Dynamic Evaluation of Project Governance in Collaborative Innovation Projects: A Case of Industry Technology Research Institute

Abstract

Collaborative innovation projects (CIPs) are a typical type of inter-organizational collaboration project to deliver innovative results, in which diverse and dynamic collaborative relationships exist among stakeholders. The project’s success depends on appropriate governance mechanisms to coordinate the relationship between stakeholders. Contractual governance and relational governance both play an important role in successful project delivery. Existing research on the static evaluation method of project governance has obvious contradictions with the dynamic characteristics of stakeholder collaboration relationships during the innovation process. In response, this study proposes a dynamic evaluation method of CIP governance that uses stakeholder networks as the evaluation object, thus filling in gaps in the literature on CIP governance and contributing to the development of governance theory. A project of the Industry Technology Research Institute (ITRI) was chosen as a case study to evaluate the effectiveness of the procedure. The results show that contractual governance and relational governance both have a strong impact on stakeholder collaboration in the whole stage of CIPs. The governance effect is determined by the changeable contractual and relational governance structure, not just the stakeholder’s power of the general understanding. The findings have implications for the governance of CIPs and mainstream project governance research.

1. Introduction

Innovation is an effective way for enterprises to increase productivity and quickly obtain economic benefits [1,2]. The increased technological and managerial complexity puts massive restrictions on innovation efficiency and lengthens the production cycle of innovation output [3,4], thus lead achieving sustainable innovation becomes a major issue. A project is an important way for organizations today to realize much of their innovation value and deliver innovative products [5]. Collaborative innovation projects (CIPs) provide a systematic model that can prompt stakeholders to integrate internal and external resources to respond to innovation process challenges [6]. A sound project organization provides organizational support for the sustainability of enterprise innovation. There has been a growing recognition that an effective governance structure can restrict the stakeholders’ behavior by determining their relationships, underlying the project’s success [7,8]. Clarifying whether governance mechanisms are applicable is crucial for enterprises to achieve sustainable innovation through CIPs.

CIPs are carried out by public research organizations and enterprises in different fields and industries to complete innovation goals and deliver innovative products [9,10]. In China, the Industrial Technology Research Institute (ITRI) is an important organization undertaking major technological innovation activities. It has extensive connections with the government, innovative enterprises, scientific research institutions, etc. The ITRI CIPs are a typical type of CIPs. In these projects, tasks and resources are allocated based on stakeholder roles to maximize project profit [11]. The stakeholder theory holds that the stakeholders involved in a project are closely related to the project’s success [12,13]. Project governance decisions must consider the relationships among the stakeholders as they restrict each other, and no one can take charge of everything [14,15].

Project governance determines the interactions between project stakeholders and the engagement of the stakeholders in project governance [16,17]. For clarity, the stakeholder relationship structure is always abstracted by a project stakeholder network constituted of stakeholders (nodes) and their relationships (edges) [18]. The project stakeholder network is formed under a specific project governance mechanism; the better the project governance, the better the project stakeholder network structure. So, the state of the network could reflect the effect of governance. Although previous studies have addressed a wide range of different applications of project stakeholder networks [19,20,21,22], using that to evaluate the project governance of CIPs is feasible but remains ambiguous.

For this, researchers typically addressed the evaluation issue through an integrated static relationship model [23,24]. Unfortunately, the inherent cross-functional nature of CIPs demands adaptable collaboration among different stakeholders [25]. The relationship between stakeholders presents a characteristic that dynamically changes with the project lifecycle since the project tasks are quite different in every project stage [26]. To ensure the effectiveness of project governance, evaluating the governance network from a dynamic perspective is of significance, which has gained less attention in the literature.

In addition, it is not easy for project governance evaluation, as the relationships among stakeholders are complicated in CIPs [27]. An effective contractual relationship ensures that the stakeholders have sufficient motivation to complete the project tasks [28], and an adequate information relationship is a guarantee to give full play to stakeholders’ capabilities [29]. Projects governance provides an environment with contractual and information relationships in which stakeholders have both motivation and capabilities to implement their behaviors, which aligns with the governance theory that contractual and relational governance coexists in the project [30,31]. Therefore, a more systematic and accurate quantification method is required to characterize the relationships among stakeholders from multidimensional relationships.

Given the dynamic and diverse characteristics of the stakeholder relationship, a research question on how to conduct comprehensive project governance evaluation for CIPs is proposed. Therefore, this paper aims to provide a novel dynamic approach to evaluating project governance from the network perspectives for CIPs. According to the evaluation results, every stakeholder could adjust the way of participating in the project to achieve a better project governance effect. For this, the relationship between contractual governance and relational governance in CIPs needs to be clarified, and the network indicators for evaluating governance structures need to be screened and readjusted according to dynamic requirements. To achieve these objectives, the project stakeholder network is divided into contractual and information networks to quantify the governance status of the project [20,32]. Social Network Analysis (SNA), a widespread network analysis method [33,34,35], is adopted as the solution technique, whose indicators are tailored and improved in our work to meet the needs of comparative analysis in different stages of dynamic evaluation. Finally, we selected a typical ITRI CIP, which involves most organizational types of collaborative innovation activities, such as government, enterprises, universities, and research institutes, and has the characteristics of typical CIPs. Therefore, the application of the network model for this project can provide suggestions for the improvement of project governance in similar projects, and also provide a template for other CIPs to use this framework for project governance evaluation.

The outline of this paper is as follows. The literature on the governance of CIP and network-based analysis is reviewed in the next section. After that, a case study of the ITRI CIP, with obvious dynamic characteristics of the governance network, will be reported in Section 3 to illustrate the application progress of the evaluation method. Section 4 reports the results of the evaluation model and analyzes the governance issues based on the evaluation results. The discussion of the findings is presented in Section 5. Finally, the conclusions, limitations, and ideas for future research are presented in Section 6.

2. Literature Review

2.1. The Governance of CIP

Collaborative innovation has become a critical way to organize large-scale, multiparty innovation, as well as to enable the transformation and upgrading of industrial structures in China [3]. Briefly, it is voluntary cooperation between organizations involving sharing, exchanging, and co-developing technologies to pursue innovative products and meet business or public service needs. It is well-accepted that a project is a valuable means for organizations to collaborate on innovation [36]. The temporary organization involves two or more firms and other types of institutions with the potential to create more value than each firm can achieve on its own. Given that collaboration across different stakeholders plays a determinant role in the innovation process [37], prior research based on principal-agent theory has focused on the importance of participants’ attributes. On the one hand, participants with heterogeneous resource bases help pull complementary resources and capabilities for better project performance; on the other hand, this also leads to a game between different stakeholders. Considering collaborating among partners with different expertise is a practical approach to improving the reliability of the innovation process [38], a myriad of studies have examined the rules of enterprise collaboration with various types of institutions, such as universities [39,40], customers, and suppliers [41], so as to understand how to control the collaboration process and what are the major factors that drive the successes and failures of such innovation endeavors [42]. However, significant hurdles remain in the implementation process, including the lack of close cooperation among partners [43], unstable operation, and low success rate [44], and existing research does not provide an effective method to evaluate the collaboration process.

CIPs encounter not only management issues, but also governance issues. Muller proposed that management is a goal-oriented activity, and governance provides an environment or framework for executing management [8]. The governance of inter-organizational projects has been the focus of previous research. It’s been treated as a process [45], a system of controls [46], and a framework of decision [47] in several streams of literature. Yet, the unified object and basis of project governance have not been formed. Turner implies that the governance of a project involves a set of relationships between the project’s management, its sponsor (or executive board), its owner, and other stakeholders [48]. It provides the structure through which the objectives of the project are set, and the means of attaining those objectives and monitoring performance are determined [49]. In addition to addressing the challenge of aligning the interests of all stakeholders and limiting opportunistic behavior, some researchers have also emphasized the role of governance in enabling stakeholders to better coordinate their work. The coordination of work is especially relevant in complex transactions, such as innovation. These projects often include multiple stakeholders who have to accomplish a complex set of interrelated tasks in a limited timeframe. Although the role and value of project governance in the integration of stakeholders has been consensus, how to evaluate the effect of project governance in the CIP has not been studied yet.

It is an adaptative solution to solve this problem from the perspective of project stakeholder relationships. Stakeholder theory argues that a project organization is accountable to a broader range of stakeholders, and the structure of the organization should be aligned with this inclusive approach [50]. It emphasizes the importance of the individual attributes of stakeholders, all stakeholders have to be identified, and their potential impact on the project will be assessed [27]. The stakeholder theory in project governance focuses on providing space for stakeholders, managing their engagement in decision making, and addressing their concerns and demands [26]. Considering that project governance establishes and maintains regulatory relationships between project stakeholders, project governance can be indirectly evaluated by judging the rationality of stakeholder relationships. According to different stakeholder relationships, stakeholder governance is further divided into relational governance and contractual governance [32]. Some studies implicated that the collaboration between stakeholders in the project is relatively stable and contract governance plays a key role [18], while other studies believe that the uncertainty and dynamics of innovation lead the relational governance in the project should play a leading role [51]. Even in a project, some studies also propose that contractual governance has a stronger impact on project performance than relational governance in the early stage [24]. Therefore, it is not comprehensive to evaluate the effect of project governance only from one aspect; combining them is essential but has received little attention in project governance research.

Dynamic is an important feature of CIPs [52], with various stages making up the whole process: project planning, technological and product development, and marketing. Dynamic capability theory proposes that in order to maintain competitiveness, organizations need to constantly integrate their resources to adapt to the rapidly changing external environment [53], which is widely used in the research of enterprises’ participation in external market activities. However, in the project scenario, whether and how the organization adjusts its governance strategy according to the changes in project tasks and organizational structure has not received sufficient attention. The decision-making framework and management principles proposed by the governance theory focus on the unity of process and scheme [19], but also lack the consideration of dynamics. However, for CIPs, there are distinct differences between project tasks and objectives at different stages. Stakeholders will assume major governance responsibilities in different project stages due to their different professional strengths [54]. These structure and relationship changes influence how the governance mechanism is applied, just as the governance mechanism shapes how the stakeholder relationship is established [55,56]. This dynamic aspect of project governance has received little attention in project governance evaluation research and is of specific interest to CIPs as its unconsolidated stakeholder relationship brings additional complexities.

2.2. Network-Based Analysis

The interrelationships among CIP tasks are complex, and the coordination and integration of various stakeholders are required to produce satisfactory project revenue [57]. Therefore, an effective governance structure among project stakeholders is necessary for project success [58]. Rowley proposed that the relationship of project stakeholders should be changed from binary to network [59]. The project governance structure can be regarded as an organizational network. The stakeholders are the nodes of the network, and the relationships among stakeholders are the edge in the network [23]. Nodes can be joined by different kinds of links in various manners, forming unique network structures [60].

SNA is a network-based analysis method commonly used in the field of project management [61] since it provides a visual network expression measure to display the various relationships clearly. It has been integrated into an effective method widely used in project management and other research fields to solve problems related to stakeholders [62]. Mitchell defined SNA as “a specific set of linkages among a defined set of persons, with the additional property that the characteristics of these linkages may be used to interpret the social behavior of the persons involved” [63]. Previous studies show that stakeholders’ performance in governance activities has a great relationship with the position of stakeholders in the stakeholder network [64]. Network indicators such as density, centrality, structural holes, and factions of the SNA are often used to evaluate the project stakeholder network [23,33,65]. From a network perspective, this research highlighted the important value of stakeholders’ status in project governance.

In summary, the SNA-based project governance evaluation research focuses on the static evaluation of the project stakeholder network [64]. The network indicators in SNA are used to conduct a one-time overall evaluation of the project stakeholder network. The essential topic of these studies is how to establish a centralized coordination mechanism to facilitate project execution. Such research seems complicated but not rigorous. For example, the tasks vary in every stage of the project, leading to various relationship combinations among the stakeholders during the different stages [66]. This phenomenon means that the governance network structures will dynamically change with the project stage. As a result, the static evaluation method is no longer suitable and cannot accurately reflect the governance problems existing at different project stages. Therefore, this study conducts research on the dynamic evaluation of project governance and design network indicators that meet the needs of a vertical comparison of dynamic evaluation based on SNA.

3. Research Methodology

Previous studies on project stakeholder networks typically follow a classical framework containing network element identification, network construction, network analysis, and network evaluation [67]. This framework can effectively identify stakeholders and quantify their status on the project stakeholder network. Following this framework, this research will use an ITRI CIP to illustrate the process of solving the problem of project stakeholder network evaluation. The framework for project stakeholder network evaluation is presented in Figure 1.

3.1. Case Selection

The instrumental case study aims to reveal a social phenomenon by studying a specific object [68]. Case selection is the primordial step for case study, “representative” and “useful variation on the dimensions of theoretical” and “transparent observable” are standard case requirements for case study [69]. Using a single case study is appropriate when studying a representative case. The case selected for this research is a CIP of ITRI in Jiangsu, a province of China. This case is a typical Chinese CIP, and it is a landmark project in the region and plays a vital role in regional development. The participants of this project cover most types of innovative organizations in China, including universities, research institutes, enterprises, etc.; also, it covers the key stages of collaborative innovation. There are obvious differences in the tasks and goals of projects in various life cycle stages, and the rapid changes in the roles and relationships of project stakeholders in different stages have brought great challenges to project governance. The analysis of this case will provide a reference for the effective governance of other CIPs of the same type, and benefit for studying the reasons for the rapid development of innovative industries in China in recent years.

This project is initiated by ITRI, aiming to develop a standard technology through cooperation with enterprises and universities, promote local industrial transformation, and enhance industrial competitiveness. ITRI signed a cooperation agreement with a local company and two universities and jointly applied for government policy support. The government recognizes the value of the project, approves the project, and provides support for the project. The company successfully obtained bank loans through this project. According to their profession, the government, ITRI, enterprises, and universities are responsible for the corresponding tasks. In this case, the high project complexities made its project governance a challenging task.

3.2. Network Elements Identification

The snowballing method is a popular method to identify stakeholders in general social network research [70], such as interpersonal relationship networks. This study draws on the idea of snowballing to identify stakeholders from inside to outside layers and uses expert judgment to achieve node identification and relationship quantification. Considering the scope of the stakeholder, the stakeholder here refers to the project actor who is constrained by the project governance mechanism and whose decisions and actions can have an impact on the project governance. Independent interviews were conducted with a total of 15 interviewees, each interview with a theme related to the project stakeholders and a corresponding interview outline (as shown in

Table 1. Key information of interviews.

No.	Role and Number of Interviewees	Specific Interview Topic	Duration
1	Project manager (1)	Project stage division and key stakeholders in the project	2 h
2	ITRI (3)	Task, main responsibilities, stakeholders with collaborative relationships, and the type of relationship	1.5 h
3	Enterprise (4)	Task, main responsibilities, stakeholders with collaborative relationships, and the type of relationship	2 h
4	Basic research university (3)	Task, main responsibilities, stakeholders with collaborative relationships, and the type of relationship	1 h
5	Applied research college (3)	Task, main responsibilities, stakeholders with collaborative relationships, and the type of relationship	1 h
6	Government (1)	Task, main responsibilities, stakeholders with collaborative relationships, and the type of relationship	-

).

First, through a 2 h interview with a project manager, we identified the ultimate responsibility stakeholders of each stage: six nominated stakeholders were identified, including ITRI, enterprise, basic research universities, applied research college, government, and bank. Then, the nominated stakeholders were further identified according to their task relationship, and the layer-by-layer identification of stakeholders was realized from the inside out. At this stage, we conducted 4 semi-structured interviews with the nominated stakeholders, including ITRI, enterprise, basic research universities, and applied research college. Moreover, interviews with the government were conducted by email. The interviewees were all responsible for this project within their organization. Each interviewee mainly responded to two questions: “Who are your related stakeholders?” and “What kind of relationship exists between you?”. Most interviews lasted more than 1 h. Finally, the reference list was adjusted by experts to avoid the problem of range spread. The names and codes of stakeholders are shown in

Table 2. Stakeholder names and codes.

Stakeholders	No.	Stakeholders	No.	Stakeholders	No.
ITRI	A	Basic Researcher (PhD student)	C12	Project Manager	G
Enterprise	B	Basic Researcher (master’s student)	C13	Administrative Assistant	G1
R&D Technology Department	B1	Applied Research College	D	Theoretical Researcher	G2
Production Department	B2	Head of Patent Development (professor)	D1	Patent Development, Prototype Designer	G3
Equipment Management Department	B3	Patent Development Researcher (PhD student)	D11	Prototype Designer, Process Designer	G4
Purchasing Department	B4	Patent Development Researcher (PhD student)	D12	Advanced Manufacturing Laboratory	H
Market Department	B5	Prototype Designer, Process Expert (associate professor)	D2	Supplier I	I
Basic Research Universities	C	Prototype Designer, Process Designer (master’s student)	D21	Supplier II;	J
Head of Basic Research (professor)	C1	Government	E
Basic Researcher (Ph.D.)	C11	Bank	F

. Due to the consideration of dynamics, the correspondence between nodes and stages is very important. The nodes that do not belong to the stage are removed through the adjustment.

Taking into account the dynamic characteristics of project governance, the stakeholders identified above will appear in different project stages, so the project lifecycle needs to be divided into several stages. We interviewed the person in charge of ITRI and the project manager first to clarify the project’s main tasks and the stakeholders in charge of these main tasks. The entry and exit of important stakeholders and the change of the person in charge of the stage task will serve as a reference for the division of stages. Finally, the ITRI CIP stage is divided into 7 stages. The project stages, tasks, milestones, and correspondence between nodes and stages are shown in

Table 3. Division of project stages and correspondence between nodes and stages.

Stage	Task	Milestone	Stakeholders
Project initiation stage	The ITRI found opportunities for the project and initiated the project in conjunction with key stakeholders to obtain government support.	The government approves the project and decides to give policy support; cooperation agreements between stakeholders are signed	A, B, B1, C, C1, D, D1, D2, E, F, G
Orientation basic research stage	Organize and verify the theories related to the project technology to obtain basic data and related materials.	Complete “F Technology Theory Analysis Report”	A, B, C, C1, C11, C12, C13, E, G, G1, G2
Applied research stage	Implement theoretical research results, and apply for related patents for product development.	Submit a patent application	A, B, B1, D, D1, D11, D12, D2, E, G, G1, G2, G3
Prototype design stage	Determine the product’s function, and design the product structure and appearance.	Complete prototype interactive design	A, B, B1, B5, D, D2, D21, E, G, G1, G3, G4
Laboratory environment trial production stage	Design product implementation plans, develop prototype processes, and conduct trial production in a laboratory environment.	Complete prototype production and expert review	A, B, B1, B2, B4, D, D2, D21, E, G, G1, G3, G4, H, I, J
Enterprise environment trial production stage	According to the corporate environment, design and manufacture processes, supporting related materials and equipment, and conduct trial production of the corporate environment.	Complete small batch trial production	A, B, B1, B2, B3, B4, E, G, G1, G3, G4, I, J
Commercialization stage	The enterprise fully absorbs relevant technologies and pushes products to the market to solve problems such as production change between products and material coordination.	Achieve product sales	A, B, B1, B2, B3, B4, B5, E, F, G, G1, G4, I, J

.

3.3. Network Construction

Ucinet and Netdraw are used for network visualization based on the identified network elements. According to the survey results, fourteen adjacency matrixes of the seven project stages that represent the contractual and information relationships of stakeholders were determined. The network group is shown in

Table 4. Project relationship networks.

Stage	Contractual Network	Information Network
Project initiation stage
Orientation basic research stage
Applied research stage
Prototype design stage
Laboratory environment trial production stage
Enterprise environment trial production stage
Commercialization stage

. In the network graph, nodes represent stakeholders, and the edges are the relationships between them.

3.4. Network Analysis

As mentioned above, the project stakeholder networks are divided into contractual and information networks. Networks analysis includes whole network analysis and key node analysis. The whole network indicators reflect the system status of all stakeholders and represent the overall effect of project governance. The network status of key nodes has a crucial impact on project success, so the evaluation of key nodes focuses on analyzing whether they are in a proper network position to play their roles [61].

3.4.1. Network-Level Analysis

The contractual network represents the relationships of incentives and constraints among stakeholders. Their behavior can be effectively controlled when they obtain reasonable incentives or constraints, and the uncertainty of their behavior is low. The uncertainty of stakeholder behavior reflects the project governance risk [71], and a high project governance risk indicates a bad governance effect. The effect of controlling the uncertainty of stakeholder behavior can be expressed by the concentration of constraints that a node has in the network. In SNA, the structural hole indicator hierarchy represents how network constraints are concentrated on actors [72], and it can be formulated as Equation (1). The hierarchy of all nodes on the network reflects the governance risk of the entire project. Therefore, the average value of the hierarchy of nodes can be used to characterize the governance risk of the project and used as the network evaluation indicator of the contractual network. The formulation of the network hierarchy is shown in Equation (2),

$$H_i=\frac{{{\displaystyle \sum }}_j\left(\frac{C_{ij}}{C/N_i}\right)ln\left(\frac{C_{ij}}{C/N_i}\right)}{N_{i}ln\left(N_i\right)},$$

(1)

where $H_i$ represents the hierarchy of node $i$, $C_{ij}$ represents the constraint degree of point $j$ to node $i$, $N_i$ is the size of the individual network of node $i$, $C$ is the sum of the constraint of every node in the individual network of node $i$, and node $j$ belongs to the set of nodes connected to node $i$.

$$H=\frac{{{\displaystyle \sum }}_i{H}_i}{n}={{\displaystyle \sum }}_i\frac{{{\displaystyle \sum }}_j\left(\frac{C_{ij}}{C/N_i}\right)ln\left(\frac{C_{ij}}{C/N_i}\right)}{n{N}_{i}ln\left(N_i\right)},$$

(2)

where $H$ is the average hierarchy of all nodes in the network, and $n$ represents the number of nodes in the network.

The information network reflects the information interaction between nodes. The more connections between nodes on the network, the faster and more accurate the information transmission. In the SNA indicators, network density, which represents the ratio between the actual network connection and the maximum possible connection, is an important indicator that characterizes the network structure. When the network density is high in an information network, participants have more channels to interact with others, which ensures the effectiveness and stability of information transfer [73]. Therefore, network density can be used to measure the information interaction among stakeholders in the project, and it can be calculated as Equation (3).

$$ND=\frac{2l}{n\left(n-1\right)},$$

(3)

where $l$ represents the number of edges in the network.

3.4.2. Node Level Analysis

(1)

Key node indicators in the contractual relationship network

The nodes’ ability to integrate resources in the contractual network can be expressed by the effective network scale of the node [74]. The effective network scale of a participant is equal to the actor’s personal network size minus the redundancy of the network. The larger the effective network scale, the wider the node’s influence on other nodes. The calculation method is expressed as Equation (4),

$$N_i^{\prime }=N_i-\frac{2t}{N_i},$$

(4)

where $N_i^{\prime }$ represents the effective network scale of node i, $t$ represents the number of edges in the individual network of node i (excluding the number of edges that are directly connected to i), and $N_i$ is the size of node i’s individual network.

It is necessary to compare the network indicators of nodes at different stages when considering dynamic analysis. Since the network scales vary in different stages, this study uses the relatively effective network scale indicator to eliminate the difference in indicators caused by the network scale, which is formulated as Equation (5).

$$N_i^{″}=\frac{N_i-\frac{2t}{N_i}}{n},$$

(5)

The status of nodes on the contractual network is related to their ability to integrate network resources and their degrees of freedom when they take action. For example, when A wants to take action, the more the opinions of the surrounding nodes need to be considered, the greater the restriction and the lower the freedom degree of his decision. In the structural hole indicators, the constraint degree indicates the ability of the node to use structural holes in its personal network, and it is evaluated by the node’s dependence on another node, as shown in Equation (6). On this basis, by analyzing the sum of the constraint degrees of all nodes that are directly connected with node A, the constraint degree of node A can be obtained. The constraint degree shows the freedom of the stakeholder’s action, which can be formulated by Equation (7),

$$C_{ij}={\left(\frac{1}{N_i}\right)}^2{(1+{{\displaystyle \sum }}_q\frac{1}{N_q})}^2,$$

(6)

where $C_{ij}$ represents the constraint degree of node j on node i, and $N_q$ is the size of node q’s individual network. Node j belongs to the set of nodes connected with node i. Node q is an intermediate node connecting node i and node j.

$$C_i={{\displaystyle \sum }}_j{C}_{ij}={{\displaystyle \sum }}_j{\left(\frac{1}{N_i}\right)}^2{(1+{{\displaystyle \sum }}_q\frac{1}{N_q})}^2,$$

(7)

where $C_i$ represents the constraint degree of node i.

(2)

Key node indicators in the information relationship network

The role of stakeholders in the project information transfer can be evaluated from three aspects: the ability to collect and publish information, the efficiency of information transfer, and the ability to control the information transfer between other nodes [75]. Accordingly, the indicators for judging the information control ability of key nodes in the information network include three aspects: (1) the number of network nodes that are directly connected to the node, (2) the distance between the node and others, and (3) the probability of acting as an intermediary for other nodes’ information transfer.

In an information network, the more nodes in direct contact with a node, the stronger the node’s ability to publish and receive information. The degree of centrality indicates the number of nodes in the network that directly contact the node. Therefore, it can be used to represent the node’s ability to receive and publish information [76]. It can be formulated as Equation (8),

$$P_i={{\displaystyle \sum }}_j^n{a}_{ij}, i\ne j,$$

(8)

where $a_{ij}$ represents the relationship between i and j and $P_i$ represents the degree centrality of node i.

When a node acts as an intermediary for information transmission between other nodes, it has the ability to control its information transfer. Information always transfers through the shortest path, and betweenness centrality is the sum of the probabilities that a node is on the shortest path of two nodes of the network, which indicates the ability of the node to control the information transfer between other nodes [77]. It can be formulated as Equation (9),

$$B_i={{\displaystyle \sum }}_j^n{{\displaystyle \sum }}_k^n{b}_{jk}\left(i\right)={{\displaystyle \sum }}_j^n{{\displaystyle \sum }}_k^n\frac{g_{jk}\left(i\right)}{g_{jk}},j\ne k\ne i, j<k,$$

(9)

where $B_i$ represents the betweenness degree of node i, $b_{jk}\left(i\right)$ indicates the ability of node i to control the information transfer between node j and node k, $g_{jk}\left(i\right)$ represents the number of shortcuts between node j and node k that pass node i, and $g_{jk}$ represents the number of shortcuts between node j and node k.

When a node transmits information to other nodes on the network, the fewer intermediate nodes it needs to pass through, the more it can avoid information delay and distortion; thus, it has higher information transmission efficiency. The closeness degree represents the sum of the shortest distance between the node and others in the network [76]. Therefore, the closeness degree can be used to indicate the efficiency of the node in transferring information. It can be formulated as Equation (10),

$$Z_i={{\displaystyle \sum }}_j{d}_{ij},$$

(10)

where $Z_i$ represents the closeness degree of node i and $d_{ij}$ represents the shortest distance between node i and node j.

Considering that the positions of nodes in different networks need to be compared in dynamic analysis, this study uses the relative comparison method of ranking to redesign the above indicators. Select the optimal and worst values of the indicator: for the degree centrality and betweenness centrality, the larger, the better, while for the closeness degree, the smaller, the better. The calculation of the relative comparison method is shown in Equation (11),

$$L_i=\frac{V_i-Worst}{Best-Worst},$$

(11)

where $L_i$ is the relative index, $V_i$ is the actual value of node i, $Best$ represents the optimal value in the network, and $Worst$ represents the worst value in the network.

The calculations of relative degree centrality, betweenness degree, and closeness degree are shown as Equations (12)–(14),

$$P_i^{\prime }=\frac{P_i-Min\left(P\right)}{Max\left(P\right)-Min\left(P\right)},$$

(12)

$$B_i^{\prime }=\frac{B_i-Min\left(B\right)}{Max\left(B\right)-Min\left(B\right)},$$

(13)

$$Z_i^{\prime }=\frac{Z_i-Max\left(Z\right)}{Min\left(Z\right)-Max\left(Z\right)}.$$

(14)

The value of the above calculation results ranges from 0 to 1. The higher the value is, the better the network status of the node.

The above three indicators represent the three measurement dimensions of the node status on the information network. To obtain the comprehensive index that expresses the overall status of the node, three indicators are weighted and synthesized. The comprehensive index can be formulated by Equation (15),

$$S=S_1{P}_i^{\prime }+S_2{B}_i^{\prime }+S_3{Z}_i^{\prime },$$

(15)

where S₁, S₂, and S₃ represent the weights of relative degree centrality, relative betweenness degree, and relative closeness degree, respectively.

4. Results

Ucinet 6.0 and MATLAB R2020a are used to calculate the network indicators to obtain the dynamic governance evaluation results of the ITRI CIP. Ucinet 6.0 is the most frequently used comprehensive analysis program for processing social network data. It will be helpful to solve the natural indicators of SNA, while for some adjusted indicators, it should be calculated by MATLAB R2020a, which is a more flexible data analysis tool.

4.1. Result of Network-Level Analysis

(1)

Contractual network

Figure 2 shows the trend of the average hierarchy of project stakeholders as the project stage changes. The average hierarchy of the contractual network is lower in the initial stage of the project and gradually increases as the project progresses. This feature is in line with the rule that the risk is high in the early stage and progressively decreases as the process goes on [78]. Compared with traditional construction projects, this feature is more obvious in CIPs because of their more unpredictable production, more uncertain progress, and more uncontrollable stakeholder relationships.

(2)

Information Network

Figure 3 shows the trend of the network density as the project stage changes. It conveys two messages about the project information situation. First, the network density is relatively low, less than 0.4 in all project stages, indicating that the actual connections between stakeholders in the network account for less than 40% of all possible connections. In innovative projects, it is of great significance to share needs and experiences through information transmission, and insufficient information is not conducive to innovation. Second, the network density is unstable throughout the whole project lifecycle. It is close to 0.4 in the prototype design stage but lower than 0.3 in the commercialization stage. The changing stages of tasks and the stakeholders who undertake the main work are the main reasons for this problem.

4.2. Result of Node Level Analysis

The ITRI CIP has public interest attributes, serving the development of local industries rather than just an enterprise. The government needs to provide a reliable implementation environment for the project and monitor the progress of the project. The enterprise that absorbs scientific and technological production is the “customer” of the project. ITRI assumes ultimate responsibility for the final results of the project. As the governance subject of the entire project, it should be responsible for coordinating the responsibilities and rights of all stakeholders. The project manager is responsible for project management and plays a vital role in the operation process of the project. Therefore, the key stakeholders in the ITRI CIP include governments, enterprises, ITRI, and project managers. Figure 4 and Figure 5 present the results of the network indicators of each key stakeholder change with the project stage.

(1)

Government node

As shown in Figure 4, in the contractual network, the government’s effective network size gradually decreases, and the degree of restriction gradually increases with the development of the project, indicating that the government’s ability to control network resources is getting weaker, and the resistance to its actions is increasing. Therefore, the government’s position in the contractual network gradually declined with the project development. Meanwhile, the same situation also occurs in the information network. Specifically, the government has a higher degree of centrality and betweenness centrality in the project initiation stage and then worsens in the following stages, and each indicator becomes the worst from the third stage. However, the closeness degree increases after the fourth stage, and the comprehensive information network status of the government is still the worst.

The results show that the government has lost the ability to restrict other stakeholders because of its bad network status, and it can only issue orders through its inherent power, not effective management methods, which is not an effective and sustainable governance mechanism [79]. Meanwhile, it can be seen from Figure 5 that the government’s various information network indicators are very low, which means that the government’s access to network information mainly depends on a few nodes, and its ability to control the transfer of information is extremely poor, which can be easily controlled. It could present an enormous risk for the government caused by information asymmetry [80]. In summary, the government has neither an effective way to restrict the behavior of stakeholders nor sufficient information to supervise the project process, leading to insufficient project participation.

(2)

Enterprise node

The network status of enterprises shows great volatility in Figure 4 and Figure 5, and it shows the same trend in the contractual network and the information network. Figure 4 shows that the enterprise has a larger network scale and a lower restriction degree at the first and last few stages in the contractual network, indicating its network status is high. In the middle stages, the situation gets worse. Figure 5d shows that the comprehensive network status index of enterprises presents a situation of sharply decreasing and then gradually increasing in the information network, and other information network indexes show the same trend in other figures of Figure 5.

As the project sponsor, the enterprise needs to establish contractual relationships with many stakeholders and obtain enough information about policy and academic support in the early stages of the project. By the end of the project, the enterprise takes responsibility for market development, and it will publish all the project information to ensure that production can be accepted by the market. Therefore, the enterprise has a higher level of participation in the early and last few project stages. In the R&D stage, in the middle of the entire lifecycle, the company is not responsible for specific scientific work, so the degree of participation is relatively low. Due to the superimposed influence of the two networks, the status of the enterprise in project governance is not optimistic.

(3)

ITRI node and project manager node

The project manager is appointed by ITRI and undertakes its management function in the project. Therefore, the parties can be seen as independent but closely related. Putting them together for analysis will be convenient for comparing their network status and will contribute to finding their governance risks.

ITRI’s network scale gradually decreases in the contractual network, and the constraint degree slowly increases, indicating that ITRI’s status in the contractual network gradually decreases. It shows the opposite trend for the network status of the project manager. Moreover, the network status of the project manager is gradually higher than that of ITRI. For the information network, the network status of ITRI and the project manager does not show obvious advantages compared with other stakeholders. In terms of dynamic trends, the ITRI network status in the information network is gradually decreasing, and the project manager’s network status is more volatile.

The indicators of the contractual network show that the project manager has a comparative advantage compared with ITRI, which will cause entrusted agency risk [81]. In the absence of effective project contractual structure constraints, ITRI and project managers can control the information flow and create information asymmetry, thereby affecting the behavior of all stakeholders involved. Some defects in the information network structure affect the ability of the project governance subject to control project information. First, the ITRI and project manager’s network status should be significantly higher than those of others, as they are responsible for project process control and the final results, but this feature is not obvious in most stages in Figure 5d. Second, it can be seen from Figure 5a that the degree centrality of each stakeholder is relatively low; there will be many “information black boxes” for project managers and ITRI because only a few nodes are connected to them, which reduces their ability to control the overall project information. Furthermore, the network indicators of the project manager are all relatively high in all figures of Figure 5, which means project managers have the advantage of asymmetric information in most information networks, leading to the excessive power of project managers.

5. Discussion

Dynamic evaluation of project governance is conducive to identifying project governance risks and improving project governance efficiency. This study uses the network-based method to analyze the dynamic process of the overall project and key stakeholders’ governance structure with the project lifecycle in a typical ITRI CIP case. On the one hand, it will provide a framework reference for the quantitative evaluation of project governance for other collaborative innovation projects, and the insights proposed will help call for quantitative methods to evaluate the effectiveness of project governance; On the other hand, the results of this research also provide typical case support for the application of governance theory in ITRI CIPs.

Traditionally, project governance evaluation relies on operational, tangible measures centered on time, cost, and financial return to determine the success of a project [82], or perceives success at a long-term, strategic level by validating indirect, social benefits of project outcomes [83]. However, Lappi’s study indicates that even though projects themselves favor measuring and visualizing success and outcomes through short-term metrics and deliverables [84], the innovation project’s success also be of interest for stakeholders’ satisfaction, including the customer, sponsor, or owner organization, and should therefore be evaluated with means closer to those of project governance. As mentioned in the case project, the government’s shortcomings in contractual governance seriously affect its process of monitoring and obtaining effective feedback on the project. However, in reality, the project does not face any cost or schedule risks at this time. This study thus provides an evaluation approach for CIP governance that does not use traditional indicators, and indicates that ensuring the stakeholders maintain a balance between relational governance and contractual governance in the project governance structure is the key to ensuring the effectiveness of CIP governance.

Under the project governance research based on principal-agent theory, the most significant risk for enterprises is the divorce between R&D and the market [85]. Information asymmetry between different stakeholders is the resource of these risks [86]. To address this problem, existing researchers use the game theory method to find a fair equilibrium solution for all parties [87]. The game between stakeholders means that one party gains at the expense of the other party, which is completely contrary to the basic characteristics of the co-creation of value in CIPs [88]. The results of this study provide evidence for the statement that the relational governance structure in a project is dynamically changing, with different stakeholders occupying the dominant position in information governance at different stages. The dynamic characteristic leads to the loss of benefits obtained by the stakeholders through creating information asymmetry in other stages. For example, if a scientific research institute cannot effectively communicate with the enterprise when it has an information advantage brought by its professionalism during the technology research and development stage, the enterprise will use its information advantage position in the marketization stage to counter it. It is not feasible to obtain benefits through manufacturing information asymmetry in ITRI CIPs. Therefore, there are limitations in using principal-agent theory to carry out research in CIPs; the roles of stakeholders are more like partners than principals and agents. Significant changes were found in terms of four key stakeholders’ network positions, both in the contractual and information networks. The analysis result of different scenarios also shows significant differences in the governance capabilities and roles of various stakeholders under different governance structures. These findings support the theoretical analysis of the dynamic characteristic of project governance [7,21]. In a rapidly changing project environment, organizations need to constantly adjust their participation strategies according to changes in external conditions [89]. Regular joint work activities provide an opportunity to exercise collaborative behaviors, and develop contractual and relational relationships for all team members without limitation to the management level. At this level, the role of the governor as a ‘coordinator’ of collaboration from the top level to the project is very important [90]. A project coordinator should be appointed to serve as the single point of accountability for the project governance by their network position in order to prevent ‘blame-shifting’ by the parties involved. The analysis of the project governance structure in this research method provides a scheme for finding the coordinator of project governance. Stakeholders in a better network position at different stages should be accountable for project governance. The standard approach to project governance is to attempt to address all of the relationships at once. While such a costly governance mechanism may not always be justified, multi-level governance will be more scientific through the evaluation of the stakeholder networks.

For these key stakeholders, their position in the network is essential for project success. The paper also illustrates why stakeholders perform poorly in good contractual or information relationships. Some stakeholders have obtained the dominant position in some networks, naturally or through the system, such as the government. However, the results of the evaluation of the government’s project governance status show that due to its poor governance structure and weak governance network status, it has not shown its power in project governance. Similarly, enterprises with good contractual governance status in the technology research and development stage have poor participation in projects due to a lack of relational governance. This finding is consistent with the conclusion that contractual governance and relational governance complement each other in static project governance research [32]. The results of this study indicate that relational governance and contractual governance are not substitutional, even in an innovation environment with rapidly changing stakeholder relationships. The management of both contractual elements [91,92] and relational elements [93], as well as the complex interplay between these elements [32,91], become important concerns for the project governance system. Replacing formal contracts with informal organizational relationships will pose risks to project governance.

This research has taken a significantly different perspective on studying project governance. The iterative process and dynamic stakeholder participation bring great complexity to CIP governance. Targeted project governance provides stakeholders with a suitable environment by modifying the contractual relationship and information relationship. This research provides a dynamic network analysis model to evaluate key stakeholders’ network status and identify the shortage of their network structure, underlying the scalpel-style and systematic improvement of project governance.

6. Conclusions and Future Research

Project governance is essential to managing innovation project implementation and delivering innovative products in multiorganization collaboration project contexts. This paper extends the understanding of the dynamic project governance mechanism in the context of CIPs by explicating the change through the network analysis, and the dynamic evaluation of project governance will guarantee the innovation process and the commercialization of innovation products. Scientific project governance will effectively enhance the sustainable innovation ability of enterprises.

The research suggests that project governance is a unity of change and unchanging. Stakeholder relationships, especially the contractual relationship and information relationship, constitute the complex organizational environment of project governance in the CIP, which is the unchanging part of project governance. The poor network status in the contractual network or the information network is the source of project governance risks. Meanwhile, the relationships between stakeholders will also change along with the project lifecycle, resulting in a change in the project governance environment. The network model proposed in this study can effectively measure the governance environment of the main stakeholders in a project, and then determine whether the existing stakeholder relationships can meet their effective governance needs. Therefore, during the project lifecycle, project governance rules need to be constantly adjusted to meet the needs of stakeholders, which is the changing part of project governance. The governance effect of CIP is determined by the changeable stakeholder relationship structure, not just the stakeholder’s power of the general understanding, and the key stakeholders can achieve the optimal project governance effect through the adjustment of the relationship. The findings of this research will shed light on the exploration of the project governance object and spur further conceptual and empirical research on project governance from the perspective of stakeholder relationships.

In practice, for other CIPs of the same type, the findings will help project stakeholders formulate appropriate participation strategies directly. The government should strengthen the control of project information and avoid forcibly interfering with the project based on its administrative status. Enterprises need to participate more during the project implementation stages and reduce the rework of the project through sufficient demand expression. Project managers that ITRI arranges need both technical and management capabilities. When a technical expert is arranged to be a project manager, there should be equipped with a management team to reduce speculation in the project. Considering the unique characteristics of the project, other CIPs can dynamically evaluate the governance level and identify project governance risks by constructing their own governance evaluation network model based on the framework proposed in this study. The case study process provides a template for them to apply the framework. Furthermore, this study also can offer some reference significance for projects in traditional industries. Although the network structure of these projects is more stable, this study can provide strategies for designing the governance network at the beginning of projects to reduce project governance risks.

Despite the research contributions and the practical implication suggestions mentioned above, our study was carried out with a few limitations. First, in terms of data acquisition, we had limited access to project meetings and documentation. There may be a problem that data cannot be cross-verified through different data resources in the process of data collection. The multiple independent interviews ensure the accuracy of the initial data and, thus, the validity of the results as applied in this context. Second, this study mainly focuses on the proposal and application of evaluation methods, while in case selection, only a typical project implemented in China is analyzed. Therefore, the conclusions obtained lack universality, and expanded future research using this method in different locations, project types, and contexts will help to generalize our findings further. Third, concerning the method design, the dynamic evaluation method based on the project stage in this research belongs to “discrete dynamics”. Future research will realize “continuity dynamics” through the integration of complex networks and improve the accuracy of the evaluation.