition with less embedding, the τ1and τ2 parameters would be higher. 1

Similarly, relationships that are extra useful to the alter may require less embedding. For instance, suppliers who derive economic benefits from entrepreneur's success (Zuckerman & Sgourev, 2006) are more likely to invest their own resources to ensure transfer and may accept a lower level of embeddedness before agreeing to resource flow. Conversely, more embeddedness may be required if the alter and the entrepreneur are competing for similar resources either directly or indirectly (e.g., due to structural equivalence) (Von Hippel, 1988). 2 The τ1 and τ2 parameters would be higher in the first case and lower in the second.

External Context

The alters’ willingness or ability to cooperate is also influenced by the organizational, industrial, institutional, and cultural contexts that encompass the entrepreneur's relationships. For instance, Stam et al. (2014) find that the need to embed connections, while important in established economies, is more important in emerging economies. In emerging economies, building trustworthiness either through relational or structural embeddedness is crucial for performance. In established economies, however, various institutions exist that encourage and enforce trustworthy behavior. This suggests that to reflect their greater need for embeddedness, the τ1 and τ2 parameters of the model would be lower in emerging economies. Similarly, Xiao and Tsui (2007) show the higher importance of structural embeddedness for ensuring resource flow in collectivistic compared with individualistic cultures. This implies a lower τ2 parameter for our analytical model in such settings.

Studies in organizational ecology, industrial clusters, and regional economics also hint at other candidates for contextual influence that would affect the required level of embeddedness. For example, the organizational ecology literature points towards a legitimation dynamic that shapes the industry's identity in the eyes of resource providers such as the investors, customers, suppliers, and potential employees (e.g., Aldrich, 1999; Hannan & Freeman, 1977; Hannan, Pólos, & Carroll, 2007). Such a dynamic indirectly increases the likelihood of resource acquisition by the participants in the industry by influencing the willingness and/or ability of other stakeholders to engage in transfer. Similarly, research in industrial clusters and regional economics highlights how being in a cluster or region that has a culture of sharing, greater mobility, or facilitating industry peer networks eases many common frictions that impede resource acquisition (e.g., Saxenian, 1996; Zuckerman & Sgourev, 2006).

Our analytical model is able to account for such wide varieties of differences among entrepreneurs, resources, and environments by tailoring the specifications of β, τ1, τ2, and CE. But, more importantly, even without adjusting these values, the framework highlights, and helps entrepreneurs address, pertinent questions about how best to allocate time and energy for networking purposes.

Results

Following standard techniques to solve the optimization problem represented in equation ((1)), we arrive at equation ((2)). This equation yields the number of valuable resources the entrepreneur is likely to reach and acquire. That is, it calculates the benefit the entrepreneur is expected to gain from his/her ego-network. This benefit depends on the number of alters (g = N − 1), the activity capacity CE, the number of mutual acquaintances zEi shared with the alters, as well as the parameter values β, τ1, and τ2.

display math((2))

To illustrate the implications of the model's solution, we compare and contrast the total benefit the entrepreneur would expect to garner through three different ego-network configurations. These three ego-networks exemplify the range of the arguments we have highlighted in this paper. Figure 5a depicts an ideal broker, where the entrepreneur is connected to otherwise unconnected alters. Figure 5c presents a fully cohesive ego-network where the entrepreneur and all his/her alters are interconnected. Finally, Figure 5b shows an entrepreneur's ego-network that meshes characteristics of both brokerage and cohesion. In this network, the entrepreneur's alters are tied to others in the same cluster but not to those in other clusters.

Figure 5.

Alternative Entrepreneurial Ego-Network Configurations

In all three networks depicted in Figure 5, the entrepreneur has g = 6 alters. In Figure 5a the entrepreneur does not share any of these alters with his/her other alters: (zEi = 0, ∀i). This specification allows us to simplify equation ((2)) into equation 3 for the ego-network in Figure 5a: inline image(equation 3). Figure 5b presents an example of many ego-network configurations where the entrepreneur shares some alters with others (∃i, st. zEi > 0). Here, each of the g = 6 alters has a tie to one other alter (zEi = 1, ∀i). The model's solution for this ego-network is: inline image(equation 4). In Figure 5c, the entrepreneur has g − 1 = 5 mutual acquaintances with each of his/her six alter(s), resulting in a fully cohesive ego-network. Placing (zEi = g − 1, ∀i), into equation ((2)) for this ego-network reduces it to: inline image.

The Model's Ability to Replicate Predictions of the Extant Literature

If we adopt the key assumptions of the brokerage literature, the model should be able to reproduce its predictions. We show this by specifying the parameters accordingly. First, we allow τ1 → ∞, i.e., resources are assumed to flow automatically, no matter how strong or weak the tie. Second, we set β = 1, i.e., mutual acquaintances are assumed to significantly decrease the value of an alter's resource to the entrepreneur. This specification yields the following solutions: inline image, inline image, and limSCE = 1 for the ego-networks in Figure 5a, 5b, and 5c, respectively. In other words, the entrepreneur is able to acquire g = 6 units of valuable resources from the 6 unconnected alters in Figure 5a. However, as the alters become more interconnected, the total benefit to the entrepreneur drops from six units in Figure 5a to three units in Figure 5b and to one unit in Figure 5c. This shows that the model can reproduce the expected results. In the analysis below, we proceed with the more general solution where parameters are not specified.

Comparison of the Benefits of Three Example Ego-Network Structures

If we compare the benefit yielded by each ego-network configuration in Figure 5a, 5b, and 5c, we notice all are increasing functions of g. 3 Taking the limit, as g → ∞ we find that:

display math
display math
display math

Comparing the first two limits to discover the relationship among the parameters β, τ1, and τ2, we arrive at the expression: βτ1 ≤ ? ≥ τ2. This indicates that it is not clear which is greater: βτ1 or τ2. When βτ1 is greater than τ2, the entrepreneur should prefer the brokerage ego-network shown in Figure 5a, which ensures the most efficient use of time and energy to achieve both reach to and acquisition of valuable resources. When βτ1 is smaller than τ2, the entrepreneur is better off with a network that meshes brokerage and cohesion, like one shown in Figure 5b. Similar conclusions can also be derived by comparing the first and third or the second and third limits. Thus, as one would expect, the values for β, τ1, and τ2 play a crucial role in deciding which ego-network the entrepreneur should prefer.

Discussion

Effects of β, τ1, and τ2 on the Entrepreneur's Design of the Ego-Network

Under the brokerage literature's assumptions that β is approximately 1 and resources always flow regardless of embeddedness, i.e., τ1 ↑ ∞, we find that βτ1 > τ2. Therefore, it is not surprising that under these assumptions, the entrepreneur should prefer a brokerage ego-network like the one in Figure 5a over either of the alternatives in Figure 5b and 5c. At the same time, the view that advocates cohesive ego-networks implicitly assumes β to be a low number around 0 and (τ2 > τ1), i.e., structural embeddedness is more important than relational embeddedness in facilitating resource acquisition. These assumptions yield βτ1 < τ2. Again, given the assumptions of this literature, it is not surprising that the entrepreneur should prefer the more cohesive ego-networks of Figure 5b or 5c over the one in Figure 5a, because both emphasize structural over relational embeddedness. As τ1 and τ2 vary with context, facilitating acquisition through either relational (vEi) or structural (zEi) embeddedness becomes more crucial.

Effects of Ego-Network Size (N) and Activity Capacity (CE) on the Entrepreneur's Design of the Ego-Network

Entrepreneurs’ choice between structural and relational embeddedness at the dyad level will also depend on the time and energy they expend on networking and on the total number of their alters. These dyadic choices aggregate into brokerage, cohesive, or meshed ego-networks at the network level. While some entrepreneurs spend only a few hours each week on networking, others spend much more (Greve & Salaff, 2003; Patel & Terjesen, 2011; Semrau & Werner, 2013). The results from the analytical model suggest that the greater the entrepreneurs’ activity capacity, the more they can afford to use relational instead of structural embeddedness to ensure resource acquisition. To the extent they can relationally embed a tie, they should, because the structural alternative would also decrease their reach to valuable resources.

In other words, when entrepreneurs have enough time and energy to embed each of their ties relationally, our model suggests that they can engage exclusively in network broadening behavior (Vissa, 2012). Having ample time and energy would allow them to relationally embed each of their ties and thereby acquire resources from each alter. In this case, the resulting brokerage ego-network structure will serve them the best. However, as the number of their connections increases, they will need more time and energy to relationally embed all ties. At some point, they will no longer have enough. When this happens, acquiring more from their current alters will become more of a concern. They have two choices at this point. They can begin to favor structural over relational embeddedness for some of their connections. By selectively choosing relational and structural embedding at the dyad level, they are effectively designing a social network that combines brokerage and cohesion at the ego-network level. Alternatively, they can prune back some of their ties until they can relationally embed all again. Consequently, for those entrepreneurs who do not have enough time and energy for networking, the more beneficial ego-networks are ones that either exhibit a mixture of brokerage and cohesion (i.e., a meshed ego-network) or are smaller than those of entrepreneurs who are equally capable but can spend more time and energy networking.

Reinterpreting Empirical Results in the Literature

To illustrate how our study and model help interpret extant findings in the literature, we briefly review results of three different empirical studies.

Patel and Terjesen (2011)

These authors find that network range (a proxy for the entrepreneur's reach to valuable resources dimension in our model) does not enhance venture performance unless it is accompanied by strong ties (relational embeddedness in our model) between the entrepreneur and his/her alters. Our study offers another rationale for this finding.

Suppose two identical entrepreneurs who differ only in their activity capacity. Let the first have more time and energy than the second to spend on her social network. Suppose further that, both have identical brokerage ego-networks and, hence, the same high levels of network range and enhanced reach to valuable resources. With more time and energy to network, the first entrepreneur is able to relationally embed more of her connections. She is likely to have an ego-network that exhibits both network range and strong ties. With less time and energy available, the second entrepreneur is less likely to relationally embed all his ties. Even though both have the same network range and can therefore reach similar sets of valuable resources, the flow of these resources is likely to be more restricted through the ties of the second entrepreneur. Our arguments explain why greater network range can help performance only if the entrepreneur actually acquires resources that greater range puts within his/her reach.

Our study suggests that the first entrepreneur has devised a more effective ego-network because it not only enhances reach to valuable resources but also assures their successful acquisition. Given the lower activity capacity (a lower CE) confronting our second entrepreneur, however, he should be willing to forego some network range (i.e., reduce reach to valuable resources) in order to create a more cohesive ego-network. This would enable him to benefit from relational and structural embeddedness simultaneously. In forgoing reach by structurally embedding some of his connections, he increases the likelihood of their acquisition. Doing so would likely more than compensate for the loss he suffers from reduced reach. We believe this offers another explanation why Patel and Terjesen (2011) find lower performance in social networks with high range but low average tie strength.

McFadyen, Semadeni, and Cannella (2009) and McFadyen and Cannella (2004)

In a different context (knowledge creation among academic researchers) that still informs entrepreneurship, McFadyen et al. (2009) also report findings that support the conclusions of our analytical model. They find a significantly negative interaction effect of density and tie strength (i.e., relational embeddedness) on knowledge creation. They conclude, just as our analytical model predicts, that brokerage ego-networks need to be coupled with strong ties to stimulate knowledge creation (McFadyen et al., p. 560).

Our findings suggest a second implication as well: Researchers with cohesive ego-networks do not need to incur the extra cost of relationally embedding their structurally embedded ties. Instead, while allowing these ties to remain weaker, they can still derive sufficient benefit from structural embeddedness (indicated by high density) to acquire the resources needed in the knowledge creation process. In doing so, they can use the freed up time to add more co-authors, just as our study suggests that entrepreneurs should do.

Using the same dataset, McFadyen and Cannella (2004) also find two important results that can be further explained by our analytical model. They report two nonlinear effects (both inverted U-shape)—of number of connections and of average strength of connections—on knowledge creation for these researchers. These results suggest that as either the number or the strength of ties increases, the activity capacity of these researchers is likely to begin constraining their networking choices. As they increase the number of their ties, the researchers begin to exhaust their time and energy allocated to networking. This yields lower average relational embeddedness, resulting in lower performance. Similarly, increasing average relational embeddedness with their ties may result in overcommitment to these connections. As a result, they would end up with relatively smaller ego-networks, which once again yield lower performance. That this activity capacity constraint may be binding is further supported by the negative correlation these authors report between the number of relations and the average strength of relations.

Our results offer two distinct choices for such underperforming researchers. If they believe they are committing more time and energy than is needed to ensure resource acquisition from their alters, they could reduce their commitment while maintaining the brokerage structure of their ego-networks. If they believe they are not overcommitting, they could structurally embed some of their connections. This would enable them to maintain or even increase their acquisition of resources while reducing the time and energy required. Either of these choices would enable these researchers to free up time and energy that they can use to expand their network and increase their reach.

Implications for Network Theory and Entrepreneurship Theory

This study contributes to the lively debates on brokerage vs. cohesive ego-networks and bridging vs. bonding social capital that exist in both the entrepreneurship and mainstream network literatures. We conclude that neither pure brokerage nor fully cohesive ego-networks but rather a meshed ego-network is more beneficial for the entrepreneur. Recently, to organize and reconcile the conflicting findings in the literature, researchers have suggested contingency-based solutions to these debates. Our conclusion is consistent with, yet different from, extant context specific arguments in the literature.

Uzzi (1996, 1997) examines how connections among firms affect a firm's performance and discusses the importance of (1) facilitating resource flow and (2) mixing arm's length and embedded ties to increase performance. The primary aim of these early studies was to specify the distinct contribution of social relations on economic behavior compared with the arguments of neoclassical economists and game theorists. Contrasting emdedded ties with those that are arm's length, purely material and weak, he demonstrated that information about such things as “relocating a production facility” was more likely to be shared with ego's embedded connections than with arm's-length ties. Uzzi, however, did not distinguish between different types of embeddedness. 4 By considering relational and structural embeddedness separately, this study offers a more nuanced portrayal of the mechanisms through which resource flow happens. In particular, we explain that structural embeddedness allows resources to flow through weak ties. Such structurally embedded ties do not demand as much time and energy for the alter and the entrepreneur to develop the mutual understanding and establish the trust required to facilitate resource acquisition. This option allows the entrepreneur to decide when it would be beneficial to structurally embed certain connections.

Uzzi (1997) concluded that to increase performance (measured as firm performance), the ego needs a mixture of arm's-length and embedded ties. Our argument further extends his conclusion in two ways. First, by differentiating between relational and structural embeddedness, we show that deciding to use one or the other represents another balancing act, besides the mixture that Uzzi suggests. Second, as he focuses mainly on relational embeddedness, diversity and value of resources is implicitly assumed. How mutual connections may reduce the value of resources within reach is not considered. We show that to increase performance, entrepreneurs need to consider not only how embedding a tie affects the flow of resources through that tie, but also the value of resources they can potentially receive.

In another set of studies, McEvily, Reagans, and Zuckerman also investigate the effect of social structure on performance (Reagans & McEvily, 2003, 2008; Reagans & Zuckerman, 2001; Reagans, Zuckerman, & McEvily, 2004). Claiming that each type of network structure may be more useful at different levels, these studies suggest another solution to the potential trade-off between brokerage and cohesion. Focusing on teams, they show that to increase overall team performance, the part of any team member's ego-network that includes other team members should be cohesive. The team members’ ties to others, i.e., to those beyond the team, however, should instead form a brokerage network. Such a network structure will not only increase mobilization of resources that the team eventually acquires (Centola & Macy, 2007), but also enhances the team's reach to a more diverse set of resources (Reagans & McEvily, 2008). In sum, they conclude that teams need a meshed network of brokerage and cohesive networks (also see Burt, 2005, for a similar argument).

Their conclusion seems similar to ours at first sight. However, our approach differs from theirs in two important ways. First, our focus is at the individual level, whereas they study a collection of actors as their unit of analysis. Their argument can be interpreted at the individual level as follows: A single individual (e.g., an entrepreneur) should have cohesive ties within himself/herself and brokerage ties with others. Simply put, although at the team level their conclusion has important implications, at the actor level it boils down to the generic conclusion offered by the brokerage literature. 5 Second, these studies implicitly assume, as is generally the case in the brokerage literature, that resources held by actors beyond the team will flow into the team without any need for embeddedness. Of course, this is at odds with what Hansen (1999) finds in his study on resource flow between departments or that Uzzi (1996) shows in his ethnographies. We claim and show in this study that resource flow is too tricky and important to be taken for granted. Some form of embeddedness is crucial to induce flow for many resources. The conclusions by McEvily, Reagans, and Zuckerman have potentially important implications for entrepreneurial teams and for the configuration of ties both within the team and beyond. Nevertheless, extensions of our results suggest that, for their connections to actors beyond the entrepreneurial team, if the team members are not able to relationally embed these ties, they may be well advised to structurally embed those ties with actors outside of the team. Only then can the entrepreneur or the entrepreneurial team ensure the acquisition of resources.

A third solution to the brokerage vs. cohesion debate relies on contextual contingency arguments (Stam et al., 2014). Stuart and Sorenson (2007, p. 213) observe “there are probably hundreds of studies that pit cohesion against brokerage in a horse race for explanatory power, with a bewildering array of contingencies to contextualize the potential benefits of one versus the other.” Similarly, Martinez and Aldrich (2011, p. 9) argue “entrepreneurs face a dilemma of whether to invest their limited resources in a small number of strong connections, a large number of weaker ties, or something in-between. We mention ‘something in-between’ because research suggests that there is no single solution to this dilemma. Rather than confirming that one type of network connection is better than the other, the literature has pointed out the contingent nature of networks’ values.” We agree with the basic premise that context is important and our study depicts a simpler route for considering context. Using our framework and model, researchers can contemplate how contextual factors may affect not only extending reach to valuable resources through brokerage or cohesive ego-networks, but also the likelihood of their acquisition at the dyad level. At the same time, before studying various contingencies, it is important to have a strong theoretical basis around which to organize findings.

By considering both dyad and ego-network levels and incorporating costs as well as benefits of designing, managing, and maintaining ego-networks and relationships, our theoretical framework and analytical model present one such organizing effort. Using the model as a baseline, researchers can investigate how contextual factors affect the various parameters of the model. Through this understanding, they can organize their findings and offer guidance to entrepreneurs regarding not only how they should approach designing and evaluating alternative ego-network structures, but also how they should think at the dyad level about adding, maintaining, or removing particular connections.

Implications for Entrepreneurs’ Agentic Behavior and the Evolution of Their Ego-Networks

Stuart and Sorenson (2007, p. 219) observe that the literature still lacks “systematic knowledge of how strategic actors construct their networks.” This study takes a small but important step in this direction. Our theory and results inform how entrepreneurs can think about evolving their ego-networks from small personal networks to larger and/or more professional ones. Knowing that relational and structural embeddedness are alternatives for enabling resource flow gives the entrepreneur a reason to take a more active role in introducing some of his/her unconnected alters. For example, as entrepreneurs pursue ties to additional investors, they can encourage them to invest (i.e., facilitate acquisition) by introducing them (i.e., forming structurally embedded ties) to existing investors (Bygrave, 1987; Lockett & Wright, 2001). This would avoid the relatively more costly step of relationally embedding these new ties. Furthermore, they may decide to form a structurally embedded tie with an alter and reduce the tie's relational embeddedness. Our results show that doing so would enable the entrepreneurs to maintain or even increase resource acquisition from the alter. At the same time, they will free up time and energy to expend on extending and growing their network by forming new ties.

Purposefully converting two relationally embedded ties into structurally embedded ones by connecting the two alters can offer another explanation for the observed high decay rates of bridge relationships. A bridge decays, i.e., a structural hole disappears, when two unconnected alters form a connection with each other and thereby close the triad with the entrepreneur (Burt, 2002). Burt (2005) sees bridge decay as a problem actors should seek to avoid or try to resolve once it happens, for example by dropping one of the now redundant connections. The results of our analytical model, however, support the possibility of employing a “tertius iungens” strategy (Obstfeld, 2005).

Any rationale for introducing two unconnected alters may at first appear counterintuitive, given the arguments that dominate the brokerage literature (see Obstfeld, 2005, for an exception). But, as our study shows, the entrepreneur can better achieve his/her dual aim of reaching valuable resources and acquiring them through just such a proactive action. Steier and Greenwood's (2000) study of the evolution of an entrepreneur's social network illustrates a case consistent with this conclusion. Early on Don Clark (the entrepreneur) was regularly spending one-to-one time with each of his initial investors. However, by the time the size of his social network doubled, managing each tie became a much more demanding endeavor. Consequently, Clark's ability to spend time with each investor to ensure he could acquire their resources and at the same time connect to new alters with other potentially valuable resources had begun to suffer. His ability to balance the dual aim of extending reach to valuable resources and assuring their acquisition was in jeopardy. Our results suggest that he should convert some relationally embedded ties to structurally embedded ones. In doing so, he converts what was essentially a brokerage network into one that meshes characteristics of brokerage and cohesive ego-networks. Doing so frees up time and energy that can then be allocated to forming and developing new ties to others.

In this way, entrepreneurs are able to play a more active and central role in shaping the evolution of their ego-networks. Notwithstanding the need to account for network cognition and other agentic concerns (Carley & Krackhardt, 1996; van Liere, Koppius, & Vervest, 2008), our findings contribute some of the much-needed micro foundations toward understanding key network evolution pathways. 6 We expect that the activity capacity of the entrepreneur, the current size of his/her ego-network, external contextual factors, and the changing type of resources he/she needs would moderate this network evolution. In particular, this study provides a basis for asserting the entrepreneurs’ agentic role as the central driver in building, growing, and evolving entrepreneurial social networks.

Limitations and Boundary Conditions

Our analytical model can serve as a basis for important extensions to the literature. At the same time, it can accommodate a wide variety of conditions. However, we would be remiss if we were to ignore its limitations. Next, we discuss these limitations and offer some suggestions for future research to circumvent them.

Potential Variation in Entrepreneurs, Entrepreneurs’ Alters, and Alters’ Resources

To keep the model traceable, we omit the possibility that some alters may be better connected than others or may possess different resources than other alters. We have also assumed that entrepreneurs are only after similar types of resources from each alter and try to accumulate as much of it as efficiently as possible. To account for such potential variation, researchers could perform two alterations. First, the β parameter in equation ((1)) can be allowed to vary per dyadic tie depending on the type of resources sought from each alter, rather than specifying one global β parameter. Second, dyad-specific τ1 and τ2 parameters can be used to account for differences in alters’ characteristics and their differing base level willingness and ability to transfer their resources. However, complicating the model in this manner would render the solution procedure difficult, if not impossible, to trace analytically. This would defeat the purpose of having a simple but powerful analytical model with generalizable conclusions. We think simulation methods such as agent-based modeling may offer more viable approaches to introduce such extensions.

From the entrepreneurship practice perspective, however, we think that our model's current set of conclusions still offer value to those entrepreneurs who face distinct alters and seek different resources. While some may see our model as a guide on how to design an effective ego-network from scratch, we think that it offers a more practical use. Our model offers even those entrepreneurs with established ego-networks an approach to assess and improve their current networks at the margin. Since one seldom has the opportunity to design an entire network from scratch, everybody, including entrepreneurs, may be more concerned with “How do I improve what I have?” and “Am I getting all I can from this contact? Am I investing too much time and energy for what I am getting? Are there resources I need that I cannot get from any of my current contacts?” This study guides entrepreneurs on how to answer these questions and helps them think about how to adjust the relational and structural embeddedness of each connection they already have. It prompts them to consider which, if any, contacts they may want to replace or at least reduce their investment in. It also helps them evaluate whether they need to free up more time and energy for new contacts through converting some of their currently relationally embedded ties into structurally embedded ones.

The Independence of Relational and Structural Embeddedness

It is not uncommon to observe that some ties are both structurally and relationally embedded. Vissa (2012) finds that entrepreneurs who perform network-deepening actions (i.e., those who prefer to build relationally embedded ties with existing contacts), later use these ties as a source of referrals for new connections. This essentially leads to structurally embedded ties. However, observing both types of embeddedness at the same time does not necessarily mean the two constructs cannot be conceived as independent or orthogonal. It is just as likely, if not more, for one form of embeddedness to occur independently of the other. Indeed, recent empirical studies suggest that relational and structural embeddedness can actually be uncorrelated (e.g., Bhagavatula, Elfring, van Tilburg, & van de Bunt, 2010).

Theoretically, we consider these two forms of embeddedness as independent for a simple reason. Only by considering the independent contributions of each to the entrepreneur's ability to reach and acquire valuable resources can we begin to understand how they affect entrepreneurs’ decision-making processes. Moreover, a careful entrepreneur can use one form or the other independently and avoid embedding a tie with both forms of embeddedness when such avoidance is appropriate. Importantly, our study provides insights for making this possible. For example, armed with the knowledge that either relational or structural embeddedness alone is sufficient to acquire resources, entrepreneurs can carefully choose how to embed a tie. They may choose not to encourage introductions among relationally embedded ties and prefer to keep them separate from each other, at least for some time. When two alters are introduced, regardless of who may initiate the introduction, they may look for ways to leverage the expected enhanced flow from either or both alters due to the greater structural embeddedness of their tie to each. Such choices will then also guide their decisions to further allocate their activity capacity by employing checks to ensure that they spend only the requisite amount of time and energy with each newly structurally embedded alter.

Conclusion

Entrepreneurs rely heavily on their network of social relationships as a crucial and primary source of all kinds of resources. Thus, their key focus in building and maintaining their ego-networks is satisfying a dual aim: increasing reach to valuable resources and convincing the owners of those resources to share or transfer them. Our review of the brokerage, cohesion, and embeddedness literatures yielded a growing number of, sometimes conflicting and confusing, observations and accounts regarding the networking behavior of entrepreneurs (see also Martinez & Aldrich, 2011, for a discussion). This prompted us to synthesize these various literatures into a coherent theoretical framework and analytical model.

We employed a multilevel approach that integrates the embeddedness of individual ties at the dyad level with the structure of the overall network at the ego-network level. The approach also incorporates the entrepreneur's activity capacity at the individual level. By considering these three different levels simultaneously, we were able to better account for the distinct mechanisms in play and to offer a resolution to the existing debate on the benefits of brokerage vs. cohesive ego-networks.

This study yields two important insights. First is the need to balance increasing reach to and ensuring acquisition of valuable resources. Entrepreneurs who fail to strike an appropriate balance can retard their entrepreneurial process. Second are the concrete steps towards striking that balance. Our results demonstrate that deciding between two alternative ways of embedding a specific dyadic tie represents a key decision in entrepreneurs’ attempt to achieve such a balance. If their activity capacity is sufficient, entrepreneurs should aim to form brokerage ego-networks and relationally embed all their ties. However, if their activity capacity constrains their ability to relationally embed all ties, they should structurally embed some ties and at the same time reduce the relational embeddedness of these ties. They should do so, until their activity capacity is no longer binding, to ensure they achieve their dual aim of increasing reach to valuable resources and acquiring them. This would mean that for those entrepreneurs who are constrained by networking time and energy an ego-network that meshes brokerage and cohesion is a more efficient ego-network structure.

Footnotes

  1. 1

    As a practical matter, a reduced need for embedding would have an effect similar to that of the entrepreneur having more time and energy available for networking.

  2. 2

    We thank our anonymous reviewers for bringing these two potential factors to our attention.

  3. 3

    This is because our simple cost function is not dependent on the number of the entpreneur's alters. It depends only on the energy expended.

  4. 4

    Uzzi emphasizes the history between two actors as the main determinant of the tie's embeddedness. While he also describes how new ties tended to be initiated by mutual contacts, his discussion of embeddedness centers around the direct relationship between the focal actor and the alter, without any reference to third parties. This usage of a tie's embeddedness mainly coincides with the relational embeddedness definition.

  5. 5

    A word of caution here is needed: Although their conclusion is not new at the individual level, the importance of their contribution lies in stressing the role of cohesion in mobilizing resources owned by group members for the benefit of the group and balancing this mobilization with the role of brokerage in bringing in more diverse resources.

  6. 6

    We thank our anonymous reviewers for highlighting the need for such an extension to study agency and evolution.

Ancillary