A case study on network orchestration in the Dutch water technology sector

(1)

How does the head cattle lead its herd in a fragmented industry?

A case study on network orchestration in the Dutch water technology sector

Msc BA Master Thesis Strategy and Innovation

By Anne Gierveld University of Groningen Faculty of Economics and Business

Msc Business Administration Strategy and Innovation Supervised by dr. R. van der Eijk and dr. W.A. Dolfsma

Groningen, December 2013

(2)

Preface

This thesis is written to complete my master degree in Strategy and Innovation at the University of Groningen. The writing of a thesis is an individual process but you cannot do it alone, therefore I would like to thank all those who helped in carrying out the research.

First, I would like to thank Aleid van Diepeveen en Paul van Koppen from the Netherlands Water Partnership (NWP) for their assistance with the selection of a relevant case. During the research, I had the opportunity to conduct my case study at the Centre of Expertise in South Africa.

Therefore I would like to thank Leo Meijer for giving me this opportunity, and his assistance in introducing me to the relevant parties.

Furthermore, I would like to thank my first supervisor, Rene van der Eijk, for his support and insights during the writing process. I also would like to thank Jasmijn en Lex for their useful feedback and comments on my thesis.

Finally I would like to thank all the interviewees for their willingness to cooperate. I experienced a lot of openness and preparedness to help.

Anne Gierveld

(3)

Executive summary

Collaboration needs some kind of coordination. This descriptive case-study investigates how a relatively central player coordinates a network constellation in a fragmented industry. This research focuses on the Dutch water technology sector since this sector is highly fragmented, consisting of many small and medium sized enterprises. The aim of this research is to provide empirical evidence on how organizations in fragmented industries collaborate in order to become active in an global market by focusing on the role of relatively central players in a network.

A relatively central player is an organization that possesses a central position in the network structure and uses this position to perform a leadership role. Social network analysis (i.e. degree centrality) is used to identify a relatively central player in this network. For this research a commercialization network is selected (n=13). The network members have a common goal;

“Match the South African water knowledge and technology needs with the Dutch water knowledge and innovations’’. Semi-structured in-depth interviews with all the network members (n=13) were conducted in order to identify which coordination mechanism this relatively central player uses to orchestrate the network constellation.

This research identifies five mechanisms that a central player uses to coordinate the exchanges in the network; 1) communication, 2) trust 3) division of benefits, 4) guarantees and 5) conflict resolution. Related to these mechanisms, specific tools are identified which the central player uses to coordinate the exchanges in the network. It seems that the type of coordination mechanisms employed are dependent on the type of relationship, the level of dependency, the specific (international) context and the type of network.

For practitioners, this case study provides an example of how organizations in a fragmented industry can cooperate to enter an international market. For scholars this research provides empirical evidence about the role of central players in (commercialization) networks, which is lacking in the academic literature.

Keywords: networks, network orchestration, leadership in networks, degree centrality

(4)

1. Introduction

As a result of the ongoing increased division of labor and specialization, organizations do not longer possess all available knowledge and skills required themselves. Therefore the need for the exchange of resources and cooperation has increased (Dolfsma and van der Eijk, 2010). The effects of cooperation have been studied in a wide variety of literature. Strategy literature emphasizes that organizations are motivated to cooperate in order to acquire resources that cannot be developed internally, but which are needed to survive in a highly competitive environment (Powell et al., 1996). The literature on learning and innovation argues that collaboration can facilitate the creation of new knowledge (Powell et al., 1996; Larsson et al., 1998 and Gulati, 1999). Additionally, literature on networks and social capital suggests that collaboration can affect the structure of inter-organizational relationships, making some organizations more central and influential compared to other organizations (e.g., Dyer 1996, Wasserman and Galaskiewicz, 1994; Hardy et al., 2003).

Despite the fact that the need for the exchange of resources and cooperation has increased, fragmented industries impose barriers to collective action (Dollinger, 1990). According to Porter (1980, p. 191) a fragmented industry is a situation in which “no firm has a significant market share and can strongly influence the industry outcome’’. In the literature, organizations in fragmented industries typically have been characterized as small, competitively weak and subject to intensive rivalry, lacking the economic power to control critical resources (Dollinger, 1990). Factors that cause fragmentation are low entry barriers, a lack of economies of scale or scope within production or distribution; diverse market needs and high exit barriers (Borch and Brastad, 2003). Collaboration is difficult in fragmented environments, and is therefore recognized as a strategic option in concentrated and oligopolistic markets (Dollinger, 1990). The low entry barriers ensure that there will be many firms, which makes information exchanges difficult. However, Dollinger (1990) dismissed the notion that a collective strategy in fragmented industries does not exist and states that there is considerable evidence that firms in fragmented industries use similar strategies. Research on the strategic behavior of companies in fragmented environments is lacking (Borch and Brastad, 2003).

This research contributes to the existing literature by providing insights in the strategic

behavior of organizations in a fragmented industry that faces certain problems due to this

fragmentation: the Dutch water technology sector. The next paragraph explains why the Dutch

water technology sector is chosen as research context.

(6)

1.1 Research context: the Dutch water technology sector

Bring in the Dutch! This was the headline of several newspapers in the United States after Hurricane Katrina devastated New Orleans in 2005. Despite the worldwide reputation on water management, the Dutch water sector faces difficulties to match the demand of the international market (Gleijm et al., 2013, NWP, 2006). One of the weaknesses of the Dutch water sector is that many parties operate independently and have no strong mutual connections (NWP, 2011).

Especially the water technology cluster is fragmented since it consists of many small and medium sized enterprises SME's (between 1.000-1.500) that are active in different parts of the value chain, e.g. suppliers, consultants, engineers and contractors (see appendix 1). In the table below (table 1) the number of jobs related to the water technology sector in the private sector are described. It seems that around 70% of the private sector has 1 to 10 employees.

Table 1: Number of jobs related to the water sector (NWP, 2012) Percentages of the private sector

(estimated) Number of employees

70% 1-10 (the average is 4)

30% 10-100

8% >100

The above numbers seem to fit the definition of Dess (1987), who states that a fragmented industry is characterized by many SME's. According to the European commission a firm can be indicated as an SME when it has less than 250 employees (see appendix 2). Using this definition, more than 92% of the private sector organizations can be indicated as SME's (see table 1).

Empirically, a fragmented industry has been defined as one in which the top-four-firm concentration ratio (CR) is 40% or less (Porter, 1980). It is estimated that in the water technology industry the contribution of the four largest firms (the CR) is about 25% of the total turnover (see appendix 2). Although the CR provides an incomplete picture since the definition does not include all the market shares in the industry, it is useful to give an indication about the level of fragmentation in an industry. Because the water technology sector has a low CR ratio, combined with the large amount of SME's, it can be concluded that this sector is fragmented.

This is confirmed by several documents (NWP 2006, 2011, 2013) and interviews with several experts (interviewees 1-5, see appendix 3).

1.2 Problem statement and research question

The international market demands total solutions (NWP, 2006), while the water (technology)

sector is not able to meet this demand due to fragmentation and insufficient collaboration

between the different parties in the value chain (NWP, 2006). Orientating interviews with

experts (interviewee 1-5 ) in the water technology sector were conducted in order to validate

the research (see appendix 3). These experts agreed that the sector is fragmented and that this

(7)

fragmentation negatively impacts the internationalization of organizations. As one of the interviewees stated; "The Dutch water sector is fragmented, it is interesting to investigate how the parties in the Dutch water sector can provide a collective offering" (interviewee 2). Currently most organizations in the Dutch water sector are operating on a solitary basis (Aqua Nederland and VLM, 2011). In the past some initiatives were taken to set-up international consortia but this did not succeed (Interviewee 1). At the same time, two foreign major international operators, Veolia Environment and Suez Environment are able to meet this demand for total solutions by acting as system integrators (Global Water Intelligence, 2011). A system integrator is a single prime contractor organization that is responsible for designing and integrating externally supplied products and services into a system for an individual customer (Davies et al., 2007). The role of the system integrators comprises two main activities 1) setting-up a network of various organizations and 2) coordinating the work of the organizations involved in the network (Rutten et al., 2009). However, since the Dutch water technology sector is fragmented, such a system integrator remains absent in the current research context .

It seems there is a high need to cooperate within the Dutch water technology sector in order to participate in water projects (interviewee 1,2 and 4). Since individual organizations often only have partially overlapping goals and interests, exchange and cooperation need some kind of coordination (Dolfsma and van der Eijk, 2010). The literature describes different mechanisms to coordinate collaboration such as strategic groups (Hunt, 1972), industry architecture (Oscan and Eisenhardt, 2009) and network orchestration (Müller-Seitz, 2012). In this research context, a relevant managerial question is how collaboration is coordinated without the presence of a 'system integrator'. However, the absence of a system integrator does not automatically mean that there are no relatively central players who can perform the role of network orchestrator.

Dhanaraj and Parkhe (2006) define network orchestration as the set of deliberate and a purposeful action undertaken by the network orchestrator as it seeks to create value and extract value from the network. In this research the term 'central player' is used for a network orchestrator, which is an organization that performs a leadership’s role in pulling together the dispersed resources and capabilities of network members (Dhanaraj and Parkhe, 2006). This research investigates the following question: How does a relatively central player orchestrate a network constellation in a fragmented industry? This research question can be divided in the following sub-questions:

1. What is the definition of a network?

2. Which types of network constellations can be distinguished?

3. Which coordination mechanisms can be used in a network constellation?

4. What is the role of a relatively central player in coordinating a network constellation?

5. Which coordination mechanisms and tools do relatively central players use to orchestrate a

network constellation?

(8)

1.3 Aim of this research

It seems that fragmented industries impose barriers to collaboration. The aim of this research is to provide empirical insight in how organizations in fragmented industries are collaborating in order to become active in an international market by focusing on the role of a relatively central player. The research results in a description of the coordination mechanisms and tools that a central player uses to orchestrate a network constellation in the Dutch water technology sector.

1.4 Relevance of this research

From a managerial point of view, this research is relevant because it investigate how a relatively central player can facilitate organizations in a fragmented industry to become active in an international market. This research also has academic relevance since network orchestration is an under-investigated topic (Müller-Seitz, 2012; Park, 1996). The current literature focuses on the orchestration of research and development (R&D) networks (e.g. Dhanaraj and Parkhe, 2006) or the role of intermediary organizations that are not part of the network (e.g. Batterink et al., 2010). This research contributes to the existing literature by investigating the role of a network orchestrator that is part of a commercialization network. Furthermore, this research provides theoretical insight in the strategic behavior of companies in fragmented environments, which is lacking in the current literature (Borch and Brastad, 2003).

1.5 Scope and domain

This research observes which coordination mechanisms a central player uses to coordinate a network constellation. This is examined in a fragmented industry, namely the Dutch water technology sector, that faces problems due to this fragmentation. This research does not study whether network orchestration is the most appropriate mechanism to coordinate the collaboration in a fragmented industry. Therefore this research does not make a comparison between different mechanisms to coordinate collaboration. Furthermore, the research provides a description of the coordination mechanisms a central player uses rather than a prescriptive advice on which coordination mechanisms a central player should use.

1.6 Structure of the paper

This research is structured as follows. Chapter 2 provides a theoretical background on network

constellations, coordination mechanisms and the role of central players, answering sub-question

1, 2, 3 and 4. Chapter 3 describes the research design and execution. In Chapter 4 the case-study

results are described, providing an answer to sub-question 5. In chapter 5 these case study

results are analyzed. Finally, chapter 6 covers the conclusion, limitations and suggestions for

future research, answering the central research question.

(9)

2. Theoretical background

Research has paid increasing attention to the benefits that accrue to organizations from their networks (Gulati et al., 2011; Park, 1996). According to Park (1996) networks can reduce variety and uncertainty by forming loosely coupled interdependencies and exercising shared division of benefits. The division of labor allows network members to specialize in the value-creating activity supported by their own distinctive competence (Miles and Snow, 1986). Furthermore, networks allow firms to economize on information costs and accelerate innovation. Since the 1990s, the network perspective has received interest from academics, policy makers and managers concerned with the promotion of the innovative capacity of companies, regions, sector or nations (Conway and Steward, 2009). For example, Zeng et al., (2010) conducted an empirical research among 137 Chinese manufacturing SME’s finding significant positive relationships between inter-firm cooperation, cooperation with intermediary institutions, cooperation with research organizations and the innovative performance of SMEs. Furthermore, Burt (1992) demonstrated that networks generate informational advantages that can in turn steer actors toward beneficial actions. Moreover, networks can reduce economic uncertainty (e.g. securing resource supplies) which is the main motivation for the buyer-supplier networks that were popular in the 1970’s (Park, 1996).

2.1 Definition of a network

Traditionally the term network refers to a group of individuals within a particular community or organization, also called 'social networks'. Nowadays, the literature on networks is also focusing on a set of relationships within a group of organizations; 'organizational networks', such as formal alliance agreements. According to Conway and Steward (2009), both social and organizational networks are important in order to understand how innovation and knowledge is shaped, diffuses and emerges. This research focuses on inter-organizational networks. The current literature on inter-organizational networks is related to;



Different theoretical perspectives e.g. economics (resource dependency theory and transaction cost economics (Pfeffer and Salancik, 1978; Williamson, 1979) and sociology (e.g. Dolfsma et al., 2009 and Uzzi, 1997)



Different forms of collaborations e.g. joint ventures, strategic alliances, inter- organizational relationships and R&D consortia (e.g. Fowler and Reisenwitz, 2013;

Grandori and Soda, 1995; Provan et al., 2007).



Different levels of analysis i.e. dyad, egocentric and network level (Provan, et al., 2007)

(10)

Networks have been defined in different ways (see table 2), for the purpose of this research, the definition of Provan and Kenis (2008) is used since this definition is suited for the selection of a relevant case without the need for an extensive preliminary research (see chapter 3: Research design). Furthermore, this definition is not limited to firms but also on other types of organizations such as; public institutions, knowledge institutions and non-governmental organizations (NGO's). The definition encompasses a wide range of relations including joint ventures and consortia.

Table 2: Different definitions from the literature

Term Definition Reference

Network “Repetitive exchanges between a set of autonomous but interdependent organizations in order to achieve individual and shared objectives”

Whelan (2011), p.267/277 Network “Groups of three or more legally autonomous

organizations that work together to achieve not only their own goals but also a collective goal”

Provan and Kenis (2008), p.231 Inter-

organizational network

“A social system in which the joint activities of at least three independent legal entities are reflexively coordinated on a repeated basis, in order that joint benefits may accrue to all the parties concerned”

Müller-Seitz (2012), p.429

In addition to this definition, the component parts of which a network comprises (i.e. actors, links, flows and mechanisms) described by Conway and Steward (2009) are used to describe the network selected in this research (Chapter 4, §4.1).

2.2 Types of network constellations

Network constellations can emerge when organizations face similar environmental conditions and joint forces (emergent) or when a triggering entity is responsible for setting up a network (engineered) (Müller-Seitz, 2012). The literature distinguishes different types of network constellations (e.g. Powell, 1990; Kenis et al., 2007; Müller-Seitz, 2012). Powell’s (1990) research on exchanges suggests that transactions can take place (a) through loose collections of individuals with impersonal and constantly shifting exchange ties, as in market, or (b) through stable networks of exchange partners who maintain close relationships (embedded ties) (Uzzi, 1997). Müller-Seitz (2012) distinguished between (a) an 'heterarchical' network in which the participants govern the network, (b) an 'hierarchical' network in which the network is governed by a lead organization, and (c) a mixed form like in a network governance organization.

Furthermore, networks can also vary in the level of fragmentation (from highly fragmented

networks with many small players to networks with only a few large players) and structural

holes, the level of embeddedeness, cluster and cliques, centralization and type of governance

(Provan et al., 2007; Conway and Steward, 2009). These dimensions are explained in the table

below (table 3).

(11)

Table 3: Types of network constellations

Dimension Definition References

Level of

fragmentation and structural holes

Network members can be connected, or a network can be brokered into fragments of unconnected organizations (e.g. cliques)

Provan et al., 2007

Level of

Embeddedness The degree to which organizational relationships are reinforced by social relationships

Conway and Steward (2009)

Cluster and Cliques Densely connected regions in a network Conway and Steward (2009) Centralization The extent to which one or a few organizations

in the network are more centrally connected than others

Provan et al., 2007

Type of governance The mechanisms used to govern and/or manage

the network Provan et al., 2007

This research focuses on the level of centralization and the type of governance. According to Provan and Kenis (2008) a network can be categorized along two dimensions, the level of brokering and the level of participation of the network members in the governance process (see table 4). This research focuses on the role that a lead organization, or central player, assumes in a network.

Table 4: Network governance forms (Provan and Kenis, 2008)

Governance form Level of brokering Participant governed versus externally governed

Shared

governance Low i.e. networks is completely governed by the organizations that comprise the network (dense and highly decentralized network)

Participant governed i.e. collectively governed by members themselves

Lead organization Medium i.e. single organization might take some key governance activities while leaving others to network members.

Participant governed i.e. a single network participant takes the role of lead organization

Network administrative organization

Highly brokered i.e. few direct organization-to-organization

interactions except some operational issues

Externally governed i.e. either voluntarily or mandated

2.3 The definition of a central player

As mentioned in the introduction, individuals only have partly overlapping goals and interests, therefore exchange and cooperation need some kind of coordination (Dolfsma and van der Eijk, 2010). Coordination is essential in networks since it reduces uncertainty and opportunistic behavior. Coordinating multiple players to accomplish a common goal is difficult because of the continuous pursuit of individual goals by the partners and the managerial complexity involved in the process (Park, 1996). Müller-Seitz (2012) conducted a literature review about the ways in which inter-organizational networks coordinate their activities for the benefit of all the parties.

In the literature different terminologies are used to describe an organization that performs a

leadership role in coordinating the network (see table 5).

(12)

Table 5: Terminologies for leadership in networks

Term References

Hub-firm Wasserman and Galaskiewicz, 1994

Triggering entity Doz et al., 2000

Strategic center Lorenzoni and Baden-Fuller, 1995 Network orchestrator Müller-Seitz, 2012

In this research, an organization that performs a leadership role is called a central player. A central player is defined as an organization that possesses prominence and power gained through a central position in the network structure and uses its prominence and power to perform a leadership role in pulling together the dispersed resources and capabilities of the network members (Dhanaraj and Parkhe, 2006). This term is related to leadership. Müller-Seitz (2012) defines leadership as the exertion of influence to 'make things happen' despite the lack of formal authority.

2.3.1 The characteristics of central players

Firms described as fulfilling the role as central player are typically large and dominant firms (Batterink et al., 2010, p.51). Nevertheless, in the case of SME innovation networks this situation is different since “SME's do not have all the relevant capabilities for fulfilling the network orchestration role successfully” (Batterink et al., 2010, p.51). In SME innovation networks, the orchestration role is often executed by a third-party that is not part of the network, a so-called innovation intermediary (Batterink et al., 2010). Such intermediaries focus on facilitating innovation by fulfilling the role of broker between cooperating SME's. Contrary to this common practice, this research focuses on relatively central players which actively take part in the network themselves.

The literature identifies different advantages and roles for central players. Inquired and Hanneman (2006) argue that actors who have many ties with other actors may be in an advantageous position, “a central actor has greater access to external assets, such as technology, money and management skills” (Gnyawau and Madhavan, 2001, p.435). Second, being at the confluence of a larger number of information sources through their ties, central actors are likely to receive new information sooner than less central actors (Gnyawau and Madhavan, 2001).

Third, high centrality implies higher status and power. Therefore, a central actor generally has

access to better and more resources and opportunities (Gnyawau and Madhavan, 2001). The

superior status and the power associated with centrality strengthen the tendency to take the

initiative (Gnyawau and Madhavan, 2001). Central actors are often third parties and deal makers

in the exchanges among others, and are able to benefit from this brokerage (Hanneman and

Riddle, 2005). Conway and Steward (2009) state that an organization with a high number of

(13)

links to others in the network (called a 'star') can play an important role in diffusing information.

2.4 The type of coordination mechanisms used by the central player

From the perspective of a central player, value must be created and extracted from the network.

Dhanaraj and Parkhe (2006) define 'network orchestration' as the set of deliberate and purposeful actions undertaken by the central player as it seeks to create value and extra value from the network. This term is sensitive to the dynamics of how lead organizations actually influence other network members (Müller-Seitz, 2012). It is important for leading actors to accurately perceive the network relations that connect people and to actively manage these relations (Balkundi and Kilduff, 2006). However, research indicates that, “even leading actors, having access to superior resources are by no means capable of actually steering the respective network entirely” (Müller-Seitz, 2012, p.437). Huxham and Vangen (2000) argue that central players can only make things happen by attempting to influence structures, processes and participants. The literature has focused attention on activities related to the role of central players in coordinating the network (e.g. Dhanaraj and Parkhe, 2006; Batterink et al., 2010, Gardet and Fraiha, 2012). Several scholars make a distinction between network design activities and ongoing network management activities. In terms of designing a network, the central player is engaged in network membership, network structure and network position (Dhanaraj and Parkhe, 2006). Once the network is created, a central player may employ orchestration processes in order to direct the core activities in the network in which it operates (Batterink et al., 2010, p.53). Furthermore, Dekker (2004) identified formal (outcome control) and informal (behavioral, control and social control) control mechanisms in collaborations. Poppo and Zenger (2002) developed and tested the idea that formal contracts and relational governance complement each other as coordination tools.

Gardet and Fraiha (2012) identified five coordination mechanisms used by a hub-firm (in their case an SME) in a network (see table 6). Coordination mechanisms are factors or processes that companies use to coordinate their transactions. In this research the coordination mechanisms discussed by Gardet and Fraiha (2012) are investigated since these provide a clear overview of the available coordination mechanisms. Furthermore, these coordination mechanisms are also discussed by other authors (see table 6).

Table 6: Coordination mechanisms identified by Gardet and Fraiha (2012) Coordination mechanisms References

Communication Dhanaraj and Parkhe (2006); Batterink et al., (2010) Trust Dekker (2004); Dhanaraj and Parkhe (2006); Gausdal and

Hildrum (2012); Uzzi (1997) Division of benefits Dhanaraj and Parkhe (2006)

(14)

2.4.1 Communication

The literature recognizes both formal and informal communication. Formal communication means that exchanges are explicit and written e.g. standardized procedures, technical reports, confidentially agreements and contracts. In informal communication exchanges are implicit and verbal and include the implementation of common teams, seminars, meetings and certain mechanisms for decision-making (Gardet and Fraiha, 2012).

In order to create value, it is essential that specialized knowledge of network members flows across organizational boundaries. Dyer and Nobeoka (2000), Dhanaraj and Parkhe (2006) and Müller-Seitz (2012) indicate the importance of knowledge-developing routines arguing that network orchestrators should design networks that allow opportunities for knowledge exchange. Knowledge mobility can help to learn from the partners and exploit resources that are made available. Dhanaraj and Parkhe (2006, p.660) define knowledge mobility as the “ease with which knowledge is shared, acquired and deployed within the network”. A central player can promote knowledge mobility within a network by using formal, semiformal informal communication tools. Table 7 provides an overview of the tools that a central player can use to enhance promote knowledge mobility.

Table 7: Overview of the communication tools (Gardet and Fraiha, 2012) Formal

communication Informal communication Semiformal communication Technical reports Implementation of common

teams Loose contracts

Standardized

procedures Seminars Reports

Confidentiality

agreements Meetings Meetings

Contracts Staff transfers Documentation

2.4.2 Trust

Several researchers indicate that trust is a crucial condition for successful collaboration in networks (Müller-Seitz, 2012 and Gausdal and Hildrum, 2012). In an empirical research (Uzzi, 1997) respondents viewed trust as an explicit and primary feature of their embedded ties.

Trust is defined as the “expectation that an exchange partner will act honestly and in good faith can alleviate fears of opportunistic behavior” (Dolfsma and van der Eijk, 2010, p.63). Trust is often found in inter-organizational relationships with a long history of prior interaction.

Scholars identify several types of trust. For example, Das and Teng (2001) make a distinction between goodwill trust i.e. expectation that another will perform in the interest of the relationship, it relates to not behaving opportunistically and competence trust i.e. expectations

Guarantees Jones et al. (1997)

Conflict resolution Mohr and Spekman (1994); Dhanaraj and Parkhe (2006);

Batterink et al. (2010)

(15)

about another's competencies to perform a task satisfactory. Nooteboom (1996) notes that trust may concern a partners' ability to perform according to agreements (competence trust) or his intention to do so (goodwill trust).

A central player can use different tools to build trust. Granovetter (1985) claims that trust cannot be intentionally created. However, several scholars believe that building trust can be facilitated (Gausdal and Hildrum, 2012). Trust can be developed when extra effort is voluntary given and reciprocated (Uzzi, 1997). This type of effort, often called 'gifts' or 'favors' might involve giving an exchange partner a preferred treatment. This component of the exchange relation is important since it enriches the firm’s opportunities and provides access to resources.

Dekker (2004) identifies risk taking and increasing interaction as trust building mechanisms, for example by joint goal setting, decision making, problem solving and partner development activities (Dekker, 2004). Close interaction results in commitment and decreases the likelihood of opportunistic behavior. Moreover, interactions enable partners to learn about another's skills and expectations and to establish standardized communication and routines (Jones et al., 1997).

Das and Teng (1998) identified risk taking and communication as trust building mechanisms.

A central player must provide leadership in “building trust levels and in communicating clear, pre-established sanctions for trust violation” (Batterink et al., 2010, p.51). Communication and proactive information exchange boost trust among network partners (Das and Teng, 1998). The table below (table 8) provides an overview of tools that can be used to build trust.

Table 8: Overview of tools to build trust

Tools to build trust References

Gifts Uzzi, 1997

Communication Das and Teng, 1998

Risk taking Das and Teng, 1998; Dekker, 2004

Interaction Dekker, 2004; Jones et al., 1997

2.4.3 Division of benefits

There may be a high variability in the outcome of network activities carried out by network

partners (Dhanaraj and Parkhe, 2006). An essential element of cooperation is the rules for

distributing or sharing the benefits or outcomes of a project (Gardet and Fraiha, 2012). Gardet

and Fraiha (2012) make a distinction between an equitable and equal division of benefits. An

equitable division of benefits means that each member’s payoffs are a function of its resources

i.e. the greater the party’s contribution to the collaboration, the greater is its payoff. An equitable

division of incentives is perceived as an incentive for the network members to work harder. On

the other hand, an equal division of benefits means that every member of a project receives an

equal share of the results, no matter how much they contribute in terms of resources and/or

expertise. According to Batterink et al., (2010, p.51), “managing value creation and appropriation

relates to the idea that in a network equitable distribution of value must be ensured”.

(16)

Appropriation refers to the degree to which firms capture the profit associated with their innovative activity (Cohen and Levinthal, 1989). Equal distribution of value brings complications in terms of free riding e.g. when a network member does not bring in the best ideas to the network but enjoys the benefits of knowledge flows and opportunistic behavior such as taking advantage of the openness of others in the network.

According to Doz et al. (2000) a triggering entity may be required to lessen the concerns of potential participants that the costs and benefits of collaboration will be shared fairly. Central players can ensure equitable distribution of value by focusing on the following processes;

procedural justice and joint asset ownership (Dhanaraj and Parkhe, 2006). Procedural justice, the fairness of the decision process, has a strong and positive impact on voluntary cooperation and it discourages hoarding of ideas. In order to create procedural justice a central player might employ different mechanisms e.g. bilateral communication and consistency in the decision- making process. Studies on innovation networks discuss equity joint ventures and patent pooling as forms of joint asset ownership. Joint asset ownership between a central player and others in the network enhances appropriability since it creates a context for joint problem solving arrangements and it enhances the commitment of actors toward shared goals and provides incentives for sharing rewards. Table 9 provides an overview of the tools to divide benefits.

Table 9: Overview of tools to divided benefits (Dhanaraj and Parkhe, 2006)

Tools Examples

Procedural justice Bilateral communication, consistency in decision-making Joint asset ownership Joint ventures, patent pooling

2.4.4 Guarantees

In a network, firms may invest in assets, procedures and training, and these act as a motivation to continue the cooperation (Gardet and Fraiha, 2012). However, the value of some investments may decrease over time, which can increase opportunistic behavior. Collective sanctioning and reputation can be considered as safeguards against opportunistic behavior (Gardet and Fraiha, 2012), “Collective sanctions involve group members punishing other members who violate group norms, values or goals” (Jones et al., 1997, p.931); These collective sanctions can range from gossip to exclusion from the network. Reputation involves an “estimation of one's character, skills, reliability and other attributes important under exchange conditions of uncertainty and customization” (Jones et al., 1997, p.932). Reputation reduces behavioral uncertainty by providing information about the reliability and goodwill of others (Jones et al., 1997). However, reputations have limitations since the information may be inaccurate. Also an overreliance on reputation may reduce new and innovative information (Jones et al., 1997).

Gardet and Fraiha (2012) differentiate direct guarantees from indirect guarantees against

opportunistic behavior. A direct guarantee may include that the non-cooperating member will

(17)

be expelled from the network and a indirect guarantee may include that a network member will be excluded from a future business opportunity. A mixture of the two tools can be used as well.

For example an organization may pay a penalty for damage and may as well be excluded from future collaboration opportunities. The table below (table 10) provides an overview of different tools for guarantees.

Table 10: Overview of the tools for guarantees (Gardet and Fraiha, 2012)

Tools Explanation

No guarantees -

Direct guarantee Immediate impact e.g. collective sanctioning Indirect guarantee Deferred impact e.g. affect reputation and future

business opportunities Mixture of indirect and indirect

guarantees Both an immediate and deferred impact

2.4.5 Conflict resolution

Due to inherent interdependencies between parties in inter-organizational relationships a certain amount of conflict is expected (Mohr and Spekman, 1994). Therefore it is important to understand how such conflicts can be resolved.

A central player can be concerned with resolving conflicts between network members. The impact of conflict resolution can be productive or destructive (Mohr and Spekman, 1994). When parties engage in joint problem solving, a mutually satisfactory solution may be reached which will in general be more constructive than the use of destructive conflict resolution techniques such as confrontation. Ideally, networks are composed in such a way that conflicts are unlikely to occur (Dhanaraj and Parkhe, 2006). However, when conflicts occur it is suggested that third- party innovation brokers can play an important role in handling them because “their experience with previous innovation networks and because of their neutral position in the network” (Batterink et al., 2010, p.64). Winch and Courtney (2007, p.751) describe an innovation broker as “an organization acting as a member of a network of actors in an industrial sector that is focused neither on the generation nor the implementation of innovations, but on enabling other organizations to innovate”. In the research of Batterink et al. (2010) on conflict handling between innovation brokers, the importance of a personal approach in an attempt to solve a conflict was emphasized. Gardet and Fraiha (2012) identified five conflict resolution tools (see table 11).

Table 11: Overview of conflict tools (Gardet and Fraiha, 2012) Conflict resolution tools Explanation

Joint resolution of

problems Members make a commitment to find a solution

Persuasion One of the members tries to persuade the other members

Pressure One of the members force the others to choose a particular option Penalty The non-conforming partner is penalized and excluded from the

network

(18)

Introduction of a third

party A mediation between the network members takes place and may include legal action

2.5 Conclusion

This chapter defines the scope of this research, provides an overview of the current literature on networks and the orchestration of networks, and answers sub-question 1, 2, 3 and 4. The theoretical concepts discussed in this chapter provide the input for the empirical research (see table 12). Chapter 3 provides an explanation of how the theoretical concepts are applied in this research in order to answer sub-question 5 (see chapter 4).

Table 12: Theoretical input for the empirical research Theoretical

concepts Description

Network (§2.1) Groups of three of more legally autonomous organizations that work together to achieve not only their own goals but also a collective goal

Components of a

network (§2.1) Actors, links, flows and mechanisms Central player

(§2.3) An organization that possess prominence and power gained through a central position in the network structure, and uses its prominence and power to perform a leadership role in pulling together dispersed resources and capabilities of network members

Coordination mechanisms and tools

(§2.4)

Communication Contracts, seminars, meetings, documentation, reports, standardized procedures, confidentially agreements, common teams, staff transfers

Trust Interaction, gifts, communication and risk taking Division of

benefits Procedural justice and joint asset ownership

Guarantees No guarantees, direct guarantees, indirect guarantees or mixture of direct and indirect guarantees

Conflict

resolution Joint resolution of problems, persuasion, pressure, penalty and introduction of a third party

(19)

3. Research design

The research consists of a literature research and empirically collected data. The literature research entails secondary data from scientific journals, researches, websites, newspapers and magazines. The scientific literature is acquired via different scientific search engines such as Google Scholar and Business Source Premier. Furthermore, data such as newspaper and magazine articles were obtained using LexisNexis and Google.

The first paragraph (§3.1) provided a general description of case studies research as a methodological research strategy. This is followed by a description of the empirical part of this research consisting of two phases. The first phase is selecting a relevant network (§3.2) and the second phase is to identify a relatively central player and investigating which coordination mechanisms and tools this relatively central players uses to orchestrate a network constellation (§3.3). Furthermore, the data collection (§3.4) and data measures (§3.5) are described. Finally, the reliability and validity of the results are discussed (§3.6).

3.1 Case study research

A case study design is employed to answer the last sub-question; which coordination mechanisms and tools do relatively central players use to orchestrate a network constellation? A case study is considered relevant when "a how or why question is being asked about a contemporary set of events over which the investigator has little or no control" (Yin, 2014, p.14). Furthermore, Eisenhardt (1989, p.534) describes a case study as "a research strategy, which focuses on understanding the dynamics present within single settings". The aim of this research is to investigate how a relatively central player coordinates a network constellation in a fragmented market. Since this involves a 'how question', a case study is the relevant research methodology.

Case studies can provide detailed information useful for an intensive analysis of a case and can

involve either single or multiple cases, and numerous levels of analysis (Yin, 2014). This

research is focusing on a single case (a specific network). Since the current literature lacks

evidence about the role of central players in coordinating exchanges inside the network, a

detailed approach is necessary. Furthermore, case studies typically combine data collection

methods such as archives, interviews, questionnaires and observations (Eisenhardt, 1989). In

this research different sources are used: documentation is reviewed, interviews are conducted

and a questionnaire is used in order to improve the construct validity. This research combines

both qualitative and quantitative data. Case studies are often seen as a synonym for qualitative

research. However, Eisenhardt (1989) and Yin (2014) show that case study evidence does not

necessarily need to originate from purely qualitative data (e.g. interviews). The evidence of this

research is based on a combination of qualitative (interviews) and quantitative (centrality

(20)

measurement) research. According to Yin (2014) this research can be classified as a descriptive case research, which is useful to illustrate events and their specific context. Case studies are not sufficient to support or reject a theory (Blumberg et al., 2005). However it is a useful approach to increase knowledge of an under-investigated topic.

3.2 The selection of a network

Since networks lack natural boundaries, network research frequently requires the setting of artificial boundaries (Conway and Steward, 2009). A researcher of networks can set these boundaries by focusing on relationships between a group of actors (socio-centered) or only include those relationship connected to a specific actor (ego-centered). It is also common for network researchers to focus on a specific type of exchanges or transactions e.g. information, friendships or power. Additionally in order to focus even more, researchers can limit a network that relates to a specific activity e.g. development of a new product.

The water technology cluster consists of a lot of small organizations (between 1.000-1.500). In order to limit the scope of the network, it is decided to investigate a 'sub-network' in the water technology sector. In order to identify an interesting sub-network suitable for this research, a few requirements were identified. First (as stated in the introduction) the definition of Provan and Kenis (2008) is used to identify a network without conducting an extensive preliminary research. According to this definition a network has the following characteristics:



Groups of three or more legally autonomous organizations;



that work together;



to achieve their own goals but also a collective goal.

Secondly, in order to identify a central player, the network must contain one or a few relatively central actors. Furthermore, it is decided to select a network with a limited scope. This research excludes loosely structured, large-scale analyses (n>100) that neglect the interplay between central players and network members (Lazzarini, 2007). Lastly, since the main objective of this research is to investigate "how a fragmented supplier base, can orchestrate the participation in water projects in a global market" (§ 1.2 p.6) a network is selected that focuses on the commercialization of products/technologies in a global market.

In order to identify a network that met these requirements, several interviews with experts

(interviewee 1, 5) in the water technology sector were conducted. The experts are all employees

of the same Dutch network organization; The Netherlands Water Partnership (NWP). This

organization aims to create a water network with all the parties active in the Dutch water sector

(see appendix 6). It is decided to select the experts from the NWP because of the large network

of NWP (200 members) and the researcher’s personal connection with the organization. A

drawback of this sampling method (convenience sampling) is that the sample might be biased

(21)

since the experts are all members of the same organization. In the preliminary research, four networks that the employees of NWP were familiar with, were identified. As a result of the orientating interviews (interviewee 1, 5) and available documentation, a network is selected that meets the requirements of this research; the Centre of Expertise (CoE) in South Africa (see table 13, appendix 4).

Table 13: Requirements for the selection of a network

Networks Convenant

Informatie Keten voor Water en Klimaat (CIWK)

Centre of Expertise (CoE) South Africa

AquaNL Water

Alliance

Requirements

Groups of three of more legally

autonomous organizations Yes Yes Yes Yes

That work together Yes Yes No Yes

Shared goal Yes Yes Yes Yes

Network size ≤100 Yes Yes Yes Yes

Presence of a relatively central

player Unknown Yes Unknown Unknown

Commercialization of products and techniques in a global market

Yes Yes No No

3.2.1 Case study: The Centre of Expertise in South Africa

The CoE is an initiative by eThekwini Water and Sanitation (EWS) and various Dutch partners.

The aim of the CoE is to "match the South African water knowledge and technology needs with the Dutch water knowledge and innovations"

¹

. Under the CoE, a series of pilot projects are executed to showcase the implementation of new technologies and knowledge at EWS. For this individual pilot project the contribution and support from various partners is combined. The CoE is officially launched on May 8th 2012, during the WISA conference in Cape Town. The CoE is funded by the Dutch government through the transition facility

²

.

The CoE is an export-improving concept that helps to improve and accelerate the market entry of innovations in the public sector (interviewee 11, 12, 14, 19). According to one of the experts in the Dutch water sector the CoE is a “unique construction that can facilitate organizations in a fragmented industry to become active in international market” (Interviewee, 5).

A strength of this concept is that the products are shown/tested in the local market (interviewee 5, 6, 7, 8, 9, 10, 11, 12, 13); "The strength of this initiative is that it is demand driven and the products are directly applied to the local market" (interviewee 5), as another network member said: "Seeing is believing" (interviewee 12). This is especially important for Dutch technologies since the technologies are (or are perceived to be) more expensive compared to other countries

1 The information in this section (case description) is from the website of the Centre of Expertise: http://www.coe.org.za/about

2 The transition facility is a cooperation between the Dutch ministries of Foreign Affairs and Economic Affairs, agriculture and innovation. In this co-operation existing budget is used to transform traditional aid relationships into mutual beneficial trade relationships.

(22)

such as China. The showcases provide evidence that more expensive solutions generate less negative effects e.g. maintaining costs and fewer nuisances during the execution (interviewee 5).

Showcases are needed to show the value of Dutch technologies (interviewee 10, 14). Another strength of this concept is that the client-firm has a leading role in South Africa; "Since our municipality is seen as an example to the other municipalities, it gives innovators an entry to the other municipalities" (interviewee 6, 7, 8). Furthermore, the CoE contributes to lower the transaction costs (NWP, 2012) by creating a platform to fast-track bureaucratic tender procedures (interviewee 5 , 15) and to structure and organize the efforts of Dutch SME’s that want to enter the South African market (interviewee 13).

3.3 Coordination mechanisms and tools used by the central player

The second phase of the empirical research was to identify a relatively central player and investigate which mechanisms and tools the relatively central player uses to orchestrate a network constellation.

3.3.1 Identifying the central player

In order to identify a relatively

³

central player, social network analysis can be applied. In network analyses, a network can be defined as a set of actors, or “nodes”, that have relationships, or “ties” (Whelan, 2011). The idea of network analysis is to map the pattern of relationships between actors and to analyze the implications of these relationships for actors in the network and the network as a whole (Whelan, 2011). The centrality of a node in a network is a measure of its structural importance (Izquierdo and Hanneman, 2007) and refers to the position of an individual actor in the network (Gnyawau and Madhaven, 2001). The level of centrality is “the extent to which the focal actor occupies a strategic position in the network by virtue of being involved in many significant ties” (Wasserman and Faust, 1994, p.72). Izquierdo and Hanneman (2007) identified three approaches to calculate the centrality of a node: based on degree, closeness and on betweenness (see table 14)

Table 14: Mechanisms to calculate the centrality Centrality Explanation

Degree More ties means being more important

Closeness Actors who are able to reach others at shorter path lengths, or who are more reachable by others at shorter path lengths, are in favored positions

Betweenness It is being in between many actors what makes an actor central

This research used 'degree centrality' in order to measure whether an organization occupies a central or more peripheral position in the network based on the number of ties with other organizations. This research only differentiates the network members based on how many

3 The term 'relative' is used since the water technology sector is characterized with the absence of a 'system integrator' or 'focal actor'.

(23)

connections they have with other network members (undirected network). There are different methods to measure the degree centrality. Linton Freeman (1979) developed basic measures for the centrality of actors. This original degree centrality approach argues that actors that have more connections are more likely to be powerful. This measurement of Freeman (1979) is used since basic measures are sufficient to support the findings from the interviews.

In this research the in-degree centrality is measured, which means that only the connections mentioned by the other network members are used to measure the centrality. By measuring the in-degree, it can be prevented that the level of centrality is based on network members own perception of the collaboration. In order to measure the degree centrality this research made use of the social network analysis software UCINET 6 (Borgatti et al., 2002). The definition of Dhanaraj and Parkhe (2006) is used to identify a central player (see table 15). This research focused on relatively central players that are part of the network. In order to find the central player, the network members were asked to indicate with which organizations they collaborate.

This research focused on collaboration regarding the commercialization of technologies/products, in which the CoE plays a role (see table 15).

In order to validate the findings of the degree centrality method to identify the central player, the following qualitative control question is asked to each member of the network; Is there a network member that possesses a central position in the network? Subsequently it is asked which network member possesses this position. All the respondents indicated one of more network members as a central player. There was a possibility that none of the respondents indicate one of the other network members to be a central player. Therefore, the following question was asked: Which network member is most suited to execute a central role in the network?

3.3.2 Coordination mechanisms and tools

The next phase of the research was to investigate which coordination mechanisms the relatively

central player uses to orchestrate a network constellation . The coordination mechanisms and

tools identified in the literature review (Chapter 2, §2.4.3) are used to develop semi-structured

questions (see appendix 8). The interview consisted of both closed and open questions (see

table 15). An in-depth interview with the more central player identified in the previous research

phase is used in order to gain detailed insight into the mechanisms central players use in order

to orchestrate a network constellation. Additional supporting interviews with all the network

members are conducted in order to validate the findings of the central player. In-depth

interviews are most appropriate for situations in which a scholar wants to obtain detailed

information from relatively few people. The seven stages of conducting in-depth interviews

listed by Kvale (1996) are followed for the interview; thematizing, designing, interviewing,

transcribing, analyzing, verifying and reporting. An important element in conducting in-depth

(24)

interview is the skill and attribute of the interviewer. A good qualitative researcher should be open-minded, flexible and responsive, patient, observant and a good listener. Several strategies were used in order to be a good listener. For example, attending fully that is giving the full attention to the interviewee and paraphrasing what the interviewee is saying to confirm to the interviewee that the interviewer is actually listening. Furthermore, the responses were audio- recorded and complemented with written notes.

Table 15: Overview of variables measured (see also appendix 8)

Variables Definition of variables How

measured⁴ References Collaboration A recursive process where people or

organizations work together in an interaction of a common goal by sharing knowledge, learning and building consensus focused on the commercialization of

technologies in South-Africa, whereby the CoE plays a role.

Closed

question (6) Dietrich et al., 2010 (p.60)

Central player An organization that possess prominence and power gained through a central position in the network structure, and uses its prominence and power to perform a leadership role in pulling together dispersed resources and capabilities of network members

Closed

question (7) Dhanaraj and Parkhe (2006)

Coordination

mechanisms The coordination mechanisms mentioned in the theoretical background (§2.4.3) that the central player uses to orchestrate a network constellation

Open questions (8-13)

Gardet and Fraiha (2012)

3.4 Data collection

The first step of this research was to select a relevant network. Multiple sources of evidence were used to select this network including interviews, documentation and reports. Interviews with experts (n=5) in the water technology sector were conducted in order to identify a suitable network (see appendix 4). Furthermore, these orientating interviews were used to validate the interview questions used in later data collection. Semi-structured, in-depth interviews with the relevant network members (n=15) were conducted in order to identify a central player and to discover which coordination mechanisms and tools the central player used in order to orchestrate a network constellation. Appendix 5 provides an overview of the respondents.

3.5 Data measures

According to Hartley (2004) data collection and analysis are "developed together in an iterative process" which can be a strength, as it allows for theory development grounded in empirical evidence (Hartley, 2004, p.329). The collected data is analyzed by making a matrix of categories (a qualitative data matrix) and placing the evidence in such categories (Bijlsma-Frankema and

4 The numbers mentioned are related to the interview questions, see appendix 8

(25)

Droogleever Fortuijn's, 1997). The matrix provides an overview of the similarities and differences in the data collected from the interviewees (see appendix 10). Furthermore, the data collected during the interviews is compared with the academic literature in order to validate the case study (pattern matching). For case analysis, pattern matching is one of the most desirable techniques. If the pattern coincides, the results can help to strengthen the internal validity (Yin, 2014).

3.6 Reliability and validity

3.6.1 Reliability

Reliability in research designs means that the operations of a study (e.g. data collection procedure) can be repeated with the same results (Yin, 2014). A limitation of in-depth interviews is the possibility of interview bias i.e. the situation where an interviewee is influenced by the interviewer. In order to increase the consistency of observations and to minimize the errors and biases in a study, the interview procedure is narrowly documented.

Developing an interview guide is an import part of qualitative research; it serves as a memory list to ensure that the same issues are addressed in every interview (Kvale, 1996).

3.6.2. Validity

External validity- The external validity of a study refers to the degree to which a conclusion is generalizable across other settings, places and times (Cooper and Schindler, 2006). Since this research only focuses on one network (case) in a specific industry, this research has a relatively low external validity. However by interviewing multiple organizations that are involved with the same central player the external validity is enhanced since the experience of multiple organizations of a common central player are assessed. A limitation of case studies is the general applicability of the study; “Case studies do not attempt to give a represented picture of an issue, they still attempt to reveal certain effects or mechanisms that are likely to occur in similar settings”

A case study on network orchestration in the Dutch water technology sector

How does the head cattle lead its herd in a fragmented industry?