Network analysis and business networks

Abstract

The most common definition of 'network' in the business literature focuses on the ties or relations that one entity has with various other entities, abstracting from the ties among those other entities or with others not directly tied to the business of interest. In contrast, economists use a definition based on graph theory that has proven useful in a wide variety of fields (physics, engineering, sociology, economics, biology, etc). In that tradition, a network is a set of entities and the relations between all elements of the set. A fraction of the business literature applies the techniques based on the graph-theoretic definition of networks, known as social network analysis. This paper discusses some exemplary business network analyses and explains additional insights that could be obtained by applying graph-theoretic approaches. The conclusions drawn from this paper illustrate how the network structure affects the market power of the members, depending on whether the transactor has an alternative. The weak links form bridges to other networks and therefore provide alternatives which can reduce market power. This paper identifies the relevance of social network analysis in the social sciences in order to provide a better understanding of business networks.

Full text

published in:
Network analysis and business networks
Journal: Int. J. of Entrepreneurship and Small Business, 2014
Vol.21, No.3, pp.303 - 316
Network Analysis and Business Networks
by Maureen Kilkenny and Nerys Fuller-Love
NCFAP, Washington, D.C. Aberystwyth University of Wales
maureenkilkenny@gmail.com nnf@aber.ac.uk
Abstract: The most common definition of “network” in the business literature focuses on the ties or
relations that one entity has with various other entities, abstracting from the ties among those other
entities or with others not directly tied to the business of interest. In contrast, economists use a definition
based on graph theory that has proven useful in a wide variety of fields (physics, engineering, sociology,
economics, biology, etc). In that tradition, a network is a set of entities and the relations between all
elements of the set. A fraction of the business literature applies the techniques based on the graph-
theoretic definition of networks, known as Social Network Analysis. This paper discusses some
exemplary business network analyses and explains additional insights that could be obtained by applying
graph-theoretic approaches. The conclusions drawn from this paper illustrate the how the network
structure affects the market power of the members, depending on whether the transactor has an
alternative. The weak links form bridges to other networks and therefore provide alternatives which can
reduce market power. This paper identifies the relevance of Social Network Analysis in the social
sciences in order to provide a better understanding of business networks.
Network Analysis and Business Networks
Introduction
Among ‘hard’ scientists like engineers or mathematicians, network analysis is a structured technique used
to mathematically analyze a graph, circuit or a “network.” A graph is formally defined as a set of nodes
or entities, and the relations between them. This is also how economists and other social scientists
formally define a business network (Borgatti, Mehra, Brass and Labianca, 2009). The entities of interest
are either individuals: people in businesses or markets; or groups such as firms, localities, or associations;
in markets or society in general. The relations of interest include transactions for goods or services;
exchanges of assets, information, or trust; and so on.
In the business literature, a business network is typically the set of people or firms with whom one
business or business person has relationships. Thus the majority of business network analyses in that
literature concern what graph theorists call ego-networks (for an excellent review, see Borgatti and Foster,

2
2003). The focus is on the relations that a single business entity ego has with various alters. The alters
may be, for example: suppliers, customers, lenders, mentors, or even competitors; in the marketplace, or
in affiliations such as a local industry association.
This differs from the approach applied by economists and the other social scientists using graph-theoretic
approaches who study the relations between and among all the entities. Figure 1 illustrates this difference
between the perspectives in the business and social science literatures concerning business networks.
Figure 1. Ego- and complete graph-theoretic versions of the network of one business
The business entity of interest in Figure 1 has exactly the same set of direct, pair-wise, first-degree, or
dyadic relationships with suppliers, customers, employees, mentors, competitors and a financier in both
representations of the network. The networks contain the same set of entities. But there is more
information about ties in the graph-theoretic one. It is particularly rich with respect to information
relations. The graph-theoretic version documents information flows—usually associated with
interpersonal ties--among the employees themselves, between an employee and one competitor, between
a competitor and one of the business’s suppliers, and even between a mentor and a competitor. The
graph-theoretic representation also documents the existence of transactions between its competitors and
its suppliers and customers. And it documents that one of the business’s employees is employed by a
competitor.
Interpersonal ties that undergird information flows are particularly important for the vitality of a business
(see Borgatti and Cross, 2003). Businesses whose employees have good interpersonal ties among
themselves retain valuable employees at lower costs. And businesses in which interpersonal information
ties are effective are better organized for innovation.
“Informal networks — also known as social networks — are especially important in knowledge-
intensive sectors, where people use personal relationships to find information and do their jobs.
This fact is supported by our own research and that of many others. One researcher who looked
at this question for more than a decade, just to give one example, found that engineers and
scientists were roughly 5 times as likely to turn to friends or colleagues for information as to
impersonal sources” (page 68, Cross, Nohria, and Parker (2002); emphasis added.)
Furthermore, knowledge of the business’s alters’ other relations provides insight into the entity’s
competitive position. When an alter with whom a business relates has no other ties, it is monopolized by

3
that business. But if the alter has just one alternative tie, the business’s monopoly power is neutralized.
By a similar token, a unique information source may be novel but it is also unverifiable. If there is at
least one other source of similar information, the information is potentially verifiable and more reliable.
Clearly it is useful to know the relations that the other entities in a business’s ego-network have with each
other and with others, in order to understand the business’s strategic opportunities or to deduce the
verifiability of its information flows.
Of course, to document the full graph-theoretic representation of a business’s network requires more than
the data about the single entity’s relations. The main point of this article is the value of collecting the data
about all the other relations. We begin with a brief review of opportunities and precedents in the
literature. We discuss exemplary applications of the qualitative, data rich ego-network approach in the
business literature. Then we introduce a few useful constructs from graph-theoretic Social Network
Analysis (SNA). Finally, we define “weak” and “strong” ties, and show how SNA is used to analyze a
small business’s strategic position or assess the effectiveness of a knowledge network.
Opportunities and Precedents
A search of the business literature using “Business Source Complete” online lists hundreds of articles
mentioning “business network.” Of those, however, just 28 also mention Social Network Analysis
(SNA). This indicates the opportunity in the business literature for analyses of business networks using
the graph theoretic toolkit called SNA (see Wasserman and Faust, 1994; and Borgatti, Everett, and
Freeman, 2002, for software). The opportunities for SNA analyses of business networks may be even
greater in the economics literature. Though there are now two excellent economic texts about network
analysis: Jackson (2010), and Easley and Kleinberg (2010), a search of the economics literature using
“Econlit” lists only one journal article about a business network that applies SNA: by Giuliani (2007) in
the Journal of Economic Geography.
Our brief chronological review of the business literature about business networks begins by noting the
context. About fifty years ago, interest in ‘business networks” grew with the understanding that a
responsible firm not only manages relations with its own employees, activities, and internal units, it also
manages how it relates to entities beyond the narrow legal ‘boundary’ of the firm (e.g. Tichy, Tushman,
and Fombrun, 1979; Anderson and Narus, 1990; Anderson, Håkansson, and Johanson, 1994). The old
idea was that the boundary of a firm excluded everything that wasn’t legally part of the firm. The
external context was called the firm’s ‘environment,’ and it was thought that a firm could not change its
environment.
The new idea was that firms can and do shape their environment by forging relationships:
“A relationship gives each firm a certain influence over the other… each firm gains control of at
least some of its environment while giving away some of its internal control.” (Anderson and
Narus, 1990).
When it was recognized that the boundary of a firm is porous and in flux, that construct became less
useful. The ego-network construct rose in importance (e.g. Eccles, 1981). In the new tradition, the ‘firm’
is its legally-delimited self, and the firm’s relationships are called its “network.” This may explain the
business literature focus on ego-networks. Furthermore, both the firm and its network were surrounded
by what was considered to be a parametric market environment. Network analysis is now, however, also
understood to be a good foundation for operationalizing strategic game theory (Thorelli, 1986; Jackson,
2010; Easley and Kleinberg, 2010).

4
Exemplary Analyses of ‘Business Networks’ in the Business Literature
Studies of business [ego-]networks are based on much more than just data about the ties the ego of
interest has with the various alters. Analysts also collect information, usually through interviews or
surveys, about alters’ characteristics, the nature of the ties, the length and intensity of the relationships,
the purpose or usefulness of the ties (e.g. Johnsen and Johnsen, 1999) the consequences of particular
types of relationships, and other information depending on the topic.
For example, Greve and Salaff (2003) studied entrepreneurial networking, in which the noun “network” is
used as the verb “to network.” Concerning ego-network direct relationships, they interviewed
entrepreneurs about the amount of time they invested to establish or maintain their ties, by phase; and
examined the roles of family members. They compared the responses of novices to experienced
entrepreneurs, and first-generation to second generation entrepreneurs. Their research confirmed that
entrepreneurs ‘network’ opportunistically. Entrepreneurs choose with whom to tie depending on what
they need to know during the different phases of their business’s development. Time invested in network
tie formation also differs from phase to phase. But Greve and Salaff did not find any difference between
the time novices and experienced entrepreneurs spend networking. Parents are included in advice
networks by second generation, but not first generation, entrepreneurs. And the role of a family member
in an entrepreneur’s network does not diminish or change over subsequent phases of the business’s
development. It would be interesting to know if entrepreneurs who maintain relations regardless of the
benefit to the business fare better or worse than entrepreneurs who network opportunistically, but the
question of the relative benefits of embedded vs arm’s length networking was not investigated.
A relative strength of the approach in the business literature is that it provides this kind of qualitative and
dynamic information about the how, who, and why businesses make and keep relationships, that is, it
provides information about the verb ‘networking.’ Although much effort has been spent by statisticians
and social scientists to develop models of network formation (e.g., Wasserman and Faust, 1994) the early
probability models assumed that potential alters were differentiated only by the numbers of ties that they
have, otherwise the alters were all the same. Thus the graph-theoretic approach had little to say about
other characteristics of the entities with whom one ties. As Rapoport (1979) said in an overview of the
graph-theoretic approach to predicting network tie formation:
"Mathematical modeling is a vehicle for absolutely rigorous reasoning and therein lies its
advantage. A disadvantage of mathematical modeling is that it necessitates a drastic paring down
of the details of the phenomena modeled...these simplifications...can impair or altogether destroy
the model's pragmatic relevance."
The shortcomings of the early graph-theoretic approach for describing the act of “networking” may
explain the dearth of graph-theoretic analyses of business networks in the business literature. (SNA has
made great strides since then, however, uncovering insights about the assortative character of real-world
networks (e.g., Schaefer, Simpkins, Vest and Price, 2011). Assortative networks display homophily, the
tendency of people to associate with people like themselves – see McPherson, Smith-Lovin and Cook (
2001). To date, most business network analyses are qualitative data rich case studies rather than
mathematical or statistical models.
The analyses in the business literature describe various types of networking (the verb), such as
subcontracting (e.g., Sachetti and Sugden, 2003), entrepreneurial networking (e.g., Johannisson,
Alexanderson, Nowicki, and Senneseth, 1994, Johannisson and Mønsted, 1997; Greve and Salaff , 2003;
Uzzi, Amaral, and Reed-Tsochas, 2007), and network management or governance (e.g., Jones, Hesterly,
and Borgatti, 1997). Some analyses assess networks according to their contribution to profits (e.g.,
Jarillo, 1988; or Hagedoorn and Schakenraad, 1994, on strategic networking). The qualitatively rich,

5
survey data-augmented approach to business network analysis has also been applied to analyze network
structure and content as well. For example, it was used to compare local to international joint ventures
(e.g., Walker, Kogut, and Shan, 1997; Gulati, 1998) and to describe industrial districts (e.g., Sacchetti and
Tomlinson, 2009).
Exemplary applications of SNA in the business literature include Lee (2005) concerning supply chain
integration, Carter, Ellram, and Tate (2007) about applications of SNA to logistics; Bjork and Magnusson
(2009) about innovation, and Chen and Tan (2009) about the networks supporting transnational
entrepreneurship. A set of publications by Nohria (1992); Cross, Borgatti, and Parker (2002) and Cross,
Nohria, and Parker (2002) report how SNA can be used to understand and better manage businesses.
Cross, Nohria, and Parker (2002), for example, confront ‘six myths’ about how information is shared and
used by employees or collaborations within organizations, show the insights available from application of
simple SNA techniques, and discuss the implications for strategic management.
A survey of SNA by Tichy, Tushman, and Fombrun (1979) provided an introduction to SNA’s lexicon
and basic techniques for the business literature. Two sociologists have contributed notably to that
lexicon: Burt (e.g. 1992) and Uzzi (1997). The influential study by Uzzi (1997) distinguished the arm’s
length ties from the embedded ties of 21 businesses in a garment district.
According to Uzzi (1997) arm’s length ties are the ones in which there is a “push for the lowest price
possible,” a need to monitor the alter for opportunistic behavior, and where the way to resolve a problem
is to exit the relationship. Embedded ties are the ones in which “personal relationship matters,” “trust is
a major aspect,” “reciprocity and favors are important,” there is “joint problem-solving,” and there are
“strong incentives for quality.” Thus, arm’s length ties are impersonal market transactions, while the
same market transactions-- between personally related or personally accountable entities--are embedded.
Uzzi was able to draw these inferences after carefully analyzing numerous open-ended personal
interviews.
The same kind of inference could also have been obtained from an SNA analysis of straightforward
yes/no answers by each of the 21 businesses to simple questions such as “do you pay (receive) the lowest
prices to(from) Business Y?” and “do you give (or receive) favors to (from) business Y?”
To apply SNA, record the yes(no) responses as 1s (or 0s) in four 21x21 tables that contain one row and
one column for each business in the network. One table should be completed to record each X’s answers
to “pay lowest price to” when X buys from Y. The cell in the row for X and the column for Y is 1 if X
reports ‘yes, they strive mainly to pay the lowest price when purchasing from Y,’ and 0 if not. The next
table records if X “receives lowest price from” when Y buys from X. A third table has 1’s for “give
favors to,” between each X and Y; and a fourth shows “receive favors from.” The relationship between
any pair of businesses X and Y could be classified as arm’s length if both X and Y answered yes to the
former question (lowest price) and no to the latter (gives/receives favors), for example. And it is
embedded if both answered no to the former and both yes to the latter question, in both directions. To our
knowledge this kind of application of SNA to distinguish arm’s length from embedded ties between firms
has not yet appeared in the literature.
Beyond the dyad/Insights from SNA
Consider the value of also knowing about the ties between alters. Figures 2a and 2b illustrate how
additional knowledge about the relations that an ego’s first-degree alters have provides a richer and more
useful analysis of a firm.

6
Figure 2a shows the simplest network a firm may be part of, a 3-entity ego network. The firm itself is the
hollow circle, and the two first-degree alters are solid circles. The firm is dyadically related or tied to the
entity shown on the left, and to another shown on the right. Figure 2b illustrates how the alters’ ties--
beyond the dyad-- determine the strategic position of the business or the effectiveness of its information
exchanges, among other things.
Figure 2a. A simple 3-entity ego-network
The diagrams in Figure 2b show how the strategic position of a business cannot really be described by the
first-degree dyadic relations alone. The firm’s first-degree network could be a subset of at least four very
different canonical networks: the circle, chain, tree, or star. If an analyst documents only the first-degree
ties of the ‘hollow circle’, the analyst may remain unaware of the salient features of the environment in
which that business actually operates. The good news is that these salient features can be identified from
just three iterations of a simple SNA-type network survey or questionnaire.
Figure 2b. Four distinct network structures that contain the same 3-entity network.
Consider the features that distinguish the four canonical network structures: circle, chain, tree, and star,
with respect to (i) the market power of the entity of interest (the hollow circle), (ii) the quality of
information sharing, and (iii) the robustness of the network to the loss of a member.
A “circle” network is the most decentralized. Every entity in the network is directly connected to exactly
two others in the network. And every entity is indirectly connected to every other (by not more than a
two-step path in this small network). Thus in a circle network, the hollow circle entity has no more market
power than any other. Information is fully shared, and because every entity has more than just one tie to
the network, it can be verified. In a latter section of this paper we will explain that if the ties in a circle
network are either unidirectional or reciprocated, then all the ties in a circle network are “bonding” ties or
“strong” ties. So the circle network is most robust to the loss of any one member of the network.
A “chain” network is more centralized. The hollow circle has power over its alter to the left, but it has less
power over its alter to the right. The central entity has no power at all. The ‘chain’ network illustrates the
subtleties of the contention that an entity’s power derives not just from being central, but by relating with
entities that have no alternatives.

7
Consider the relative bargaining power of the three interior entities in the chain, when all relations are
exclusively pair-wise, that is, when an entity relates to just one other at a time. The exclusion mechanism
refers to this kind of competitive situation. Yamagishi, Toshio, Gilmore, and Cook (1988) conducted a set
of experiments which showed that entities tied to peripheral entities have the highest bargaining power,
while peripheral and central entities have the least. In Figure 2b, though the central entity has the same
number of alternatives as the 2nd and 4th in the chain, both of the center’s alternatives are themselves
powerful. That makes the center of the chain weaker. The insight is that an entity’s capacity and
performance depends not only on its own position but also on the positions and relations of the others in
the network.
With respect to information flows, the chain network can be efficient, especially if the ties are all
reciprocal ties (two-way). But even if all the ties are reciprocated, the chain network is sensitive to the
loss of any one of the interior entities. It is the least robust of the four canonical types to the loss of any
of three entities.
A ‘tree’ network is more centralized than the ‘chain’ or the ‘circle’. The power of the hollow circle
entity, however, is lower in the ‘tree’ network than in the ‘chain.’ This is of course a consequence of the
power that the central entity, to the right of the hollow circle, has in the tree network. In the ‘chain’ the
center was powerless because all of its alters had alternatives, but in the ‘tree’ the center is more powerful
because two of the center’s three alters have no alternatives. Note that Information flows may be
impeded by the center. Finally, the tree network is robust to the loss of the three peripheral entities, but
not to the loss of its center (or the hollow circle)
The ‘star’ network is the most centralized of all four canonical types. The hollow circle in the center (the
business entity of interest) has monopoly power over all the others. The star network structure can be
efficient with respect to information flows and problem solving if the central entity is very competent. It
can be disastrous if the center lacks competence or refuses to intermediate. All of the information flows
with peripheral entities are unique, none can be verified. Finally, the star is robust to the loss of all but
the central entity; and it is the most sensitive to the loss of that one entity.
With respect to knowledge flows and problem solving, early research by Bavelas (1950) and by Leavitt
(1951) concluded that the fastest performing network structures were those in which the distance of all
nodes from an ‘integrator’ was the shortest – i.e., that the star network would be the most efficient.
Subsequent research has found that Leavitt’s conclusions depended on other factors such as the simplicity
of the problem to be solved and the aptitude of the central entity/integrator. Fuller-Love and Thomas
(2004) found that networks were very important for information sharing and problem solving and Radner
(1993) suggests that decentralized or ‘loose’ groups are more efficient.
Finally, the last point of this section is that all those additional insights are easy to extract from a simple
questionnaire about of the relational ties among and between the business entity of interest, its alters, and
their alters. Another possible reason why the SNA approach is not often applied in the business literature
is the desire to exogenously define the set of entities to be surveyed. Alternatively, a ‘snowball’ survey in
three iterations could be conducted; e.g. Illenberger, Kowald, Axhausen and Nagel (2011), Figure 3.

8
Figure 3. A network delimited by two iterations of a ‘snowball’ sample, showing all the ties reported by
egos in the first iteration, but only the bridging ties of the alters in the second iteration; Illenberger, at al
(2011), working paper online.
For a ‘snowball’ survey, in the first iteration the analyst starts with one (or more) egos, describes them,
and identifies their alters. In the second iteration, the analyst describes the first iteration’s alters and
identifies their alters as well. The process of documenting the ego-networks of each alter can be repeated
for a third iteration or more. To obtain insights about the market power of a small business, three
iterations may be necessary, but not more; as illustrated in Figure 2b. Illenberger, et al (2011) discuss the
sampling bias inherent a snowball survey approach, and describe one way to correct for it. Rothenberg
(1995) provides a useful critical survey; see also Goodman, 1961.
The Strength of the Graph-Theoretic Definition of “Weak Ties”
Ties that are uniquely essential for the connectedness among subsets of a network are called “bridges”
(Granovetter, 1973). A bridging tie is one that makes a network have fewer, larger subsets (called
components) when it is present than absent. In Figure 3, components appear as asterisk-like stars, and the
links between the asterisk-like stars are the bridging ties. If those linking ties did not exist, there would
be more structural holes (Burt, 1992) in Figure 3. In Figure 2b, all of the ties in the chain, tree, and star
networks are bridging ties. None of the ties in the ‘circle’ network is a bridging tie. The five entities will
remain connected (as a ‘chain’) if any one of the ties is dropped. Each of the ties in the circle network is
reinforced by another tie. Reinforced ties are strong ties. Bridging ties are weak ties.

9
Granovetter’s (1973) conjecture about ‘the strength of weak ties” is widely cited even in the business
literature that doesn’t apply SNA. But the implied definitions are not standard. It behooves us to clarify
the meaning of ‘strong’ and ‘weak’ network ties. Entities i and j are:
 Weakly connected if they are linked by a series of ties regardless of their direction;
 Strongly connected iff there is a directional path from i to j and a path from j to i.
In sum, a connection is strong if there is more than one way for the substance of the relation to flow
between the connected entities. Thus according to graph-theoretic SNA, tie strength comes from
reinforcement by at least one other tie (Iacobbucci, 1994; citing Harary, Norman and Cartwright, 1965;
see also Easley and Kleinberg, 2010). Granovetter (1973) wrote that
“the hypothesis which enables us to relate dyadic ties to larger structures is: the stronger the tie
between A and B, the larger the proportion of individuals in S [the set of all entities in an
exogenously delimited network] to whom they will both be tied, that is, connected by a weak or
strong tie. This overlap in their friendship circles is predicted to be least when their tie is absent,
most when it is strong, and intermediate when it is weak.” (page 1362; parenthetic clarification
added).
Thus Granovetter postulated ‘if strong then reinforced,’ and “if reinforced then strong,” which in formal
logic means that strong ties are a wholly-contained subset of reinforced ties. This is precisely the graph
theoretic definition. Therefore, in SNA the strength of a tie is not the value or intensity of the relation
(even though sometimes value and intensity are measured to deduce or indirectly measure tie strength).
The strength of a tie is directly measured by the presence of at least one reinforcing tie. If there are no
reinforcing ties, a tie is weak. If there is at least one reinforcing tie, a tie is strong.
There is considerable confounding of unreinforced ties with low value, low frequency, or low intensity
ties in the business literature. This may arise from a confusion of cause with effect. Granovetter (1973)
explained that tie strength is a consequence of contact frequency, reciprocity, emotional intensity, and
intimacy. Reciprocated ties are, by definition, reinforced strong ties.
It may also arise from the approach that Granovetter took to measure tie strength for his 1973 paper.
Granovetter could not collect network data about all ties between all the entities in one network. That is
because the network boundaries were undefined. He could not know all of the entities in each
interviewee’s network until he had completed all the interviews. So he simply asked the individuals
about the frequency and intimacy of their ties with the persons through whom they learned of new job
possibilities. From their responses he inferred tie reinforcement or the lack of it, to distinguish strong
from weak ties. Granovetter deduced that the strength of weak ties is that weakly-tied contacts provide
more unique or innovative information (see also Boorman, 1975). Equally confounding to the profession
may have been Burt’s (1992) argument that ‘empirically, strength has two independent dimensions:
frequent contact and emotional closeness” page 19; a claim that Burt elaborated throughout his work.
Qualitative information about the nature, intensity, frequency, or intimacy of contact can be just as
difficult to collect—if not more difficult-- as quantitative network tie data through a three-iteration
snowball survey. Qualitative information is certainly more difficult to analyze systematically, especially
if interpersonal comparisons are involved. For example, a shy person might rank as ‘weak’ a relationship
that a gregarious person may rank as ‘strong.’ There’s no ambiguity if the SNA approach is used. A tie
is either reinforced or it is not. This clarity is a strength of the graph-theoretic approach to documenting
weak vs strong ties. Other strengths are revealed when quantitative SNA techniques are applied to
analyze a business network. For example, Krackhardt and Stern (1988) argue that bonding or strong ties
enable an organization to weather crises.

10
An SNA analysis of the information and knowledge flows among employees of a firm can provide very
tangible and useful management information. Cross, Nohria, and Parker (2002) or Borgatti and Cross
(2003) provide excellent examples. They explain how to document knowledge flows among
collaborators by asking a simple set of questions of each employee in a firm, then applying SNA. Cross,
Nohria, and Parker (2002) explain the efficiency benefits of identifying and filling the structural holes in
a firm’s knowledge network, linking two subsets of previously non-communicating collaborators. They
explain the benefits of identifying and correcting information bottlenecks, and other common problems.
Summary
In the business literature a business network is usually a firm (ego) and the entities with whom it has
various relations (alters). Case studies or surveys of business egos and their alters provide insights about
the formation, nature, and consequences of a business’s ties with other entities; locally and
internationally. To network analysts in other disciplines— anthropology, biology, computer science,
economics, engineering, mathematics, physics, political science, psychology, sociology, statistics, and—
an ego network is but a subset of the whole network, both inside the entity and beyond --the structure in
which the entity operates.
To understand the competitive position of a firm or to better manage information flows within a firm it is
useful to know about the relations among and between the alters, and to know the relations of the alters as
well. A large toolkit of quantitative techniques known as Social Network Analysis could be applied to
document and analyze small business networks, but such applications are less common in the business
and economics literature. This paper briefly surveyed the business literature to understand reasons for the
relatively narrow approach to network analysis. Then we explained an easy way to document a small
business’s network using snowball sampling and SNA, discussed the implications of canonical network
structures, and formally defined the ‘weak tie’ construct.
References
Anderson, J. C., Håkansson, H., and Johanson, J. (1994) “Dyadic Business Relationships within a
Business Network Context” Journal of Marketing 58(Oct):1-15.
Anderson, J. C. and Narus, J., (1990) "A Model of Distributor Firm and Manufacturer Firm Working
Partnerships," Journal of Marketing, 54 (1): 42-58.
Bavelas, A. (1950) “Communication patterns in task-oriented groups” Journal of the Acoustical Society of
America, 22:725-739.
Björk, J.and Magnusson, M. (2009) “Where Do Good Innovation Ideas Come From? Exploring the
Influence of Network Connectivity on Innovation Idea Quality” Journal of Product Innovation
and Management 26:662–670
Boorman, S. (1975) “A combinatorial optimization model for transmission of job information through
contact networks” Bell Journal of Economics 6:216-249.
Borgatti, S..P., Everett, and Freeman, L. (2002) Ucinet for Windows: Software for Social Network
Analysis. Harvard, MA: Analytic Technologies.
Borgatti, S. P. and Foster, P. (2003) “The Network Paradigm in Organizational Research: A Review and
Typology” Journal of Management 29(6) 991–1013.

11
Borgatti, S. and Cross, R. (2003) “A Relational View of Information Seeking and Learning in Social
Networks Management Science 49(4):432-445.
Borgatti, S. P., Mehra, A., Brass, D.J. and Labianca, G. (2009) “Network Analysis in the Social Sciences”
Science 323(Feb):892-95.
Burt, R. S. (1992) Structural Holes: The Social Structure of Competition Harvard University Press,
Cambridge MA.
Burt, R.S. (2004) “Structural Holes and Good Ideas” American Journal of Sociology 110(2):349–99.
Carter, C., Ellram, L. and Tate, W. (2007) “Network Analysis in Logistics Research” Journal of Business
Logistics 28(1):137
Chen, W., & Tan, J. (2009). Understanding transnational entrepreneurship through a network lens:
Theoretical and methodological considerations. Entrepreneurship Theory and Practice, 33(5),
1079-1091.
Cross, R., Borgatti,S. Parker, A., (2002) “Making Invisible Work Visible: Using Social Network Analysis
To Support Strategic Collaboration” California Management Review 44(2):25-46.
Cross, R., Nohria, N. and Parker A., (2002) “Six Myths About Informal Networks — and How To
Overcome Them” MIT Sloan Management Review Spring:67-79.
Cross, R. and Parker, A., (2004) The Hidden Power of Social Networks, Harvard Business School Press,
Boston, MA.
Easley, D. and Kleinberg, J. (2010) Networks, Crowds, and Markets: Reasoning about a Highly
Connected, World Cambridge University Press, New York, NY.
Eccles, R. G. (1981) “The Quasifirm in the Construction Industry,” Journal of Economic Behavior and
Organization, 2: 335–357.
Fuller-Love, N., & Thomas, E. (2004). Networks in small manufacturing firms. Journal of Small Business
and Enterprise Development, 11(2), 244-253.
Giuliani, E. (2007). The selective nature of knowledge networks in clusters: evidence from the
wine industry. Journal of Economic Geography, 7(2), 139-168.
Goodman, L.A. (1961) "Snowball sampling" Annals of Mathematical Statistics 32:148-70.
Granovetter, M. (1973) “The Strength of Weak Ties” American Journal of Sociology 78:1360-1380.
Greve, A. and Salaff J. W., (2003) “Social Networks and Entrepreneurship” Entrepreneurship Theory
And Practice Fall:1-22.
Gulati, R. (1998) “Alliances and Networks,” Strategic Management Journal, 19(4): 293–317.
Hagedoorn, J, and J. Schakenraad (1994), 'The Effect of Strategic Technology Alliances on Company
Performance,” Strategic Management Journal, Vol. 15, 291-309

12
Harary, F., Norman,R., and Cartwright, D., (1965) Structural Models: Introduction to Theory of
Directed Graphs, New York: Wiley and Sons.
Iacobbucci, D. (1994) “Graphs and Matrices” Chapter 4 in Wasserman, S., and Faust, K. (1994) Social
Network Analysis: Methods and Applications Cambridge, Cambridge University Press; pages 92-
166.
Illenberger, J., Kowald, M., Axhausen, K.W., Nagel, K. (2011) “Insights into a spatially embedded
social network from a large-scale snowball sample” Working paper http://svn.vsp.tu-
berlin.de/repos/public-svn/publications/vspwp/2010/10-10/submission-27-03-2011.pdf
Jackson, M. (2010) Social and Economic Networks, Princeton University Press.
Jarillo, J.C.(1988) “On Strategic Networks,” Strategic Management Journal 9(1):31-41.
Johannisson, B.; Alexanderson, O.; Nowicki, K.; and Senneseth, K. (1994) "Beyond Anarchy and
Organization: Entrepreneurs in Contextual Networks." Entrepreneurship and Regional
Development 6(4):329-356.
Johannisson, B. and Mønsted, M. (1998), “Contextualizing Entrepreneurial Networking: The Case of
Scandinavia” International Studies of Management and Organization 27(3):109-136.
Johnsen, R.E., Johnsen, T.E., (1999) “International market development through networks: The case of
the Ayrshire knitwear sector.” International Journal of Entrepreneurial Behaviour and
Research, 5(6):297-312.
Jones, C., Hesterly, W. and Borgatti, S. (1997), “A General Theory of Network Governance: Exchange
Conditions and Social Mechanisms” The Academy of Management Review 22(4):911-945.
Krackhardt, D., and Stern, R.M. (1988) “The structuring of informal organizations and the management
of crises” Social Psychology Quarterly, 51:123-140.
Leavitt, H. (1951) “Some Effects of Communication Patterns on Group Performance” Journal of
Abnormal Social Psychology 46:38-50.
Lee, P. (2005) “Measuring Supply Chain Integration: A Social Network Approach” Supply Chain Forum,
An International Journal 6(2):58-67.
McPherson, J. M., Smith-Lovin, L., & Cook, J. M (2001) “Birds of a feather: Homophily in social
networks,” Annual Review of Sociology, 27: 415–444.
Nohria, N. (1992), ‘Is a network perspective a useful way of studying organizations’, in Nohria, N., and
Eccles, R.G. Networks and Organisations: Structure, Form and Action, Harvard Business School
Press.
Radner, R. (1993) “The Organization of Decentralized Information Processing,” Econometrica
61(5):1109-46.
Rapoport, A. (1979) "Some problems relating to randomly constructed biased networks," Perspectives on
Social Network Research 119:37.

13
Rothenberg, R. B. (1995) “Sampling in Social Networks” Connections 18(1):104-10.
Sacchetti, S., & Sugden, R. (2003). The governance of networks and economic power: the nature and
impact of subcontracting relationships. Journal of economic surveys, 17(5), 669-692.
Sacchetti, S. and Tomlinson, P. (2009) “Economic Governance and the Evolution of Industrial Districts
Under Globalization: The Case of Two Mature European Industrial Districts” European Planning
Studies 17(12):1837-1859.
Schaefer, D., Simpkins, S., Vest, A., and Price, C., (2011) “The contribution of extracurricular activities
to adolescent friendships: New insights through social network analysis.” Developmental
Psychology, 47(4):1141-1152.
Thorelli, H.B. (1986) “Networks: Between Markets and Hierarchies” Strategic Management Journal
7(1):37-51.
Tichy, N.M., Tushman, M.L., Fombrun, C., (1979) “Social Network Analysis for Organizations” The
Academy of Management Review, 4(4):507-519.
Uzzi, B. (1997) “Social structure and competition in interfirm networks: The paradox of embeddedness”
Administrative Science Quarterly 42(1):35-67
Uzzi, B., Amaral, L. and Reed-Tsochas, F., (2007) “Small-world networks and management science
research: a review” European Management Review 4:77–91.
Walker, G., Kogut, B. and Shan W. (1997) “Social Capital, Structural Holes and the Formation of an
Industry Network” Organization Science 8(2):109–25.
Wasserman, S. and Faust, K. (1994) Social Network Analysis: Methods and Applications, Cambridge,
Cambridge University Press.
Yamagishi, T., Gilmore, M. and Cook, K., (1988) “Network Connections and the Distribution of Power in
Exchange Networks” American Journal of Sociology 93(4):833.