Content uploaded by Andrew D. James
Author content
All content in this area was uploaded by Andrew D. James on Dec 26, 2023
Content may be subject to copyright.
Formal versus informal knowledge
networks in R&D: a case study
using social network analysis
James Allen
1
, Andrew D. James
1
and Phil Gamlen
2
1
PREST, Manchester Business School, University of Manchester, Oxford Road, Manchester M13
9PL, UK. andrew.james@mbs.ac.uk
2
ICI Group Technology, Wilton Centre, Wilton, Redcar TS10 4RF, UK
The existence of informal social networks within organizations has long been recognized as
important and the unique working relationships among scientific and technical personnel have
been well documented by both academics and practitioners. The growing interest in knowledge
management practices has led to increased attention being paid to social network analysis as a
tool for mapping the nature and membership of informal networks. However, despite the
knowledge-intensive nature of research and development (R&D) activities, social network
analyses of the R&D function remain relatively rare. This paper discusses the role of informal
networks in the development, exchange and dissemination of knowledge within the R&D
function. A case study using social network analysis is used to compare and contrast formal and
informal knowledge networks within ICI. Marked differences between the informal organiza-
tion and ICI’s formal structures for knowledge exchange are revealed and a series of insights
into the working habits of technical staff are presented. The implications for managers are clear:
through a better understanding of the informal organization of R&D staff, they can more
successfully capture and exploit new ideas; more efficiently disseminate information throughout
the function; and more effectively understand the working habits and activities of employees.
1. Introduction
The importance of nurturing, accumulating
and efficiently deploying knowledge re-
sources for competitive advantage and value
creation is well understood. One theme emerging
out of knowledge management research, however,
is the failure of many managers to successfully
comprehend, support and ultimately exploit the
informal exchange of knowledge assets within
their organizations (Krackhardt and Hansen,
1993; Hansen et al., 1999; Cross and Parker,
2004; Anklam, 2005; Bryan and Joyce, 2005).
These studies highlight a crucial distinction be-
tween the formal organizational structures by
which companies attempt to manage and direct
the transfer of knowledge and the complex in-
formal social networks through which it flows in
practice (Cross and Parker, 2004).
This paper examines the effectiveness of the
formal organizational structures used within re-
search and development (R&D) to encourage
knowledge transfer. The paper argues that such
structures should be allowed to develop orga-
nically based upon an understanding of informal
networks rather than be designed and imposed
in a top-down manner. The issues raised are
examined through a case study of the distributed
R&D function of ICI PLC. Using social network
analysis, the case study describes the nature of
informal problem-solving networks within R&D,
and demonstrates how these compare and
contrast with equivalent formal organizational
structures.
R&D Management 37, 3, 2007. r2007 The Authors. Journal compilation r2007 Blackwell Publishing Ltd, 179
9600 Garsington Road, Oxford, OX4 2DQ, UK and 350 Main St, Malden, MA, 02148, USA
2. Context
2.1. Knowledge networks and the R&D
function
There is a long history of research that demon-
strates the inefficient nature of traditional corpo-
rate structures in enabling efficient knowledge
transfer within the R&D function. Allen (1977)
noted that technical staff maintain their own con-
tacts and sources of information, which may be
completely distinct from the organizational net-
works prescribed by senior managers. Further-
more, as the boundary of the modern company
has changed through the introduction of new
information technologies and increasing globaliza-
tion, formal channels of communication rarely
accurately reflect the working relationships be-
tween individuals (De Meyer, 1991; Cross et al.,
2002a, b). The myriad of personal communications
and ties which in reality disseminate knowledge and
information between individuals constitute the in-
formal social networks within R&D (Cross et al.,
2002a). These cross both organizational and geo-
graphical boundaries and they are crucial to the
ongoing work of scientific and technical employees
and the ability of the firm to innovate (Cross and
Parker, 2004). Understanding the manner in which
informal networks within R&D are formed, how
they are structured and how they function is there-
fore potentially critical to achieving success (Cross
et al., 2002b).
Research into the nature of informal commu-
nication and networks within R&D offers a series
of insights into the characteristics of informal
relations between scientific and technical person-
nel. Summarizing a series of studies of the com-
munication habits of R&D staff in science and
engineering stretching back to the early 1960s,
Tom Allen of the Massachusetts Institute of
Technology (1971; 1977) highlights the unique
nature of working relationships between technical
staff. He reports that technical employees are up
to five times more likely than other staff to turn to
a person, rather than a data source, to obtain
information important to their work (Allen,
1977). This is in significant part attributed to the
complex issues of trust that exist between scien-
tists. R&D workers tend to build very strong trust
relationships with the peers with whom they
collaborate and they are likely to turn to them,
and not to an alternative source, be it personal or
data, for assistance when it is required. As Allen
observes: ‘When there is social contact between
any two individuals, the probability of technical
communication is significantly higher then when
no social contact exists’ (Allen, 1977, p. 207).
Research on ‘Communities of Practice’ (CoPs)
documents how workers with similar working or
research interests often group together within an
organization (either physically, or increasingly in
the case of widely distributed firms, by making
use of information technologies) to collaborate
and share information and experiences (Wenger,
1998; Hildreth et al., 2000; Orlikowski, 2002).
Allen also notes how physical location has a
particularly strong effect on relationships between
technical staff. He observes that collaborative
projects based on regular face-to-face discussions
were significantly more successful than those
based on less frequent, often distant, communica-
tions (Allen, 1971). Building upon Allen’s re-
search, De Meyer (1983, 1991) reaffirms the
importance of face-to-face contact in R&D work.
At the same time, research examining network-
ing activities within R&D project teams recognizes
the key knowledge-sharing roles of a few specific
individuals within collaborative networks (Allen,
1971, 1977; Allen et al., 1971; Tushman, 1978;
Katz and Tushman, 1979; Tushman and Scanlan,
1981a, b; Katz and Allen, 2004a, b). The role of
‘boundary-spanning individuals’ and ‘technologi-
cal gatekeepers’ are highlighted. Boundary-span-
ning individuals are technical staff who adopt a
linking role by operating between separate net-
works or groups, providing a route for knowledge
transfer between them (Tushman and Scanlan,
1981a, b). This is of particular relevance in studies
of distributed R&D structures, where effective
communication between separate units requires
individuals to transfer knowledge and information
across organizational boundaries within the firm.
Similarly, the role of technological gatekeepers in
enabling knowledge flow has been highlighted
with regard to sources of knowledge that are
external to the firm (Allen, 1971).
2.2. Formal knowledge management
Studies of knowledge management and the
‘knowledge-based organization’ recognize that
although internal knowledge assets are vital in
driving commercial performance, success is con-
ditional on their being effectively exchanged and
exploited (Blacker, 1995; Tsoukas, 1996; Daven-
port and Prusak, 1998; Leonard, 1998; Sparrow,
1998; Teece, 1998; Hansen et al., 1999; McAdam
and McCreedy, 1999; Metcalfe and James, 2000;
Nonaka and Teece, 2001; Sanchez, 2001).
James Allen, Andrew D. James and Phil Gamlen
180 R&D Management 37, 3, 2007 r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
Companies have devoted substantial resources
to knowledge management with the aim of facil-
itating improved knowledge exchange within the
organization, and supporting and exploiting ex-
isting accumulated tacit knowledge and data
sources. Their motivation is the development
and implementation of new structures, policies
and methods by which to better access, codify,
organize and manage internal knowledge and
information within the firm (Coombs and Hull,
1998). Firms with a decentralized R&D function
have adopted a wide range of approaches in an
attempt to enable and support horizontal knowl-
edge transfer, communication and cooperation
between separate units, including specific at-
tempts to increase collaboration within R&D.
Katz and Allen (2004c) document a selection of
attempts made by companies to aid and encou-
rage networking between technical personnel,
including the Monsanto Company’s ‘Monsanto
Technical Community’ and 3M’s ‘Proprietary
Company Fairs’ (Katz and Allen, 2004c).
Encouraging networking between technical per-
sonnel displays a recognition that a suitable
organizational culture and the promotion of so-
cialization can act to improve productivity (De
Meyer, 1991). De Meyer also highlights the intent
of many firms to nurture a ‘family atmosphere’
between R&D staff, using methods such as tem-
porary assignments within partner laboratories,
encouraging travel between sites and regular
cross-function training programmes (De Meyer,
1991). Bryan and Joyce (2005) cite further exam-
ples of organizational structures created to im-
prove inter-functional communication, such as
internal joint ventures, task forces and study
groups (Bryan and Joyce, 2005).
Nevertheless, Bryan and Joyce (2005) argue
that the benefits have been limited despite the
large investments made by companies. They sug-
gest that this is because the true value of knowl-
edge is not recovered through its management but
rather from the actual practice of its creation and
exchange. They argue that in order to enable
knowledge transfer to occur more extensively
focus should be placed on the structuring of
knowledge and its transfer within the firm so as
to reduce the organizational obstacles that im-
pede its flow.
2.3. Formal and informal structures
A division has therefore become apparent between
how knowledge transfer is formally organized by
companies, and how it occurs within firms in
reality. This distinction between formal and infor-
mal organization is nothing new. Burns and Stalker
(1961) distinguished the ‘formal structure’ of the
organization (its well-defined management systems
and structures) from the ‘informal structure’ or
‘private organization’ (the processes by which
individuals communicate on issues not directly
laid down and governed by management) (Burns
and Stalker, 1961). Chandler (1962) introduced the
concept of social networks as the essential struc-
tures upon which both formal and informal com-
munication and knowledge transfer are based. He
defines formal social networks as those that are
prescribed and forcibly generated by management,
usually directed according to corporate strategy
and mission. In contrast, informal social networks,
or emergent networks, are unsanctioned and un-
governed organic structures connecting a poten-
tially unbounded group of individuals (Mintzberg,
1973; Tichy, 1981). In the context of the firm, these
informal networks extend not only internally but
also externally across organizational boundaries.
They include the working relationships, collabora-
tions and exchanges of knowledge between indivi-
duals which are not found in organizational
structures, but are the result of the personal in-
itiative of employees (Cross and Parker, 2004). As
well as the personal initiative of employees, of
course, organizations outside the firm may seek
to ‘recruit’ the firm’s R&D employees to become
members of professional associations, scientific
bodies, advisory groups and visiting faculty in
university departments.
Managerial social network studies have sought
to establish the extent of these informal networks
within organizations and assess how the informal
organization compares with the formal structures
prescribed by management (Cross et al., 2001,
2002a, b; Cross and Parker, 2004). Critically, they
observe that the patterns of collaboration and
communication revealed in informal networks are
significantly different from the formal organiza-
tional structures implemented by managers. Thus,
formal organizational structures fail to reflect
accurately the true nature of social relationships
and the dynamics and dependencies between staff.
This can be to the detriment of efficient knowledge
exchange within the firm (Cross et al., 2002b).
The key observations arising from existing
studies of informal networks are well summarized
by Cross and Parker (2004). In brief, studies of
informal relationships using social network
analysis have revealed critical disconnections be-
tween business groups or individuals, the often
Formal versus informal knowledge networks in R&D
r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
R&D Management 37, 3, 2007 181
unexpected locations of key individuals; who act
to facilitate communication and link disparate
groups, and significant instances of important
personnel failing in their formal task of enabling
and supporting collaboration. Senior managers
are often revealed by network analysis studies
to be peripheral figures in informal networks.
Equally, those who are found at the hub of an
informal network are frequently individuals who
otherwise go unrecognized, and therefore unsup-
ported in the firm. Consequently, one of the great
strengths of informal network studies is that they
enable firms and managers to diagnose the true
nature of collaborations within their organization
and thereby better support and facilitate effective
working. Cross and Parker (2004) emphasize the
importance of a well managed and understood
informal network to successful organizational
performance. Network studies can also reveal
the effect of merger and acquisition activities on
the interaction of newly combined workforces or
allow monitoring of the effect of organizational
changes made by companies on their employees.
2.4. Barriers to knowledge exchange
Studies of informal networks are also instructive in
revealing how the structure of organizations can
impede knowledge flow. The geographical and
organizational boundaries separating staff are often
cited as factors affecting communication and colla-
boration within firms. However, informal network
studies have the power to confirm or deny such
claims and enable companies to act in a targeted
manner in response. In a study of the informal
networks between R&D staff distributed across
several countries, Cross and Parker (2004) observed
that national boundaries indeed formed an obstacle
to the transfer of knowledge between them.
Leonard (1998) proposes that the ideal of well-
diffused and widespread knowledge is particularly
threatened by the tendency of organizational
boundaries, such as those between divisions or
functions, to result in the formation of what she
terms ‘islands of knowledge’ within the firm. This
point is emphasized by Brown and Duguid (2001),
who highlight that knowledge-intensive work is
generally conducted in a manner removed from
that prescribed by organizational charts and for-
mal procedures, and can therefore be threatened
by strict adherence to such structures.
Thus, despite efforts to ‘flatten’ organizational
structures through the removal of management
tiers and the promotion of decision-making at
lower levels of the firm, concerns remain that
formal organizational structures continue to im-
pede, rather than to aid, knowledge transfer
(Brown and Duguid, 2001). The traditional M-
form organizational structure and its many variants
are characteristically rigid and inflexible, thereby
hindering horizontal communication between
functions, or the separate businesses of multidivi-
sional firms (Daft, 2004; Bryan and Joyce, 2005).
The issues afflicting hierarchical systems are symp-
tomatic of the ‘mechanistic’ management systems
identified by Burns and Stalker (1961). These
mechanistic systems are characterized by vertical
interactions, systems of superiority management
and a focus upon local, rather than broad sources
of knowledge, experience and skills. This demon-
strates how the flow of knowledge within the
corporation can be restricted through bureaucracy
and the use of rigid frameworks for reporting and
sharing knowledge assets. Vertical ‘silos’ of em-
ployees are separated by functional boundaries, or
in the case of multiunit firms, by business group,
increasing duplication of resources, reducing effi-
ciency and critically impeding the exchange of
knowledge assets.
2.5. The complex ecology of organizations
The operational restrictions generated by a purely
hierarchical organizational structure are well re-
cognized. Blacker (1995) observes that to realize
the additional value and potential of knowledge
assets, organizations must actively support and
facilitate their dynamic exchange through im-
proved organizational approaches to knowledge
sharing and increased dialogue among personnel.
However, rather than abandon traditional ap-
proaches in favour of more organic systems of
management, various retrofit organizational
structures have been used to adapt hierarchies
and forge linkages between similar activities
across the firm (Brown and Duguid, 2001; Bryan
and Joyce, 2005). Matrix structures associate
professionals horizontally on shared product or
competency lines across functional, divisional or
geographic boundaries (Daft, 2004; Bryan and
Joyce, 2005). The approach integrates separate
areas of knowledge on specialist subjects other-
wise isolated from one another by vertical man-
agement systems. Modern management and
organizational practices are, however, increas-
ingly more dynamic than matrix structures, which
retain much of the rigidity of M-form structures,
reflected by infinite variations of formal colla-
borations. These approaches are sometimes col-
James Allen, Andrew D. James and Phil Gamlen
182 R&D Management 37, 3, 2007 r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
lectively termed ad hoc organizational overlays
(Bryan and Joyce, 2005) and demonstrate com-
mitments made by companies to reduce excessive
organizational layering and promote decision
making and knowledge sharing at lower hierarch-
ical levels (Cross et al., 2002b).
This brings about what Brown and Duguid
(2001) label a ‘complex ecology’ of the organiza-
tion, whereby firms outwardly retain their M-form
structure, while developing processes within for
knowledge sharing and innovation. In practice,
however, retrofit structures are often found to be
cumbersome, inefficient and slow to respond to
changes in the business environment (Bryan and
Joyce, 2005). Difficulties are frequently experi-
enced in attempting to move knowledge within
the firm (Brown and Duguid, 2001). Leonard’s
(1998) warning of the danger of creating ‘islands
of knowledge’ is relevant here, as imposed groups
may become isolated from the wider corporation
and as a result develop tendencies to rebuff other
collaborations and only focus upon highly specific
knowledge areas. The establishment of separate
structures within the strategic business units
(SBUs) of multidivisional firms similarly risks
enabling what Prahalad and Hamel (1990) term
the ‘tyranny of SBUs’. By granting SBUs greater
autonomy through full divisional control of their
own functions, units frequently concentrate on
their own strategic focus at the expense of devel-
oping competences and sharing knowledge on a
group-wide basis (Prahalad and Hamel, 1990).
3. Research methodology
As the understanding of informal relationships
has become increasingly recognized as a useful
tool for managers, the need to map and document
informal networks precisely has become para-
mount. Various researchers within the knowledge
management community (Tichy et al., 1979;
Georghiou et al., 1988; Cross and Parker, 2004;
Dahl and Pedersen, 2005; Mote, 2005) have there-
fore used a methodology developed within the
field of sociology termed ‘social network analysis’
(Scott, 1991; Wasserman and Faust, 1994).
To investigate the themes discussed in the
previous section, a research project was con-
ducted, centred on a social network analysis of
staff working within the R&D department of ICI
PLC, the large multidivisional chemicals com-
pany. The sample frame consisted of 152 senior
R&D personnel. These were drawn from the 400
R&D staff worldwide in the top four levels of
ICI’s hierarchy. The 152 individuals comprised
the total membership of ICI’s ‘Science Ladder’
and ‘Expert Groups’. The Science Ladder and
Expert Groups are ICI’s principal cross-func-
tional knowledge-sharing structures. We describe
them in the next section and here it is sufficient to
say that we selected members of these groups
because the formal role of these structures is to
promote knowledge sharing. The R&D staff were
surveyed by electronic questionnaire to determine
the nature of the informal problem-solving net-
works between them. In so doing, those ques-
tioned were asked who they most often turned to
in their daily work activities for help in thinking
through a new or challenging scientific or techni-
cal problem (Cross and Parker, 2004).
1
Each
respondent was asked to nominate up to five
people in answering the question, rating each
nomination by frequency and mode of contact.
This followed the methodology for co-nomina-
tion set out in Nedeva et al. (1996).
Of the targeted 152 individuals, 130 responses
were received, representing a response rate of
86%. The response data were processed using
the UCINET software package (Borgatti et al.,
2002) and the network maps or ‘sociograms’
presented in the following section were developed
using the NetDraw utility (Borgatti, 2002). Socio-
grams represent the network as a series of nodes,
which denote individuals, connected by linear
ties, indicating the presence of a relationship
between individuals.
2
In the networks analysed,
ties are directional, with an arrow head indicating
the direction of nominated collaborative choices.
Software packages such as those used in this
investigation automatically transform raw net-
work statistical data to generate sociograms. In-
dividuals with the most ties to others are generally
placed at the centre of the network, and are
known as focal nodes. The software groups re-
lationship clusters and will equalize the length of
ties where possible.
Individuals who act as an important connec-
tion between two or more clusters within the
network are referred to as ‘boundary spanners’.
Conversely, nodes with relatively few ties con-
necting them to other individuals in the network,
and who do not form part of a critical path
between other members, are located at the edge
of the network. Individuals who form part of a
survey population but who do not have any ties to
other members of the network are referred to as
‘network isolates’.
It is important to emphasize that the scope of
this investigation was limited to examination of
Formal versus informal knowledge networks in R&D
r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
R&D Management 37, 3, 2007 183
the informal problem-solving network within ICI,
i.e. collaborations between colleagues within the
company rather than with persons external to the
company. From an organizational perspective,
the mapping of how employees interact infor-
mally in their working collaborations within the
company can tell us a great deal about the
effectiveness of formal structures. Of course, ex-
ternal networks and links with the scientific com-
munity are very important for research scientists
but the aim of our study was to compare and
contrast formal versus informal knowledge net-
works in the R&D function of a firm.
4. Knowledge networks and the R&D
function at ICI
ICI is a multinational chemicals company, head-
quartered in the United Kingdom, which devel-
ops and manufactures the specialty effects
chemicals used in a wide variety of products,
including paints, foods and fragrances. The ICI
Group comprises the four international busi-
nesses of National Starch, Quest, Uniqema and
ICI Paints. In addition, small localized operations
are maintained in the Group’s Regional and
Industrial businesses (ICI, 2004).
The present-day ICI Group exists as a result of
an extensive series of mergers, acquisitions and
divestments conducted over the past two decades
(Morrison, 2003). Historically, the company,
formed in 1926 as Imperial Chemical Industries,
was a producer of bulk chemicals. Throughout
the 1980s and 1990s, however, it engaged in a
period of major strategic change; electing to leave
the bulk chemicals market in favour of producing
effects chemicals. This was achieved through the
divestment of its heavy chemicals business to
various parties, and critically, in 1997, the acqui-
sition of Unilever’s Specialty Chemicals Division
for d4.9 billion (ICI, 2004).
4.1. Formal organizational structures for
knowledge exchange
4.1.1. Location and structure
R&D in ICI is decentralized across its constituent
businesses according to a third-generation model
of R&D organization (Roussel et al., 1991; Ste-
venson, 2002). This approach is aimed at achiev-
ing a synergy between business and technology
strategy, such that the majority of R&D work
performed within each particular business is tai-
lored to its individual market needs (Gamlen
et al., 2003). ICI’s businesses are distributed
worldwide and their individual R&D facilities
are often widely separated. In keeping with
third-generation models, ICI maintains a small
central technology function with a hybrid mission
to oversee and guide R&D activity within the
businesses and facilitate inter-business technology
transfer (Stevenson, 2002; Gamlen et al., 2003).
The corporate centre consists of the Group
Technology Board, whose role it is to promote
best practice and knowledge sharing between
businesses via Group-wide technology platforms;
a small Group Technology Office (GTO), which
organizes and facilitates collaboration and also
plays an important role in seeking to reduce cases
of R&D duplication across the businesses; and
two specific skill centres focused upon longer-
term, general scientific research on capabilities of
importance to all group businesses. These are
the Strategic Technology Group (STG) and the
Measurement Science Group (MSG).
4.1.2. Knowledge transfer between businesses
The distribution of R&D within the ICI Group’s
businesses, supported by the corporate centre, is
displayed graphically by the company as a ‘pla-
netary map’ (see Figure 1). Direct business-to-
business knowledge transfer and linkages between
the dispersed R&D centres are actively encour-
aged as a way of generating company-wide benefit
from knowledge generated within individual busi-
nesses. The explicit route for these transfers is
termed the Business Linking Programme (BLP),
implemented with the intention of achieving tech-
nological synergies across the Group (Stevenson,
2002; ICI, 2005c). Communication pathways and
central funding are provided by the Technology
Board to support inter-business knowledge trans-
fer. A corporately funded Strategic Research
Fund also exists to support research that is of
interest to multiple ICI businesses and pursued in
partnership with major universities.
ICI has also introduced a series of cross-func-
tional knowledge-sharing groups. These are in-
tended to reflect its technology strategy and
promote the exchange of knowledge and informa-
tion assets from business to business (Stevenson,
2002; Gamlen et al., 2003). These cross-functional
knowledge sharing groups are:
The ‘Megathemes’: four steering groups com-
prised of senior staff from all ICI businesses,
which guide research and practice in the group
innovation platforms by directing and nurtur-
James Allen, Andrew D. James and Phil Gamlen
184 R&D Management 37, 3, 2007 r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
ing the Expert Groups for which they are
responsible.
The ‘Expert Groups’: teams of expert scientists
and technologists also drawn from across the
spectrum of ICI companies, who collaborate in
working on the specific subject areas selected
by the technology strategy as being beneficial
to the whole group. Expert Groups are charged
with the role of coordinating research efforts
on their subject area, and working in colla-
boration with other groups where appropriate.
The ‘Expert Networks’: prescribed knowledge-
sharing networks in the fields of formulation,
sensory research and delivery systems. The
expert networks consist not only of specialist
scientists but also of commercial participants
from the Group businesses. Rather than work
on specific technical issues, they are tasked
with advising the direction of R&D activities
in their field and stimulating new business
ideas.
In addition to these formal groups, ICI also
maintains a structured network of expert scien-
tists working within its various R&D centres. The
company’s ‘Science Ladder’ ranks appointed ex-
perts within each business over four levels accord-
ing to their seniority.
4.2. Informal networking within the ICI
group
Social network analysis was used to map the
informal problem-solving network between em-
ployees in ICI’s R&D function (see Figure 2).
Individual staff members are shown as ‘nodes’ –
single points on the diagram, and the relationships
between them as ‘ties’ – lines joining the nodes.
The key shown in Table 1 assists in the reading of
the diagram. The identities of the four businesses
have been anonymized at the request of ICI.
A number of important observations can be
made from the network:
(1) Technical communications on R&D issues
remain predominantly within the individual
businesses and the network demonstrates that
the vast majority of problem-solving relation-
ships exist only between individuals working in
the same business. There are very few working
collaborations between individuals in different
businesses. Notably, of 42 reciprocated nomi-
nations in the network, 38 are between employ-
ees working within a single business unit.
(2) When personnel do go outside their own busi-
ness boundaries for assistance on a technical
problem, it is generally not to peers in the other
Group businesses, but to staff working in the
corporate centre. This is particularly striking
when members of the centre are removed from
the network (see the second sociogram shown
in Figure 2). Without members of the centre,
the four businesses are revealed to have
almost entirely self-contained networks,
with little or no contact with the other
businesses in the Group.
(3) Even the ties to the centre on a per-business
basis remain generally lower than might be
N
N
NS
S
SC
C
C
R
R
R&
&
&D
D
D
BUSINESS TO BUSINESS
TECHNOLOGY BOARD
GROUP
TECHNOLOGY
OFFICE
STG, MSG
P
P
Pa
a
ai
i
in
n
nt
t
ts
s
s
R
R
R&
&
&D
D
D
U
U
Un
n
ni
i
iq
q
qe
e
em
m
ma
a
a
R
R
R&
&
&D
D
D
Q
Q
Qu
u
ue
e
es
s
st
t
t
R
R
R&
&
&D
D
D
Figure 1. ICI planetary map showing intended knowledge transfer within research and development (Stevenson, 2002).
Formal versus informal knowledge networks in R&D
r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
R&D Management 37, 3, 2007 185
predicted, given the scale of the network.
Statistical testing revealed that only in the
case of ties between individuals in Business
‘C’ and colleagues in the corporate centre are
the number of relationships observed to be
proportionally greater than the expected nor-
mal level based on the number of staff sur-
veyed in each business.
3
(4) The informal network reveals that direct
business-to-business collaborations are extre-
mely limited. Only a handful of ties were
observed between businesses:
Business A
(Corporate Centre)
Business B Business C
Business D Business E
Business B Business C
Business D Business E
Figure 2. The informal problem-solving network within ICI’s research and development function, shown top with, and below
without, the inclusion of members of the Corporate Centre.
James Allen, Andrew D. James and Phil Gamlen
186 R&D Management 37, 3, 2007 r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
Just one tie was reported between employees of
Businesses ‘C’ and ‘B’, and similarly between
pairs ‘C’ and ‘D’ and ‘C’ and ‘E’.
Only two ties exist between Businesses ‘B’ and
‘D’, both of which originate from a single
individual in Business ‘B’.
There are no ties at all between staff working in
Businesses ‘B’ and ‘E’.
None of the inter-business ties observed were
reciprocal.
(5) Many of the individuals in the informal net-
work are isolated from their colleagues in the
rest of the function. In some cases, indivi-
duals, or small groups, are completely dis-
connected from the wider local business
networks.
(6) It is striking that a number of the formally
designated ‘experts’ surveyed are located on
the fringes of the network, only loosely con-
nected to the wider community.
4.3. Informal networks within formally
prescribed structures
The informal problem-solving networks among
only those employees who comprise the member-
ships of ICI’s Science Ladder and Expert Group
communities were also examined. Many of the
observations highlighted in the wider Group net-
work are also revealedto be true within the formally
prescribed structures the company maintains.
4.3.1. Science Ladder informal network
The informal problem-solving network among
members of the Science Ladder is displayed in
Figure 3. Members who neither reported relation-
ships with their fellow members nor were nomi-
nated by others in the community (and therefore
have no ties within the network) are displayed as
unconnected nodes in the top left-hand corner of
the sociogram. In this instance, the direction of
nominations is also shown by arrowheads on the
ties between nodes in the network. Reciprocated
nominations are highlighted by darker ties.
Examining the network:
(1) Experts appointed to the Science Ladder can
be seen to cluster closely together with collea-
gues on the Ladder within their own business.
Very few cross-boundary relationships are
observed; when a solution to a problem needs
to be found, it is rare that any member
contacts a peer in one of the other Group
businesses.
(2) Eleven of the 49 members of the Ladder did
not nominate any of their colleagues when
surveyed and were not themselves nominated.
Over a fifth of the Ladder therefore makes no
use of its potential as a structure for knowl-
edge exchange.
(3) The clustering of members by business does
not result in the formation of strong informal
networks among personnel from a single
business. Instead, fragmented clusters can be
seen to form between members of businesses
‘C’ and ‘E’ and members from the remaining
businesses display no network cohesion.
(4) A single individual, node A4, performs a vital
role in acting as a boundary-spanning indivi-
dual within the network. If he or she is
removed then the informal network within
the Ladder is wholly fragmented.
4.3.2. Expert Group informal network
The lack of strong collaboration among experts
appointed to the Ladder and clustering by business
observed in the overall Group network could
potentially be attributed to the fact that the highly
specific technical knowledge that individuals in
the R&D function are likely to seek is often
located solely among their own company staff.
For example, individuals are likely to work on
projects with shared specialist knowledge. How-
ever, ICI has explicitly set out to encourage
Table 1. Key to network diagrams in Figures 2 and 3
Node colour
Black Surveyed individuals
Grey Nominated individuals
outside of the survey
population
1
Node shape
Circular Business A (corporate centre)
Square Business B
Upright triangle Business C
Diamond Business D
Inverted triangle Business E
1
These were non-surveyed individuals who were nominated
by those questioned as being colleagues whom they turned
to for help in solving scientific or technical problems. These
individuals were R&D personnel, but they were not surveyed
themselves as they were not members of ICI’s Science
Ladder or Expert Groups, and did not therefore fall within
the expert community which was targeted for analysis. They
were not themselves surveyed due to the limitations of time
and resource in the investigation. Further research could
extend the analysis to include the non-surveyed individuals
in order to achieve a ‘snowball’ sample i.e. progressing the
survey until all members of the population have provided
a response.
Formal versus informal knowledge networks in R&D
r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
R&D Management 37, 3, 2007 187
collaboration across the Group and seeks to
exploit opportunities for technology and knowl-
edge synthesis. It has already been discussed how
the company maintains a series of cross-functional
Expert Groups, whose members are intended to
collaborate with one another on scientific and
technical subjects of shared value and importance.
R&D personnel selected as members of the Expert
Groups have extensive and leading experience in
the appropriate area of research (Gamlen et al.,
2003), and can reasonably be considered to share
the subject of their group membership in their
working environment. Hence, mapping the infor-
mal problem-solving networks of the individual
populations of these groups provides further
insight into the communication habits of expert
R&D staff, revealing the extent to which they
communicate with their formally designated
peers.
Figure 4 displays the network surrounding the
members of one of the Expert Groups that were
mapped. The diagram shows not only the relation-
ships between group members but also those that
members have with R&D staff outside of the group
(an ‘ego-net’). The node shapes again indicate the
Business to which each individual belongs. However,
the key shown in Table 2 outlines a change in the
coding of nodes from the previous diagrams shown.
Examining the ego-net:
(1) The informal network revealed between mem-
bers of the group is demonstrative of a trend
witnessed in the majority of the groups
mapped. Where there is collaboration be-
tween members, it occurs within clusters of
experts based within the same business.
(2) There is very little communication on pro-
blem-solving issues between group members.
In several instances, even members drawn
from within the same business nominated no
collaborations with their colleagues in the
Expert Group.
(3) In terms of knowledge dissemination to the
businesses, the informal network connected to
Figure 3. The informal problem-solving network among members of ICI’s Science Ladder.
James Allen, Andrew D. James and Phil Gamlen
188 R&D Management 37, 3, 2007 r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
the group reveals some positive signs – several
members are frequently cited by peers within
their own businesses.
(4) However, a significant number of members
were not nominated by any other surveyed
individuals, including D24, E13, B12 and A16.
(5) Conversely, it is noted that two individuals
closely linked to the Expert Group, nodes
A13 and C46, are highly cited by those in
close proximity to group members, but are
not themselves members of the group.
4.4. Informal networks at business level
The insight offered by a social network analysis
approach is well demonstrated by the informal
networks generated by the survey among mem-
bers of each of the ICI Group businesses. Figure 5
shows the informal network among the formally
designated ‘experts’ within the R&D function in
Uniqema. Uniqema was formed relatively re-
cently through the merger of several former ICI
and Unilever divisions following the acquisition of
USC in 1997. Uniqema has R&D activities in
several countries (Morrison, 2003). Individual
nodes in the network have been shaded according
to the country in which they are located. Table 3
which precedes the network, shows further amend-
ments to the original key detailed in Table 1.
Briefly examining the network within Uniqema:
(1) The informal network among employees is
well consolidated. There are very few isolated
members across the business. There are sev-
eral focal, highly cited, individuals at the
heart of the network and a broad spread of
collaborative activity across the business.
(2) Considering the various geographical loca-
tions of the employees in the network, it can
be seen that despite the wide distribution of
Figure 4. Ego-Net for an example Expert Group.
Table 2. Amendments to key to network diagrams used
in Figure 4
Node colour
Black Expert Group members
Grey Non-members
Formal versus informal knowledge networks in R&D
r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
R&D Management 37, 3, 2007 189
staff, nearly all are well integrated into the
network. Indeed, in many cases, individuals
have reported collaborative activity with col-
leagues based in other countries. There is little
obvious clustering by geographical location of
employment as might be expected.
5. Analysis
5.1. Distinction between the formal and
informal organization
The social network analysis of the informal pro-
blem-solving networks within R&D at ICI reaf-
firms a key theme within organizational and
network studies. The informal networks within
the firm are markedly different from the formal
organizational structures implemented by the
company in attempting to encourage the
exchange of knowledge between its technical
personnel.
This distinction is most clearly revealed in the
disparities observable between the formal model
used by ICI to describe intended routes of knowl-
edge transfer between its constituent businesses
and the highly business-centric informal network
that was found in practice. In contrast to the
strong connections between the four Group busi-
nesses envisaged in the model, very little colla-
boration was in fact found to occur across
business boundaries. Instead, problem-solving
activity was highly concentrated within the indi-
vidual businesses. The lack of any reciprocal ties
linking personnel across organizational bound-
aries is also revealing as it suggests that even those
few business-to-business relationships that do
exist are themselves fragile.
Figure 5. Informal problem-solving network within Uniqema research and development.
Table 3. Amendments to key to network diagrams used
in Figure 5
Node colour
Black Employees based in the Netherlands
Dark grey Employees based in the United States
Light grey Employees based in Britain
White Employees based at other global sites
James Allen, Andrew D. James and Phil Gamlen
190 R&D Management 37, 3, 2007 r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
The role of the small central technology func-
tion was revealed to be critical in linking the
research activities of the Group businesses, pro-
viding a channel for communication on areas of
shared strategic interest and the exchange of
knowledge between the businesses. This is clearly
demonstrated if the centre is removed from the
network, as is shown in Figure 2.
That the broader network does not reflect ICI’s
formal plan for knowledge transfer between its
businesses is perhaps unsurprising given the re-
sults of previous research, and specifically the
series of case studies presented by Cross and
Parker (2004). The cases they discuss show how
the informal networks were structured differently
from their equivalent formal structure. Indeed,
Cross and Parker’s study of information flow
between staff in a distributed R&D group dis-
plays a pattern of communication strikingly simi-
lar to that revealed by the network analysis
carried out within ICI. Cross and Parker show
that the dispersed researchers worked in tightly
clustered groups according to their country of
operation and argue that this is a common feature
of the collaborative behaviour of technical staff
(Cross and Parker, 2004). The pattern of informal
knowledge networks within ICI bears direct
comparison with the grouping by business wit-
nessed in Cross and Parker’s study. R&D staff
appear to associate themselves with others who
share their own organizational identity and phy-
sical location.
A variety of explanations can be put forward
for the fragmentation within the individual busi-
ness clusters observed in this investigation. This
may be the result of staff working in different
countries and in separate research centres as well
as the complex recent history of the company.
The constituent ICI businesses are geographically
and culturally independent from one another,
the result of a number of acquisitions and the
restructuring of operations in the late 1990s. Quest
and Uniqema, acquired from Unilever in 1997,
were themselves the products of reorganization
within Unilever itself in the years preceding the
deal. The resulting impacts on cultural identity
therefore remain fresh, and may have contributed
to a resistance towards greater interaction with
the wider ICI Group.
Uniqema is a notable exception. The informal
network among Uniqema R&D staff is very well
integrated despite their location at a series of
geographically dispersed sites. This may be a
function of effective post-acquisition integration
management. Uniqema, unlike the other ICI
businesses, had embarked on a deliberate and
wide-ranging integration programme designed to
develop a business-wide cultural identity after its
formation. This appears to have paid significant
dividends. Uniqema contradicts the general sup-
port that this investigation gives to Cross and
Parker and Allen’s work, and merits further
analysis of the impact of national or geographical
boundaries on informal R&D networks.
The ICI case reflects the claim made by Cross
and Parker (2004) that informal networks are
essentially invisible to managers, and as such do
not mirror what is shown on formal organiza-
tional charts. The importance of the distinction
between the formal and informal organization
made by Chandler (1962) and Burns and Stalker
(1961) is therefore reaffirmed. Formal organiza-
tional structures are described as being prescribed
and predictable. The informal network mapped in
the analysis is an essentially organic structure that
has emerged out of the independent actions and
working habits of R&D personnel. It is worth
noting, however, that the social network analysis
revealed no evidence of the existence of strong
communities of practice within ICI’s R&D func-
tion. Dorothy Leonard asserts that there is a
danger of creating ‘islands of knowledge’ by
structuring the flow of knowledge too tightly
around specific areas of knowledge focus or
collaborative projects (Leonard, 1998). In ICI,
any islands of knowledge occur on a business by
business basis, rather than by specialty or areas of
shared interest.
The fragmented nature of the formally pre-
scribed Expert Groups provides support for
claims in research on knowledge networks that
there exists a lack of managerial understanding of
how best to use informal networks (Cross and
Parker, 2004). The differences between formal
structures and the informal organization are fre-
quently attributed to the fact that although man-
agers are aware of the importance of informal
networks in the production and realization of
knowledge (De Meyer, 1991; Katz and Allen,
2004c; Bryan and Joyce, 2005), they fail to exploit
them effectively as a mechanism for the exchange
of knowledge assets (Cross and Parker, 2004;
Cronin, 2005). The personal relationships and
cultural factors that drive informal working rela-
tions are often seen by managers as being ‘im-
precise and difficult to manage’ (O’Reilly and
Tushman, 1997, p. 200). By consequence, senior
management often show a relatively poor ability
to build effective and supportive formal structures
around informal networks (Cross et al., 2002b) as
Formal versus informal knowledge networks in R&D
r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
R&D Management 37, 3, 2007 191
indeed is apparent in the case of ICI. Viewed from
a strategic perspective, these observations indicate
a significant gap between what companies under-
stand to be the knowledge-sharing networks
within R&D, and how knowledge is informally
transferred in practice.
5.2. Importance to management of
mapping informal networks
This gulf in understanding reinforces a further
significant observation made in previous research
namely that companies can only begin to fully
comprehend the nature of their informal organi-
zation through the comprehensive mapping of
social networks within their organization (Cross
and Parker, 2004; Anklam, 2005). Significant
disparities can be observed in this study between
ICI’s informal problem-solving networks and its
formal structures of knowledge dissemination.
However, the process of conducting a social net-
work analysis can itself be viewed as an important
first step in correcting them. Cross argues that
studies of informal networks are vitally important
in improving managerial understanding of the
true working nature of their organization (Cross
et al., 2001, 2002a, b; Cross and Parker, 2004).
The informal problem-solving network docu-
mented here reveals several aspects of knowledge
transfer and working relationships within ICI’s
R&D function, both broad and specific, that
would otherwise be extremely difficult to recog-
nize. For example, a select few personnel within
the network can be seen to be acting as ‘boundary-
spanning individuals’, as defined by Allen (1977)
and others. These individuals are the few nodes in
the network through which knowledge is trans-
ferred between the Group’s businesses. Of the six
ties between Business D and the other Group
businesses for example (ignoring ties with the
corporate centre), three enter the company
through one member of staff. He or she, like the
other boundary-spanning individuals observed in
the network, is then in turn strongly connected to
members of his or her own business by a large
number of ties.
The difficulty in accurately reflecting informal
relationships in formal organization structures is
reflected by the fragmented problem-solving net-
works revealed within ICI’s Science Ladder and
Expert Groups. The implication of this lack of
collaboration between members of the formal
networking structures is that the prescribed mem-
berships do not reflect the preferred working
relationships of the individuals concerned. Cross
and Parker reflect in their study of a distributed
R&D group that despite attempts to promote
collaboration by the company involved, ‘people
still relied upon those that they knew and trusted,
and not a database of self-proclaimed experts’
(Cross and Parker, 2004, p. 16). Bryan and Joyce
(2005) further observe that ad hoc structural devices,
which the specialist groups used by ICI strongly
resemble, ‘serve only to complicate the organization
further’ (Bryan and Joyce, 2005, p. 23).
6. Implications for management practice
While managers in large companies are often
aware of the ‘informal’ networks in their organi-
zations it is rarely that they consciously seek to
understand and manage them, either running the
formal organizations and informal networks in
parallel, or better still fusing them together. The
distinction between formal and informal net-
works is clearly less important in companies
with few employees. Managers principally have
control of structures, processes and resources
(people and money). Social network analysis
suggests a number of questions where the manger
can take action.
Are the key individuals found in the informal
networks those who would be expected to be in
key positions?
If not, do these individuals understand the
organization’s expectations of their roles, and
do they need coaching to fulfill them better?
Should individuals who have been unexpect-
edly found to be key in the informal networks
be more integrated into the formal networks?
Are there sufficient key nodes/individuals for
the needs of the organization, or are knowledge
flows being constrained by insufficient links?
What level of duplication is required to cope
with the movement/loss of key nodes/indivi-
duals?
Where is the organization particularly vulner-
able to the loss of a key node/individual and is
there adequate succession planning?
Where individuals hold particular positions
e.g. communities of practice, expert networks,
Science Ladder appointments, are they linked
to the appropriate individuals to carry out that
role?
Are key individuals sources of knowledge per
se, or are they brokers of access to sources of
knowledge?
James Allen, Andrew D. James and Phil Gamlen
192 R&D Management 37, 3, 2007 r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
Informal networks are strongest as a result of
face to face contacts, both in the initiation and
in the upkeep. Are key individuals encouraged
and funded to ensure that they are regular
meetings at different physical locations? Do
these meetings include enough time to build a
social community as well as complete the
necessary business?
This particular case study provokes discussion
of many of these issues. But the specific circum-
stances raise additional questions. For example
the ICI organization has arisen from substantial
restructuring following a major programme of
acquisitions and divestments. Conscious action
is required to bring about the integration of
organizations that have shared no networks pre-
viously, formal or informal. Attention is inevita-
bly focused on the formal organization initially
and the informal networks evolve much more
slowly if left to their own devices. The mainte-
nance of a central technology function has un-
doubtedly been both effective and critical in
brokering contacts and knowledge flows. The
retained central function was found to play an
important role in providing a linking mechanism,
knowledge assets and a central capability that
would otherwise be absent were the function
wholly decentralized and R&D exclusively pur-
sued on a business by business basis.
Creating an effective Science Ladder for the
organization is substantially easier and faster if
the appointees can build a sense of community
with their colleagues on the Ladder and build an
identity as a grouping quite distinct to their
individual positions in separate businesses. Again,
management has a clear responsibility to enable
this to happen.
7. Conclusions
The findings of the social network analysis carried
out in ICI, and the comparison of formal and
informal organizational structures that this en-
abled, highlight several issues regarding the orga-
nization of the R&D function for successful
knowledge transfer.
Significantly, the informal problem-solving net-
work within ICI’s R&D function was found to
differ significantly from the formal structures put
in place by the company to manage knowledge
transfer. The implication which may be drawn,
when this case study is considered together with
the existing literature, is that it is difficult for
management teams to fully comprehend and ex-
ploit the nature of informal relationships within
R&D. There is likely to be a significant difference
between what managers presume to be occurring
in terms of knowledge exchange, and what occurs
in practice.
For multi-business firms with a distributed
R&D function, such as investigated in this in-
stance, the informal network suggests that pro-
blem-solving activity between separated R&D
centres and the inherent knowledge sharing which
this requires may be highly limited. Instead, this
study found that technical personnel appear to
collaborate most closely with those in close orga-
nizational and geographical proximity to them,
rather than with colleagues located in other
businesses or regions.
In the case of the ICI study, formally prescrib-
ing the memberships of various collaborative and
knowledge-sharing structures was found to result
in fragmented groups and little collaboration on
problem-solving issues. It is considered likely that
such structures might be more effective and pro-
ductive if their membership were instead advised
by social network analysis studies. Such an ap-
proach would potentially reveal more fruitful
collaborative relationships and areas where
Group wide collaborations may be nurtured and
extended.
The most significant implication to be drawn
from this investigation therefore is that firms and
R&D managers stand to gain significantly by
conducting social network analysis, such as has
been performed in this project, to map the in-
formal social networks within their organizations.
The benefits of performing social network analy-
sis studies are clear. If the true extent and
membership of an informal network is under-
stood, then it may be supported and nurtured to
increase the wider effectiveness and innovatory
capacity of the firm (Cross and Parker, 2004).
By the investigation of informal networks it is
possible to identify critical personnel who may
otherwise go unrecognized. This includes both
technology gatekeepers and boundary-spanning
individuals, but may also include staff who
may be acting as bottlenecks to knowledge
transfer (Anklam, 2005). The impact of the po-
tential departure of personnel can then be more
effectively planned for by management, and
the under-use of valuable human resources
avoided. Studying informal networks can
also reveal structural gaps and thereby provide
insight for successfully implementing formalized
networks.
Formal versus informal knowledge networks in R&D
r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
R&D Management 37, 3, 2007 193
References
Allen, T.J. (1971) Communication networks in R&D
Laboratories. R&D Management,1, 1, 14–21.
Allen, T.J. (1977) Managing the Flow of Technology.
Cambridge, MA: MIT Press.
Allen, T.J., Piepmeier, J.M. and Cooney, S. (1971) The
international technological gatekeeper. Technology
Review,73, 5, 5–10.
Anklam, P. (2005) The Social Network Toolkit: Build-
ing Organisational Performance Through Collabora-
tive Communities. London: Ark Group.
Blacker, F. (1995) Knowledge, knowledge work and
organizations: an overview and interpretation. Orga-
nization Studies,16, 6, 1021–1046.
Borgatti, S.P. (2002) Netdraw: Graph Visualization
Software. Harvard, MA: Analytic Technologies.
Borgatti, S.P., Everett, M.G. and Freeman, L.C. (2002)
Ucinet for Windows: Software for Social Network
Analysis. Harvard, MA: Analytic Technologies.
Brown, J.S. and Duguid, P. (2001) Structure and
spontaneity: knowledge and organization. In Non-
aka, I. and Teece, D.J. (eds), Managing Industrial
Knowledge: Creation, Transfer and Utilization. Lon-
don: Sage, pp. 44–67.
Bryan, L.L. and Joyce, C. (2005) The 21st-century
organization. The McKinsey Quarterly,2005, 3, 21–29.
Burns, T. and Stalker, G. (1961) The Management of
Innovation. London: Tavistock.
Chandler, A. (1962) Strategy and Structure. Cam-
bridge, MA: MIT Press.
Coombs, R. and Hull, R. (1998) Knowledge manage-
ment practices and path-dependency in innovation.
Research Policy,27, 237–253.
Cronin, B. (2005) Business Network Research Group.
Available at: www.gre.ac.uk/schools/business/
research/BNRG/index.htm (accessed July 2005).
Cross, R., Borgatti, S.P. and Parker, A. (2002a) Mak-
ing invisible work visible: using social network ana-
lysis to support strategic collaboration. California
Management Review,44, 2, 25–46.
Cross, R., Nohria, N. and Parker, A. (2002b) Six myths
about informal networks – and how to overcome
them. MIT Sloan Management Review,43, 3, 66–75.
Cross, R. and Parker, A. (2004) The Hidden Power of Social
Networks. Boston: Harvard Business School Press.
Cross, R., Prusak, L., Parker, A. and Borgatti, S.P.
(2001) Knowing what we know: supporting knowl-
edge creation and sharing in social networks. Orga-
nizational Dynamics,30, 2, 100–120.
Daft, R.L. (2004) Organization Theory and Design.
Mason, OH: Thomson/SouthWestern.
Dahl, M. and Pedersen, C. (2005) Social networks in
the R&D process: the case of the wireless communica-
tion industry around Aalborg, Denmark. Journal of
Engineering and Technology Management,22, 75–92.
Davenport, T.H. and Prusak, L. (1998) Working
Knowledge: How Organizations Manage What They
Know. Boston, MA: Harvard Business School Press.
De Meyer, A. (1983) A study of technology transfer in
small belgian manufacturing firms, PhD. Thesis,
University of Ghent, Ghent.
De Meyer, A. (1991) Tech talk: how managers are
stimulating global R&D communication. Sloan
Management Review,32, 3, 49–58.
Gamlen, P., Allan, H., Young, M., Lau, L., Megson, S.
and Mahmoud, A. (2003) Growing Guan Xi: how
the centre ensures businesses get value from global
communities of practice. In 5th Asia Pacific Knowl-
edge Management Conference, Hong Kong, 27–28
November 2003.
Georghiou, L., Giusti, W.L., Cameron, H.M. and
Gibbons, M. (1988) The use of co-nomination ana-
lysis in the evaluation of collaborative research. In
Van Raan, A.F.J. (ed.), Handbook of Quantitative
Studies of Science and Technology. Amsterdam: Pub-
lisher: Elsevier Science Publishers B.V., pp. 275–290.
Hansen, M.T., Nohria, N. and Tierney, T. (1999)
What’s your strategy for managing knowledge?
Harvard Business Review, 106–116.
Hildreth, P., Kimble, C. and Wright, P. (2000) Commu-
nities of practice in the distributed international envir-
onment. Journal of Knowledge Management,4, 1, 27–37.
ICI (2004) The ICI Factbook 2004/05. London: Imper-
ial Chemical Industries PLC.
ICI (2005) Investor relations, our strategy: why a
group? Available at: http://www.ici.com/ICIPLC/ici-ir/
strategy/why_a_group.htm (accessed August 2005).
Katz, R. and Allen, T.J. (2004a) How project perfor-
mance is influenced by the locus of power in the
R&D matrix. In Katz, R. (ed.), The Human Side of
Managing Technological Innovation. Oxford: Oxford
University Press, pp. 198–214.
Katz, R. and Allen, T.J. (2004b) Managing dual
Ladder systems in RD&E settings. In Katz, R.
(ed.), The Human Side of Managing Technological
Innovation. Oxford: Oxford University Press.
Katz, R. and Allen, T.J. (2004c) Organizational issues in
the introduction of new technologies. In Katz, R. (ed.),
The Human Side of Managing Technological Innova-
tion. Oxford: Oxford University Press, pp. 450–477.
Katz, R. and Tushman, M.L. (1979) Communication
patterns, project performance, and task characteris-
tics: an empirical evaluation and integration in an
R&D setting. Organizational Behavior and Human
Performance,23, 139–162.
Krackhardt, D. and Hansen, J.R. (1993) Informal
networks: the company behind the chart. Harvard
Business Review,71, 4, 104–111.
Leonard, D. (1998) Wellsprings of Knowledge: Building
and Sustaining the Sources of Innovation. Boston,
MA: Harvard Business School Press.
McAdam, R. and McCreedy, S. (1999) The process of
knowledge management within organizations: a cri-
tical assessment of both theory and practice. Knowl-
edge and Process Management,6, 2, 101–113.
Metcalfe, J.S. and James, A. (2000) Knowledge and
capabilities: a new view of the firm. In Foss, N. and
James Allen, Andrew D. James and Phil Gamlen
194 R&D Management 37, 3, 2007 r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
Robertson, P. (eds), Resources, Technology and
Strategy: Explorations in the Resource Based Per-
spective. London: Routledge.
Mintzberg, H. (1973) The Structuring of Organizations.
Englewood, NJ: Prentice-Hall.
Morrison, M.J. (2003) Realising value from acquired
technological resources: the challenge of post-acqui-
sition technology management. PhD Thesis, Man-
chester University, Manchester.
Mote, J.E. (2005) R&D ecology: using 2-mode network
analysis to explore complexity in R&D environ-
ments. Journal of Engineering and Technology Man-
agement,22, 93–111.
Nedeva, M., Georghiou, L., Loveridge, D. and Cam-
eron, H. (1996) The use of co-nomination to identify
expert participants for technology foresight. R&D
Management,26, 155–168.
Nonaka, I. and Teece, D.J. (eds) (2001) Managing
Industrial Knowledge: Creation, Transfer and Utiliza-
tion. London: Sage.
O’Reilly, C.A. and Tushman, M.L. (1997) Using cul-
ture for strategic advantage: promoting innovation
through social control. In Tushman, M.L. and An-
derson, P. (eds), Managing Strategic Innovation and
Change: A Collection of Readings. Oxford: Oxford
University Press, pp. 200–216.
Orlikowski, W.J. (2002) Knowing in practice: enacting
a collective capability in distributed organizing. Org-
anization Science,13, 3, 249–273.
Prahalad, C.K. and Hamel, G. (1990) The core compe-
tence of the corporation. Harvard Business Review,
68, 3, 79–91.
Roussel, P., Saad, K. and Erickson, T. (1991) Third
Generation R&D: Managing the Link to Corporate
Strategy. Boston: Harvard Business School Press.
Sanchez, R. (ed.) (2001) Knowledge Management and
Organizational Competence. Oxford: Oxford Univer-
sity Press.
Scott, J. (1991) Social Network Analysis. London: Sage.
Sparrow, J. (1998) Knowledge in Organizations: Access
to Thinking at Work. Sage: London.
Stevenson, R. (2002) ICI’s new model R&D. Chemistry
In Britain, 26–29.
Teece, D.J. (1998) Capturing value from knowledge
assets: the new economy, markets for know-how, and
intangible assets. California Management Review,40,
3, 55–76.
Tichy, N.M. (1981) Networks in organizations. In
Nustrom, P. and Starbuck, W. (eds), Handbook of
Organizational Design, Vol. 2. New York: Oxford
University Press, pp. 225–247.
Tichy, N.M., Tushman, M.L. and Fombrun, C. (1979)
Social network analysis for organizations. Academy
of Management Review,4, 4, 507–519.
Tsoukas, H. (1996) The firm as a distributed knowledge
system: a constructionist approach. Strategic Man-
agement Journal,17, 11–25.
Tushman, M.L. (1978) Technical communication in
R&D laboratories: the impact of project work char-
acteristics. Academy of Management Journal,21,4,
624–645.
Tushman, M.L. and Scanlan, T.J. (1981a) Boundary
spanning individuals: their role in information trans-
fer and their antecedents. Academy of Management
Journal,24, 2, 289–305.
Tushman, M.L. and Scanlan, T.J. (1981b) Character-
istics and external orientations of boundary spanning
individuals. Academy of Management Journal,24,1,
83–98.
Wasserman, S. and Faust, K. (1994) Social Network
Analysis. Cambridge: Cambridge University Press.
Wenger, E. (1998) Communities of Practice: Learning,
Meaning and Identity. Cambridge: Cambridge Uni-
versity Press.
Notes
1. Each respondent was provided with guidance in how
to interpret the terminology of the question. This
took the form:
The term ‘scientific or technological’ is deliber-
ately broad. It may cover your daily work, an
area of research you are involved in or an issue of
special technical interest.
By ‘problem’ is meant any challenging or impor-
tant aspect of your work.
Thus, the ‘problems’ which individuals responded
on were not necessarily of equal significance, but
were all in respect of everyday work.
2. Equally weighted ties were used in order to simplify
the analysis. Instead of weighting ties by frequency
we chose to identify and comment on reciprocity of
relationships as a means of identifying strong ties.
Frequency data could be integrated into a further
analysis of the data.
3. Observation based upon a w
2
test of the number of
ties observed to the centre in relation to the various
business populations surveyed. Details are con-
tained in Appendix A.
Formal versus informal knowledge networks in R&D
r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
R&D Management 37, 3, 2007 195
Appendix A
The w
2
calculation used to investigate ties to the corporate centre by business is detailed below:
Table A1. Chi-squared statistical test
Business Population surveyed Number of ties to centre
B15 4
C45 30
D27 10
E21 6
OpE(¼np)OE(OE)
2
/E
B 4 0.14 7 3 1.29
C 30 0.42 21 9 3.86
D 10 0.25 12.5 2.5 0.50
E 6 0.19 9.5 3.5 1.29
n¼50 1.00 50 0.0 w
2
¼6.94
O/E
B 0.57
C 1.43
D 0.80
E 0.63
w
2
, chi-squared value. A value of 6.94 suggests a significant degree of diversion from the normal values. This is confirmed by the
ratios of observed to expected number of ties shown below: O, observed ties; n, total number of ties; p, proportion of total (O/n);
E, expected number of ties.
James Allen, Andrew D. James and Phil Gamlen
196 R&D Management 37, 3, 2007 r2007 The Authors
Journal compilation r2007 Blackwell Publishing Ltd
