This article was downloaded by: [University of Arizona]

On: 21 January 2013, At: 19:54

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Innovation and Development

Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/riad20

Dynamics of technological innovation

network and its impact on
organisational learning and innovation
performance
a
Ju Liu
a
TaSTI (Research Unit for Science, Technology, and Innovation),
University of Tampere, Tampere, Finland
Version of record first published: 19 Apr 2012.

To cite this article: Ju Liu (2012): Dynamics of technological innovation network and its impact on
organisational learning and innovation performance, Innovation and Development, 2:1, 159-174

To link to this article: http://dx.doi.org/10.1080/2157930X.2012.663584

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-

conditions

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation
that the contents will be complete or accurate or up to date. The accuracy of any
instructions, formulae, and drug doses should be independently verified with primary
sources. The publisher shall not be liable for any loss, actions, claims, proceedings,
demand, or costs or damages whatsoever or howsoever caused arising directly or
indirectly in connection with or arising out of the use of this material.
 Innovation and Development
Vol. 2, No. 1, April 2012, 159 –174

Dynamics of technological innovation network and its impact on

organisational learning and innovation performance
Ju Liu∗

TaSTI (Research Unit for Science, Technology, and Innovation), University of Tampere, Tampere,
Finland

This paper examines the 10-year-long dynamics of a technological innovation network (TIN)
Downloaded by [University of Arizona] at 19:54 21 January 2013

of a textile company in China. It opens the black box of the firm and provides a whole picture of
the firm’s internal and external innovation network. It addresses two research questions: how
did the structure of the company’s TIN evolve and what was the impact on the organisational
learning and innovation performance of the company. The case study reveals that the
connectedness of the TIN and the centrality of an actor in the TIN are positively related to
more intensive, more frequent interactive learning, and also to better technological
innovation performance. Nevertheless the relation is not linear. It finds that peripheral actors
may also be important to the organisational learning and innovation performance, with their
power not from their network position but from other capabilities they have.
Keywords: technological innovation network; organisational learning; network structure;
network dynamics; innovation performance; China

1. Introduction
Innovation network has been firmly in the research agenda of innovation studies. Understanding
the role of knowledge and organisational learning in fostering or inhibiting innovation has also
become crucially important (Lam, 2005). Research has been done based on various observations
and in the light of diverse theories. But there are still two major gaps in innovation network
research. First, there is an improper separation of research on intra-firm innovation networks
and inter-organisational innovation networks. The former usually takes individuals as actors
and focuses on how individuals in the organisation link up with each other when innovating.
The latter usually takes the organisations as actors, focusing on the relationships among organis-
ations and leaves each of the organisations as a black box without touching any relationships
within the organisation. The separation of research on intra- and inter-organisational innovation
network may lead to insufficient understanding of innovation networking as a process with both
internal and external aspects (Liu and Baskaran, 2009; Liu, 2011). Second, there is a lack of longi-
tudinal research on innovation networks on account of the difficulty in collecting network rela-
tional data over time.

1) This paper attempts to analyse the dynamics of the technological innovation network
(TIN) of an innovative textile company in Southwest China from 1997 to 2007. It
opens the black box of the company by taking the different functional departments as

∗
Email: ju.liu@uta.fi

ISSN 2157-930X print/ISSN 2157-9318 online

# 2012 Taylor & Francis
http://dx.doi.org/10.1080/2157930X.2012.663584
http://www.tandfonline.com
 160 J. Liu

actors of the intra-firm innovation network on the one hand, and introduces the outside
organisations into the network analysis on the other. Thus, the innovation network in
this paper is a combination of intra-firm innovation network consisting of functional
departments and groups inside the company, and the inter-organisational innovation
network consisting of outside organisations. I used social network analysis to map and
measure the relationships between different departments and organisations inside and
outside the company.

The main research questions are:

(1) How did the TIN of the case company evolve over the different stages of technological
innovation?
(2) What was the impact of the structure of TIN on organisational learning for technological
Downloaded by [University of Arizona] at 19:54 21 January 2013

innovation and on the performance of innovation in these different stages?

2. Taxonomy of key concepts

In this paper, technological innovation includes product innovation, consisting of new or better
material goods as well as new intangible services, and process innovation which involves new
ways of producing goods and services (Edquist, 2005).
Organisational learning in this paper is about competence building of the firm. I consider
organisational learning as a process which takes place through activities performed by individ-
uals, groups and organisations as they gather, interpret and store information, plan and perform
new actions, and implement changes. It is considered as a conscious attempt on the part of organ-
isations to retain and improve competitiveness, productivity and innovativeness in uncertain tech-
nological and specific social circumstances.
The TIN in this paper is combined by intra-firm network and inter-organisational network. In
the intra-firm network, the nodes are groups of people who serve in different functional depart-
ments such as marketing, finance, R&D and human resources (HR) within the firm. In the
inter-organisational network of this research, the nodes are formal structures that are consciously
created and have an explicit purpose (Edquist and Johnson 1997). They are organisational actors,
such as educational and scientific research institutes, non-governmental investment institutions,
customers, suppliers, competitors and so on. Ties are the relationships between the nodes. Rela-
tional ties between nodes are channels for organisational learning, such as improving actions and
building competence through better knowledge and understanding. In this research the ties are
weighted and undirected. Lundvall (2007) argued that the impact of innovation on economic per-
formance will typically depend upon changes in ‘people’ or ‘orgware’ which refers to how people
relate to each other within organisational borders, and ‘socware’ which refers to how people relate
to each other across organisational borders. The ‘orgware’ and ‘socware’ are interpreted as the
structural attributes of intra-firm network and inter-organisational network in this paper.
Network dynamics in this paper refers to the evolving or changing of the structure of the
network, such as breaking or making of ties. I attempt to take snapshots of the TIN in different
stages during its evolutionary process in the past 10 years. I believe that an existing network struc-
ture can only be properly understood and explained by considering the process that led to it.

3. Literature review
Innovation network research by a social network approach is based on an assumption of the
importance of relationships among actors who interact with each other. The social network
 Innovation and Development 161

perspective includes theories, models and applications that are described and studied in terms of
relational concepts or processes. There is now a consensus that the central principles underlying
the network approach are a matter of common sense. These are that network structure provides
opportunities for or constraints on individual action, and that the power of individual actors
lies in its relative position in the network rather than the characteristics of the actor itself (Wasser-
man and Faust, 1994).
The existing literature on innovation and innovation network is inconsistent with the structure
of innovation network and its impact on organisational learning for innovation and innovation
performance.
Concerning the relation between network density and innovation, Sabatier and Jenkins-Smith
(1993) argued that dense structures tend to reinforce inertia and can be hostile to innovation;
while McCarthy (2008) arrived at the opposite conclusion that a dense network helped spread insti-
tutional norms of behaviour and it also facilitated the transfer of information which is important to
Downloaded by [University of Arizona] at 19:54 21 January 2013

innovation. The latter studied the process of network evolution and transformation in health insur-
ance in Ireland and found that the network was characterised by high centrality, high density and
weak ties, and cost efficiency was a key outcome. The positive relation between density of
network and learning is also justified by the research by Tsai (2006) on the impact of website struc-
ture on customer knowledge flow and innovation in the internet marketing in Taiwan.
About the relation between network centrality and innovation, Herminia (1993) argued that
network centrality is the strongest determinant of individual involvement in innovation efforts
for administrative innovation but not for technological innovation. Ebadi and Utterback (1984)
studied 117 Sea Grant research projects and found that on the one hand, frequency of communi-
cation on the individual level affects technological innovation and so do centrality and diversity;
on the other hand, network cohesiveness, centrality and diversity of communications on the organ-
isational level were positively related to technological innovation. In particular, the high correlation
between network cohesiveness and technological innovation implies that cohesiveness is an impor-
tant factor in the success of technological innovation. Sandstrom and Carlsson (2008) conducted a
comparative case study of four networks within the higher education sector and proposed that an
efficient and innovative policy network consists of a heterogeneous set of actors that are centrally
and densely integrated. Furthermore, the level of centralised integration promotes the function of
prioritising. Differing from the research aforementioned, Bell (2005) modelled the innovativeness
of Canadian mutual fund companies as a function of their geographic location and of centrality in
networks of managerial ties (informal ties between managers) and institutional ties (formal ties
between their firms). He found that location in the industry cluster as well as centrality in the man-
agerial tie network, which refers to informal and personal connections between managers of the
firms, enhances firm innovation while centrality in the institutional tie network, which refers to
formal connections between the firms, does not.
The inconsistency of the findings may come from the network characteristics, such as the type
of actors (individual, group, organisation), type of ties (formal, informal, both formal and infor-
mal). It may come from the measurement of innovation performance (patent, new products, new
processes, etc.), as well as the measurement of organisational learning (process measurement,
such as frequency of interaction, intensity of interaction, etc.; outcome measurement, such as
increased productivity). It may also come from the different social, political, cultural, economic
and industrial context that the networks are embedded in. In this sense, the inconsistency of the
findings actually provides an opportunity to summarise different models of interplay between
network structure, organisational learning, and innovation performance in different contexts so
as to better understand the drivers and hindrances of using the network to promote innovation.
This paper adopts a social network approach to study the dynamic history of a firm’s techno-
logical innovation network. It tests the basic assumption of a social network approach and tries to
 162 J. Liu

find out the pattern of network structure and its impact on organisational learning and innovation
performance in the context of an underdeveloped region in China.

4. Research methodology
4.1 Case study
When studying the complex dynamics of a TIN, there is no single flawless method. Methodologi-
cal individualism cannot be applied to processes where knowledge and learning are central
(Arrow, 1994). Therefore, phenomenology methodology,the case study method and social
network analysis were adopted in this paper. I tried to use multiple tools to see the dynamics
of the TIN clearly and deeply.
The case study method was used to understand the major issues surrounding the organis-
ational learning and technological innovation in theGrace Corporation – a Chinese firm selected
for the study. In organisational research, the case study method is one of the frequently adopted
Downloaded by [University of Arizona] at 19:54 21 January 2013

research methods and the appropriateness of the method is well documented (Eisenhardt, 1989;
Pettigrew, 1990). Different sources of evidence are utilised, including questionnaires, interviews,
direct observation, archives and statistics. In the data collection phase, I used a semi-structured
questionnaire, which was to collect the relational data of the TIN of the Grace Corporation.
Overall, six visits were made during the period December 2004 to January 2008 and 25 interviews
were conducted. The interviewees included the President and Chairman of the Board, the
Vice-general Manager, the directors of the middle-level management team from seven different
sections – the Science and Technology Department, the Domestic Marketing Department, the
International Marketing Department, the Strategic Planning Department, the Intellectual Property
Right (IPR) Office, the HR Department and the Real Estate Company, and individual engineers
and workers. Typically each interview lasted for about one to two hours at the old location of
Grace as well as the new site. The interview phases lasted for about eight non-consecutive
weeks. All the interviews were well recorded but not taped since the informants were reluctant
to share their views in front of a tape recorder. An agreement was reached with Grace over
business confidentiality. Informal discussions with the members of the organisation provided
me with a better understanding of the important themes underlying the firm’s practice of organ-
isational learning and technological innovation. In the data clarification and complementation
phase, I contacted the Grace’s managers via email correspondence and telephone discussions
for further information and data, and to clarify unclear points from the previous interviews.

4.2 Social network analysis

The social network analysis (SNA) was adopted as an analytical tool in this research in consider-
ation of the interaction-dependence of technological innovation and social attribute of organis-
ational learning. Sociologists in general would agree that power is a fundamental property of
social structures and is inherently relational. Hanneman and Riddle (2005) argued that power
in social networks may be viewed either as a macro property which describes the entire population
or as a micro property which describes relations between actors. Burt (1992) and Wasserman and
Faust (1994) also defined two approaches in network analysis with one approach focusing on the
whole of the network itself and another on particular actors in a network.
According to the dichotomy found in network analysis, I used two groups of important par-
ameters in this paper: the first is to describe the power of the TIN structure as a whole, and the
second is to identify and understand the position which reflects the opportunities and constraints
of a specific actor in the network. These two groups of parameters are network properties and
node properties. The network properties include density, centralisation, efficiency and diversity,
 Innovation and Development 163

while node properties refer to node centrality, including Freeman (1979)’s degree and Bonacich
(1972)’s Bonacich power, closeness and betweenness. These measures have come to dominate
empirical usage to identify the most important actors within the network (Borgatti, 2005).
Each measure has various relative advantages and disadvantages concerning its use.
The software of UCINET was used to map the Grace’s TIN and detect the properties of the
network in order to understand its pattern and dynamics.

5. Case description
This paper is based on observations of innovation activities from 1997 to 2007 in a textile
company named the Grace Corporation.
The Grace Corporation is located in Yibin, a city in the province of Sichuan in Southwest China.
It grew out of a small chemical fibre factory founded in 1984. Until 1997 it was a small factory on
the edge of bankruptcy. 1997 was a milestone in the history of the Grace marked by the change of
Downloaded by [University of Arizona] at 19:54 21 January 2013

senior management and the invention of a revolutionary technology called ‘2S’ which can double
the production at very low cost. Since then, the Grace experienced a high growth at an average
annual rate of 35%. Now it is one of the world’s largest manufacturers of viscose filament yarn,
rayon embroidery thread and hand knitted garments with total assets of 3.9 billion RMB and
12,000 employees. The world market share of its products reached 29% in 2007.1
There are many outstanding occurrences related to technological innovation in this company
which make a break from the common trend in China. First, in this company the annual R&D
investment as a percentage of sales increased from 3 to 9% in the past seven years, a figure far
higher than the average level of 0.2 to 0.5% in China’s textile companies. Second, this
company benefits significantly and continuously from a technological innovation – ‘2S’,
which is very easy to be imitated. This unique technological innovation has strongly supported
the high growth of this company in the past 10 years. Third, they have 114 patents compared
with the average of eight patents for the import and export enterprises in the province of Sichuan.2

6. Case analysis
To study the dynamics of the TIN of the Grace, I first divided the ten-year-long history of the
Grace’s technological innovation into three different stages. Second, I took snapshots of the
TIN’s socialgram in each of these stages to visualise the evolution of the TIN’s structure.
Third, I detected the structural properties and analysed the dynamics of the TIN over these
three stages to see how the dynamics of the TIN’s structure influenced the organisational learning
and innovation performance of the company.

6.1 Three stages of the Grace’s technological innovation history

I divided the Grace’s technological innovation history according to the performance of its tech-
nological innovation including the performance of product innovation and process innovation.
The percentage of the sale of new products to the total sale is used to measure the product inno-
vation performance. The number of successful process innovation projects is used to measure the
process innovation performance. The new products refer to the products which were developed
and produced by new technological elements and new design, or those which have obvious
improvement in structure, material and process compared to the existing ones, or in general
those which have obvious improved performance or expanded functions. Taking into consider-
ation the nature of chemical fibre products, I set a period of three years of expiry for the new pro-
ducts from the year they were first introduced into market. The successful process innovation
projects refer to those whose outcomes have been put into practice. In order to compare these
 164 J. Liu

two different categories of performance and also to have a comprehensive measurement of tech-
nological innovation performance as a whole, I adopted the effectiveness coefficient method to
change the performances with different measurement into index to make it comparable and super-
imposeable. The model of these indexes is as follows:

dj (x) = (f j (x) − f jmin )/(f jmax − f jmin ) × 40 + 60 [ [60, 100], j = 1, 2, . . . , p

p

D=( dj (x))1/p
j=1
dj(x) - single index, i.e. index of product/process innovation performance
D - multi objective, i.e. index of technological innovation performance
Downloaded by [University of Arizona] at 19:54 21 January 2013

According to the technological innovation performance (see Figure 1) as well as my discus-

sion with the Grace’s managers and the President, the history of the Grace’s technological inno-
vation is divided into three stages: 1997– 1999, 2000– 2004 and 2005 – 2007.
The first stage is from 1997 to 1999 and is named the ‘elementary stage of technological inno-
vation’. It witnessed three historical events. The first was the change of top management. The
current Chair and President Feng Tao was assigned by the local government. The second was
the invention of the historically important technology ‘2S’, which is an innovation process that
makes it possible to double the production at a very low cost. The third was the massive recruit-
ment of 600 new employees. This directly led to a blood-transfusion-like organisational change.
Most of the current mid-level managers are from this group of people. This stage was a turning
point, changing an old factory from the verge of bankruptcy to an innovative enterprise. However,
the innovation performance was not high or significant yet in this stage.
The second stage is from 2000 to 2004 that is called as the ‘growing stage of technological
innovation’. It is also called the ‘first spring of technological innovation’ by the Grace. The
main characteristics of this stage are the invention of a large number of influential and profitable
technologies, and the rapid growth fuelled by the prosperous technological innovations. There
were two symbolic events: a boom of patenting including the key technology ‘2S’, the establish-
ment of the Science and Technology Department and the IPR Office (both of them report directly
to the President), and the launching and implementing of a policy which significantly reward the
actors and activities of technological innovation. I observed a drop in the percentage of new
product sale to total sale during this period. However, this was not because of a slow down in
technological innovation but because of a market reason. During the 2003 – 2004 period, 2S

Figure 1. Technological innovation performance of Grace from 1997 to 2007

 Innovation and Development 165

technology was illegally imitated by all the main competitors of the Grace and its sale of new
products was significantly influenced by this infringement.
The third stage is from 2005 to 2007. I named this stage the ‘plateau stage of technological
innovation’ of the Grace. In this stage, the performance of product and process innovation
further increased but then dropped slightly and stayed at a stable level. The main development
during the later period of this stage included technological-innovation-fatigue of the employees
and lack of technological talents as a result of their movement from technological positions to
managerial positions.

6.2 The evolution of the network structure of the Grace’s TIN and its impact on
organisational learning and innovation performance
I identified eight internal actors, which are different functional departments and groups inside the
company, and ten external actors, which are different organisations outside the company (see
Downloaded by [University of Arizona] at 19:54 21 January 2013

Table 1). To examine the connections between inside actors, I asked the question: ‘Do the differ-
ent functional departments and groups contact each other when involved in technological inno-
vation activities?’ To measure the importance of the connections to the company’s
technological innovation, I asked the respondents to score the importance of connection which
is represented by the five-level Likert item. I also carried out the same examination and measure-
ment to the connections between the inside actors and outside actors, as well as the connections
between the outside actors. The types of the connections are divided into three categories: (i)
Open information sharing: accessing knowledge without the need to pay; (ii) Goods acquiring:
purchasing machine, patents, license, service, software, expertise, etc.; (iii) Innovation
cooperation: actively participating in joint innovation projects with other organisations.
We can see the evolution of the Grace’s TIN in three different stages from the snapshots
shown in Figures 2, 3 and 4. Analysing network properties of the TINs over the three stages
(Table 2) together with the innovation activities and performance (Figure 1) of each stage pro-
vided some interesting findings. These are as follows:

The connectedness and compactness of the Grace’s TIN kept increasing over the three stages of
technological innovation while organisational learning and innovation performance shows an S
curve at the mean time.
The density of the TIN increased significantly from the elementary stage to the growing stage and
then increased slightly in the plateau stage. The same tendency can also be seen in the change of
compactness which is represented by distance-based cohesion. At the same time, in reality, the
performance of technological innovation in the Grace became very significant in the growing

Table 1. Abbreviation of actors.

Inside Actors Outside Actors
PRD Production Department UNI Universities
FIN Financial Department RI Research Institutes
HR Human Resource Department SPL Suppliers
RD R&D Department CST Customers
MKT Marketing Department CPT Competitors
PCH Purchasing Department GOV Government
LOG Logistic Department INV Investors
ST Science &Technology Department CSL Consulting Companies
IA Industrial Associations
LAW Legal Services Agencies
 166 J. Liu
Downloaded by [University of Arizona] at 19:54 21 January 2013

Figure 2. Socialgram of the TIN in the elementary stage (1997–1999)

Figure 3. Socialgram of the TIN in the growing stage (2000–2004)

 Innovation and Development 167
Downloaded by [University of Arizona] at 19:54 21 January 2013

Figure 4. Socialgram of the TIN in the plateau stage (2005– 2007)

Table 2. Network properties of the Grace’s TIN over the three stages
Density Efficiency
Centrality Diversity
Avg. Std Network Average Number of
value dev centralisation distance Compactness actors
Elementary 0.6601 1.3439 32.08% 1.485 0.333 12
stage
Growing stage 1.2745 1.6967 44.12% 1.434 0.696 17
Plateau stage 1.5294 1.7566 43.68% 1.451 0.775 18

stage compared to the elementary stage (Figure 1). Then the performance reached another new
peak and after that stopped growing in the plateau stage. I attribute the prosperity associated
with the technological innovation in the growing stage to significant and increased opportunity
and frequency of interaction between people. This was confirmed by the interviews with the
Grace’s President and managers. Since 2000, the Grace has implemented a so-called inside-
out-and-outside-in policy with which outside experts were invited to the company for technologi-
cal supervising or training and inside employees were sent out to visit companies or study in the
universities. In 2000, the Grace constructed an everybody-participating internal technological
innovation network. The employees from different functional departments with different knowl-
edge and skills were encouraged to talk to and work with each other. The project teams – a dis-
tinctively new structure whose members are from different departments – were created. These
project teams are very flexible in size and functions. The hybrid-matrix-like project teams can
be flexibly involved in big and small scale of innovation projects. To support and coordinate tech-
nological innovation projects, the Science and Technology Department (S&T) was set up in that
 168 J. Liu

year. The S&T Department together with the HR Department coordinated with different depart-
ments arranging personnel for technological innovation projects and these projects were given
different priority according to the firm’s business strategy. The team leaders of the projects
with high priority had the advantage to get particular people whom they wanted. As the S&T
Manager remembered:
At that time S&T was a new department which was probably the first one in the chemical fibre
industry in China. My job was to help people from both outside and inside, connect to each other
and work with each other. The outcome was great. We not only had good performance in tech-
nological innovation but also trained our employees and enhanced our technological
competences.
The further increase of density and decrease of network centralisation in the plateau stage,
which means more dense and more widespread interaction between actors, did not go together
Downloaded by [University of Arizona] at 19:54 21 January 2013

with a further increase of technological innovation performance. On the contrary, the performance
decreased. In an intensively connected network, there are stable relationships and mutual trust
among actors. According to knowledge management theory, such stable relationships and
mutual trust is significantly important for sharing a common standard and routines which will
further help in the exchange of know-how and tacit knowledge among actors. The learning
process is collective because it requires collaboration of different people with different capabili-
ties and knowledge. However, when the intensive connections exist for a certain period of time,
when everybody knows more and more about what the others know, the interaction or exchange
will become less and less productive and innovative. This can be explained by the S-curve of
group learning (Gersick, 1991). The explanation has been justified by the Grace managers.
The Vice President of Production commented:

Even though the statistics of technological innovation performance was still good in 2007, I have
already felt a kind of innovation fatigue. People are not as enthusiastic as before. When it comes
to developing inside potential of the employees, it becomes more and more difficult. When it
comes to cooperating with outsiders, the number of cooperation is increasing but the effects are
still doubtful.

Based on such observation in the Grace, I argue that more frequent organisational learning
results from a more connected and compacted network. In such a network more people can
cross shorter social distance to learn from each other. When the connectedness of an organis-
ational structure increases, it may indicate an increase of the extent of resource-sharing and
cooperation (Powell et al., 1996). Resource-sharing and cooperation form the main activity in
organisational learning. More frequent interaction between people provides more opportunities
for organisational learning in the network, consequently the performance of technological inno-
vation improves. But in terms of density it does not necessarily mean the higher the better. When
density increases to a certain level, the effectiveness of organisational learning may go down
because of the S-curve of group learning.

The size and diversity of the TIN significantly increased from the elementary stage to the growing
stage and thus improved organisational learning and innovation performance
The TIN of the Grace got more actors involved in the growing stage. One more insider and four
more outsiders joined in.
Among the five newcomers, the S&T Department (ST) acts as the only new functional depart-
ment inside the company. It was more like a knowledge broker than a knowledge resource for
organisational learning in the Grace. I observed that it has been in the core of the TIN after its
 Innovation and Development 169

establishment in 2000. Its degree, betweenness, and closeness have been in the highest group
(Table 2).
The other four newcomers are all outsiders, namely industry associations (IA), consulting
companies (CSL), legal services agencies (LAW), and competitors (CPT). The type of connection
with these four outsiders is either for open information which means accessing knowledge
without the need to pay, or acquisition of goods, such as purchasing machines, patents, license,
service, expertise, and so on. Although there was no active participation in joint innovation pro-
jects with these outsiders, the connections with them were still important for the Grace to enrich
their knowledge base and foster their organisational learning as these actors are of totally different
background and specialty from the Grace. But as the President of the Grace commented:

I learned something from our outside co-operators but far from enough. The effectiveness of learning
depended on how much the Grace could absorb and to what extent they could develop common goals
and find common interest between the Grace and its partners.
Downloaded by [University of Arizona] at 19:54 21 January 2013

I argue that the diversity of actors provides diversity of knowledge sets for organisational
learning. Given the increasingly specialised and professionalised nature of the knowledge on
which they are based, firms are path-dependent (Pavitt, 2005). Knowledge specialisation is a
must for the organisations to effectively acquire diversified knowledge of technologies. It is
hard for every single organisation to gain various specialised knowledge without relevant knowl-
edge background (Grant and Baden-Fuller, 2004). When the fusion of technology becomes more
and more typical in the occurring and developing process of technologies, fewer and fewer single
individuals or organisations will have the capability to innovate in isolation. Organisations with
broader networks are exposed to more experiences, various competencies and added opportu-
nities (Beckman and Haunschild, 2002). By having access to a more varied set of activities,
experiences and collaborators, companies broaden the resource and knowledge base that they
can draw on (Powell and Grodal, 2005). This argument was justified by the Grace’s managers
who said that since the growing stage more outsiders contributed to Grace’s technological inno-
vation by knowledge sharing and transfer, even though there were no joint technological inno-
vation projects with most of them, with the exception of the customers and suppliers.

6.3 The change of the actors’ position in the Grace’s TIN and its impact on organisational
learning and innovation performance
Analysing node properties of the TINs over the three stages (Table 3) together with the innovation
activities and performance (Figure 1) of each stage, I found some interesting findings. These are
as follows.

The more central an internal actor becomes, the more interactive learning it does, and the better
innovation performance it has.
When the centrality of the Science and Technology Department (S&T) and the R&D Department
(R&D) is higher than that of the Production Department (PRD) and the Marketing Department
(MKT), the performance of product innovation will be better than that of process innovation,
and vice versa.
First, I respectively discuss the change of the centrality of the five most centralised actors,
namely the PRD, RD, ST and MRK. The node centrality is the most frequently used parameter
to measure the positional power of an actor in the network. Second, I discuss the overall
change of the centrality of them as a whole.
 170
J. Liu
Downloaded by [University of Arizona] at 19:54 21 January 2013

Table 3. Node properties of the TIN of the Grace in the three stages
Nrm Degree Nrm Bonacich Power Nrm Betweenness Nrm Closeness
ELM GRW PLA ELM GRW PLA ELM GRW PLA ELM GRW PLA
MKT 16.471 47.059 52.941 3.819 6.425 6.165 2.42 3.775 6.203 13.71 43.59 85
FIN 25.882 45.882 44.706 6.001 6.264 5.206 0.795 10.195 6.323 13.821 48.571 85
RD 18.824 44.706 65.882 4.365 6.104 7.672 1.445 8.732 11.813 13.821 48.571 100
PRD 42.353 64.706 69.412 9.82 8.834 8.083 7.926 8.67 8.014 14.05 48.571 94.444
LOG 14.118 11.765 17.647 3.273 1.606 2.055 0.808 0.315 0.672 13.6 39.535 68
PCH 25.882 25.882 25.882 6.001 3.534 3.014 2.136 2.565 1.768 13.821 42.5 70.833
HR 28.235 35.294 35.294 6.547 4.819 4.11 3.65 1.958 1.983 14.05 43.59 73.913
ST 0 47.059 64.706 0 6.425 7.535 0 3.257 11.813 44.737 100
RI 7.059 14.118 12.941 1.637 1.927 1.507 0.699 0.604 0.174 13.6 41.463 60.714
UNI 7.059 15.294 15.294 1.637 2.088 1.781 0 0.315 0.174 13.281 40.476 60.714
GOV 30.588 35.294 29.412 7.093 4.819 3.425 3.65 2.252 0.082 14.05 43.59 62.963
IA 0 12.941 16.471 0 1.767 1.918 0 0.092 0.082 0 36.957 58.621
 Innovation and Development 171

The Production Department (PRD) continually took the most central position in the core of
the TIN of the Grace over the three stages, but its position slightly changed away from the
very centre in terms of betweenness and closeness from the elementary stage to the growing
stage and then the plateau stage. The degree and Bonacich power of the PRD have been the
highest among all the actors over the three stages of the Grace’s technological innovation (see
Table 4). However, its highest position in betweenness and closeness at the time of the elementary
stage and growing stage was overtaken by the RD and the ST in the plateau stage. That means that
over the three stages, the PRD not only has had the biggest number of connections with the actors
in the network, but also has been connected to the biggest number of powerful actors in the
network. But it is no longer the actor who has the most opportunities to be a broker and no
longer the one which is the closest to the others in the network at the time of the plateau stage.
The PRD’s central position is consistent with Grace’s core competence in low-cost-manufacturing
based on technological innovation which I identified in my previous research (Liu et al., 2009).
Downloaded by [University of Arizona] at 19:54 21 January 2013

The Marketing Department (MKT) has been one of the core actors in the TINs over the three
stages. Since the growing stage, the degree, betweenness, closeness and Bonacich power of the
MKT have ranked top three (see Table 4) despite the fact that the increase of its centrality was
not as good as that of the RD and the ST from growing stage to plateau stage. The MKT has
been an important knowledge broker between Grace and its customers. The information, such
as what the customers need, the benefit and problems they have when using the products of
the Grace, the suggestions they have for the Grace to improve the performance of their products,
was collected and conveyed by the MKT. The Grace’s customers are mainly located in the east
and south of China. The MKT has the strongest connection with customers among all the
other actors of the TIN.
The R&D Department (RD) and the S&T Department (ST) became more and more central
over the three stages. The absolute number of degree, betweenness, Bonacich power and close-
ness of the RD and the ST kept increasing throughout the three stages. The rank of the degree,
betweenness, Bonacich power and closeness of the RD kept increasing over the three stages.
The rank of the degree and Bonacich power of the ST remained the same since it was established
in the growing stage, while its rank of betweenness and closeness increased.
To summarise, first there is a continuous increase of degree, betweenness and closeness cen-
trality of almost all the inside actors over the three stages. The only obvious counter change hap-
pened to the Financial Department (FIN). But according to my investigation, the function of the
FIN in technological innovation is mainly collecting and sharing financial information. So, the
change of the centrality of the FIN did not have significant influence on organisational learning
for technological innovation in the TIN. Second, there is a change in the relative position of the
four key actors. The RD and the ST took the central position in betweenness and closeness. They
also had a significant increase in both degree and Bonacich power, while the most centered PRD
just had a slight increase in degree and even a decrease in Bonacich power.
It is interesting to see the delicate change of the relative position of the ST and the RD on the
one side, and that of the PRD and the MKT on the other side. By comparing Figure 1 and Table 4,
I I found that when the centrality of the PRD and the MKT is higher than that of the ST and the
RD, the performance of process innovation will mostly be better than that of product innovation.
When the centrality of the ST and the RD is higher than the PRD and the MKT, the performance of
product innovation will mostly be better than that of process innovation. When it comes to tech-
nological innovation activities, the ST and the RD work more with science, while the PRD and the
MKT work more with experience. Product innovation in chemical fibre industry is more science
based and process innovation in the industry is more experience based. An actor who is in a more
advantaged position has more opportunities and less constraint to communicate with the others in
the network. The outcome of the interaction between this actor and the other actors depends on the
 172 J. Liu

actor’s absorptive capability. Cohen and Levinthal (1990) refer to absorptive capacity as the
ability to recognise the value of new, external information, assimilate it and apply it to commercial
ends. They argued that the ability to evaluate and utilise outside knowledge is largely a function of
the level of prior knowledge. Actors’ knowledge background demonstrates what they can learn or
learn more efficiently and effectively from their co-operators. It also demonstrates what will be
integrated and transformed into actor-embedded knowledge. This finding can also be considered
as evidence that network accelerates organisational learning.
I argue that the more central the actor is in the TIN the more interactive learning the actor does
and the better technological innovation performance this particular actor has.

The peripherisation of an external actor in the TIN did not necessarily lead to the decrease of its
importance for organisational learning and innovation performance.
Downloaded by [University of Arizona] at 19:54 21 January 2013

Government was the core actor in the elementary stage from 1997 to 1999. Since then government
has been moving out of the centre of the TIN. It gives us an impression that in general government
was peripherised in the TIN. Especially from the growing stage to the plateau stage, the degree of
government decreased. But we should be careful when explaining the network dynamics only by
degree centrality measures because they only take into account the positional power and ignore
other powers of the actor which can be crucial to the existence and functioning of the network.
The role of government in the Grace’s TIN changed from leading and supervising technologi-
cal innovation to supporting and encouraging innovation in the company. This change was ver-
ified by both the Grace’s managers and a Director of the Economy Committee of Yibin City. Since
2004, the local government has decided to adjust its role from a supervisor to a supporter. The
government tended not to directly interfere in the daily operation of the state-owned company
but indirectly back them up, such as provide favourable policy and legal support, invest in
local vocational and higher education, increase common awareness of the importance of inno-
vation in society, etc. In this sense, the government had less direct connection within the
network but it worked more on maintaining the institutional environment for the existence and
functioning of the Grace’s TIN. The Grace considered that the support of the local government
is one of the most important factors for its success.
Therefore, I argue that a peripheral actor with low centrality may also be indirectly important
to the organisational learning in the TIN. The power of such actors does not come from their pos-
ition but, for example, from the capability to create a good environment for the existence and
functioning of the network, such as the case of government. This finding challenges the basic
assumption of social network analysis that the power of an actor only lies in its position in the
network.

7. Conclusion and discussion

In this paper I studied the 10-year-long dynamics of the TIN of an innovative textile company in
Southwest China. I tried to find out how the dynamics of the TIN impacted on organisational
learning and innovation performance. The main findings are as follows.
Connectedness of the network is positively related to more intensive and frequent interactive
learning and also to better technological innovation performance, but the relation between con-
nectedness and organisational learning is not linear. More connected and compacted network
results in more frequent organisational learning because more people can cross shorter social dis-
tance to learn from each other. When the connectedness reaches a certain level, organisational
learning will be less and less efficient because of the S-curve effect of group learning. Then
the technological innovation performance of the TIN may stop growing or grow at a lower rate.
 Innovation and Development 173

The diversity of the network is positively related to more effective interactive learning and to
better technological innovation performance. The diversity of the actors provides diversified
knowledge access to organisational learning, and increases the possibility of generating new
knowledge and creating new combinations of existing knowledge for technological innovation.
Higher centrality of an actor is related to more interactive learning of this particular actor in
the TIN, while lower centrality of an actor does not necessarily mean less importance of this actor
in the TIN. The more central the actor is in the TIN, the more interactive learning the actor will
experience and the better the technological innovation performance will be. A peripheral actor
may also be important to the performance of technological innovation with the power not from
their position but from other capabilities such as the capability of maintaining the institutional
environment in which the network is embedded.
The implication of this paper is twofold. First, from the managerial point of view, to boost
innovation, firms can enhance the relations among actors in the TINs and introduce more hetero-
Downloaded by [University of Arizona] at 19:54 21 January 2013

geneous actors to the TIN so as to improve organisational learning for innovation. But the positive
impact of networking on innovation is not infinite. Second, from the policy point of view, gov-
ernment can help firms to innovate by both leading and supervising their innovation activities,
and by providing proper institutional environment for firms and their collaborators to play in.
However, which approach is better depends on the specific context in which the TIN is embedded.
The findings of this paper echo the literature about the positive correlation of density, diversity
and centrality of network and interactive learning, as well as the performance of technological
innovation. The paper also challenges the basic assumption in innovation network research by
social network approach that the power of the actors is only from their position in the network.
As all case studies are somewhat weak in terms of external validity, it will be bold to draw
conclusions for policy suggestion or to summarise rationales for theory building based on one
case study. However, this case study adds some new knowledge to better understand the compli-
cated pattern of social interaction within the technological innovation network.

Acknowledgements
The author is grateful to Cristina Chaminade for her excellent suggestions and to the two anonymous referees
for their inspiring comments. Thanks to Ying Liu, Tao Feng, Yi Wang and other interviewees in Grace for
their continual support to the research in the past seven years.

Notes
1. Source: Publicity Department, Yibin Grace Group Co., Ltd.
2. Source: Publicity Department, Yibin Grace Group Co., Ltd. Report of the Soft Science Project of State
Intellectual Property Office of the People’s Republic China ‘Investigation and Case Study of The Situ-
ation of Intellectual Property Rights In Sichuan Import & Export Enterprises’.

References
Arrow, K.J. (1994) Methodological individualism and social knowledge. American Economic Review, 89(2),
pp. 1–9.
Beckman, C., and Haunschild, P. (2002) Network learning: The effects of partner’s heterogeneity of experi-
ence on corporate acquisitions. Administrative Science Quarterly, 47(1), pp. 92–172.
Bell, G.G. (2005) Clusters, networks, and firm innovativeness. Strategic Management Journal, 26(3),
pp. 287–95.
Bonacich, P. (1972) Factoring and weighting approaches to status scores and clique identification. Journal of
Mathematical Sociology, 2(1), pp. 113–120.
Borgatti, S.P. (2005) Centrality and network flow, 27(1), pp. 55– 71.
 174 J. Liu

Burt, R.S. (1992) Structural Holes: The social structure of competition. (Cambridge, MA: Harvard
University Press).
Cohen, W.M., and Levinthal, D.A. (1990) Absorptive capacity: A new perspective on learning and inno-
vation. Administrative Science, 35(1), pp. 128–152.
Ebadi, Y.M., and Utterback, J.M. (1984) The effect of communication on technological innovation.
Management Science, 30(5), pp. 572–586.
Edquist, C. (2005) Systems of innovation: Perspectives and challenges, in: J. Fagerverg, D.C. Mowery and
R.R. Nelson (eds) The Oxford Handbook of Innovation (New York: Oxford University Press).
Edquist, C., and Johnson, B. (1997) Institutions and organisations in systems of innovation, in: C. Edquist
(ed.) Systems of Innovation: Technologies, institutions and organizations (London and Washington:
Pinter/Cassell Academic).
Eisenhardt, M.K. (1989) Agency theory: An assessment and review. Academy of Management Review, 14(1),
pp. 57–74.
Freeman, L.C. (1979) Centrality in network: Conceptual clarification. Social Networks, 1(3), pp. 215–239.
Gersick, C. (1991) Revolutionary change theories: A multilevel exploration of the punctuated equilibrium
paradigm. The Academy of Management Review, 16(1), pp. 10–36.
Downloaded by [University of Arizona] at 19:54 21 January 2013

Grant, R.M., and Baden-Fuller, C. (2004) A knowledge accessing theory of strategic alliances. Journal of
Management Studies, 44(1), pp. 61–84.
Hanneman, R.A., and Riddle, M. (2005) Introduction to Social Network Methods (Riverside, CA: University
of California), Available at: http://faculty.ucr.edu/~hanneman/.
Herminia, I. (1993) Network centrality, power, and innovation involvement: Determinants of technical and
administrative roles. Academy of Management Journal, 36(3), pp. 471–502.
Lam, A. (2005) Organisational innovation, in: J. Fagerverg, D.C. Mowery and R.R. Nelson (eds) The Oxford
Handbook of Innovation (New York: Oxford University Press).
Liu, J. (2011) Regional institutional environment and its impact on intra-firm and interorganisational inno-
vation networks: A comparative case study in China and Switzerland, CIRCLE Working Paper Series:
WP 2011/07, Sweden.
Liu, J., and Baskaran, A. (2009) The university-industry-government relationship of technological innovator
networks in different national innovation systems: A comparative case study. African Journal of Science,
Technology and Development, 2(2&3), pp. 63–90.
Liu, J., Baskaran, A., and Li, S. (2009) Building technological innovation based strategic capabilities at
firm level in China: A dynamic resource based view case study. Industry and Innovation, 16(4– 5),
pp. 411–434.
Lundvall, B.A. (2007) National innovations systems: Analytical concept and development tool. Industry-
and Innovation, 14(1), pp. 95–119.
McCarthy, B. (2008) The evolution and transformation of networks: A study of private health insurance in
Ireland. Irish Journal of Management, 29(1), pp. 87–113.
Pavitt, K. (2005) Innovation process, in: J. Fagerverg, D.C. Mowery and R.R. Nelson (eds) The Oxford
Handbook of Innovation (New York: Oxford University Press).
Pettigrew, A. (1990) Longitudinal field research on change. Organization Science, 1(3), pp. 267–292.
Powell, W.W., and Grodal, S. (2005) Networks of innovations, in: J. Fagerverg, D.C. Mowery and R.R.
Nelson (eds) The Oxford Handbook of Innovation (New York: Oxford University Press).
Powell, W.W., Koput, K.W., and Smith-Doerr, L. (1996) Inter-organisational collaboration and the locus of
innovation: Networks of learning in biotechnology. Administrative Science Quarterly, 41(1),
pp. 116–145.
Sabatier, P.A., and Jenkins-Smith, H.C. (eds) (1993) Policy Change and Learning: An advocacy coalition
approach (Boulder, CO: Westview Press).
Sandstrom, A., and Carlsson, L. (2008) The performance of policy networks: The relation between network
structure and network performance. Policy Studies Journal, 36(4), pp. 497–524.
Tsai, Y.C. (2006) Effect of social capital and absorptive capability on innovation in Internet marketing.
International Journal of Management, 23(1), pp. 157–167.
Wasserman, S., and Faust, K. (1994) Social Network Analysis: Methods and applications (New York:
Cambridge University Press).