European Management Journal Vol. 21, No. 5, pp.

578–587, 2003
Pergamon  2003 Elsevier Ltd. All rights reserved.
Printed in Great Britain
doi:10.1016/S0263-2373(03)00107-5 0263-2373/$30.00

Management Focus

Conditions for
Knowledge Sharing in
Competitive Alliances
MAURA SOEKIJAD, Delft University of Technology
ERIK ANDRIESSEN, Delft University of Technology

This paper examines the conditions for successful alliances or networks, in order to gain competitive
knowledge sharing and learning in inter- advantage from exchanging and creating (new)
organisational alliances. In order to improve com- knowledge, even with their competitors (Mowery et
petitive advantage, organisations try to learn and al., 1996; Hamel et al., 1989; Hamel, 1991).
develop knowledge in alliances. But how do they
‘succeed’ in these processes? We present several fac- In the situation where competing organisations co-
tors that are important for the ‘successful’ knowl- operate in an alliance, one would suspect a certain
edge sharing and learning in alliances where com- ‘co-opetitive’ strategy, as an organisation must some-
peting organisations co-operate. Whereas much how simultaneously co-operate and compete with
alliance literature seems to have a bias towards con- the same parties (Brandenburger and Nalebuff, 1996).
ditions at the inter-organisational level, we suggest In particular, organisations somehow have to man-
that attention should be paid to inter-personal con- age their knowledge sharing while competing (see
ditions as well. Our two case studies located in The also Harding, 2002). When trying to learn from each
Netherlands present some inter-personal conditions other, organisations can meet what is described as a
for ‘successful’ knowledge sharing and learning ‘learning dilemma’ (Larsson et al., 1998). Based on
in alliances. game theory, Larsson et al. (1998) argue that
 2003 Elsevier Ltd. All rights reserved. (competing) organisations are likely to meet an inter-
organisational learning dilemma when they are part
Keywords: Knowledge sharing, Learning, Co-opeti- of an alliance in which the purpose is to learn from
tion, Inter-organisational alliances and with each other. For each individual organisation
it is ‘rational’ ‘to pursue the maximum organisational
share of the joint learning by taking more than it
gives’ (p. 288). While at the same time, this lack of
Introduction openness towards the other reduces ‘the total amount
of joint learning from which the organisation
In order to develop and maintain a strong competi- attempts to appropriate its share’ (ibid.). Therefore,
tive position it is important for an organisation to the collective knowledge development and mutual
have adequate knowledge available. As Powell (1998) learning in the competitive alliance are limited by the
states: ‘the core capabilities of organisations are (natural) opportunistic behaviour of organisations. In
based increasingly on knowledge-seeking and turn, the co-operation as such can even be destroyed.
knowledge-creation’ (p. 228). Additionally, focusing Larsson et al. (1998) give many arguments why such
on their core competencies, organisations seem to a learning dilemma is likely to occur in the situation
have grown more dependent on each other. In parti- where competing organisations co-operate and in
cular, they are dependent on each other’s knowledge particular try to learn and share knowledge together.
and capabilities (Khanna et al., 1998). Because organ-
isations are more knowledge intensive and depen- Thus, when competing organisations co-operate in an
dent on each other, an important way to survive is alliance, it seems likely that learning might become
through co-operation (e.g. Powell, 1990; Inkpen, 1998; a difficult process. Our research aims to provide
Gulati, 1999). Organisations therefore participate in insight into whether learning and knowledge sharing

578 European Management Journal Vol. 21, No. 5, pp. 578–587, October 2003
 CONDITIONS FOR KNOWLEDGE SHARING IN COMPETITIVE ALLIANCES

indeed take place in competitive alliances. More and inter-organisational (alliance) learning,
importantly, we tackle the question: what are the especially from a strategic perspective. We dis-
conditions for learning in this specific context? There- tinguish three main streams of literature on learning
fore, this paper identifies factors that are important and knowledge sharing in alliances. Firstly, organis-
for successful knowledge sharing in alliances of com- ations can learn IN alliances how to improve their
peting organisations. Success is implied when indi- operations, strategy, competencies, skills or capabili-
viduals, groups or organisations have indeed learned ties (Huber, 1991; Mowery et al., 1996; Inkpen and
in the alliance1. We will continue in the next section Dinur, 1998). This can be achieved through accessi-
with our view on learning and knowledge sharing. bility, but more importantly through the actual
Additionally, we use alliance literature to illustrate internalisation of a partner’s skills or capabilities
what can be learned in alliances. In the subsequent (Hamel, 1991). Secondly, a network of organisations
section we present an overview of conditions for can learn AS alliances. This stream of literature
learning that are mentioned in the alliance literature, focuses on the inter-organisational level of learning.
after which we will complement these insights with It is believed that alliances as a whole can benefit and
results from our case studies. learn from organisational participation and processes
(Lane and Lubatkin, 1998; Kraatz, 1998; Levinson and
Asahi, 1995; Khanna et al., 1998). Finally, a third part
of literature can be distinguished in those who have
Learning and Knowledge Sharing in found that organisations can learn ABOUT alliances
Alliances (Gulati et al., 2000; Teece, 1994; Eisenhardt and Mar-
tin, 2000). Powell (1998) for instance states that
(successful) participation in, and management of
We focus particularly on alliances that have learning alliances are ‘key drivers of a new logic of organising’
and knowledge sharing as an important objective, (p. 231). In this stream, organisations might want to
and in which potential competitors co-operate. This acquire knowledge that can be used for the manage-
section will explain respectively how we see ment of alliances in general. It can also help them
alliances, competition and learning and knowledge improve the design of new future alliances.
processes in alliances.
The three streams of literature provide different
In this paper we will refer to an alliance when indicat- (motivational) reasons for organisations to partici-
ing a network that two or more (autonomous) organ- pate in alliances, such as to learn alliancing skills
isations have formally agreed upon to co-operate in, (Kale et al., 2001). In order to understand whether
to achieve a common objective. We focus particularly learning of these and other skills takes place
on those alliances in which (potential) competing (successfully), we need to consider the conditions
organisations co-operate, instead of alliances that for it.
consist of suppliers or customers (Huang et al., 2002).
We consider organisations competitors when they
produce and market products or services in the same
market. These products or services can vary,
although they are often the same ‘type’ of products
Conditions for Learning and
or services2. Knowledge Processes in Alliances

In the context of these competitive alliances, we In this section we cluster conditions for learning and
specifically look at learning and knowledge sharing. knowledge sharing in alliances as mentioned in the
In general, knowledge sharing can often be con- literature on alliances. The first range of conditions
sidered as a first and necessary step in a learning pro- can be described as organisational characteristics.
cess. People can share knowledge and experiences, Each organisation in an alliance must expect to
which they later can also apply in their own situation. receive a certain added value from knowledge shar-
Knowledge can be transferred, distributed or newly ing (high pay off) (Axelrod, 1984; Gulati et al., 2000;
created3. We use the expression of knowledge shar- Inkpen and Dinur, 1998) and in general must be wil-
ing for all of these three processes, while learning in ling and able to share knowledge (Larsson et al., 1998;
our view occurs particularly during the application Inkpen and Crossan, 1995; Doz, 1996). In this view
and use of knowledge. We consider learning as an an organisation must be motivated and have a high
activity that concerns an entity, which can be an indi- intent towards co-operation and learning. Besides
vidual, group, organisation, or even an industry or that, the organisation must be able to communicate
society as a whole (Huber, 1991). We leave room for and it needs to be highly transparent and receptive
a multi-level view. We are interested in both knowl- (Hamel, 1991). Organisational capabilities to share
edge sharing and learning, and address these pro- knowledge and learn can also be explained through
cesses simultaneously. the concepts of learning effectiveness (Inkpen and
Dinur, 1998) or (relative) absorptive capacity (Cohen
In general, much research on knowledge sharing and and Levinthal, 1990; Lane and Lubatkin, 1998; Lane
learning in alliances has focused on organisational et al., 2001). A final organisational characteristic that

European Management Journal Vol. 21, No. 5, pp. 578–587, October 2003 579
CONDITIONS FOR KNOWLEDGE SHARING IN COMPETITIVE ALLIANCES

influences learning concerns an organisation’s ability sary link between the individual and organisational
to have access to knowledge (Powell, 1998; Inkpen level of knowledge sharing and learning. Lofstrom
and Dinur, 1998). Access can be gained through ties (2000) suggests that future research on alliances
or personal networks of key individuals for instance would benefit from a multi-level approach. She
(McEvily and Zaheer, 1999; Lofstrom, 2000). explains: ‘adding the micro-level factors of individ-
uals’ networks and knowledge increases our under-
The second cluster of conditions consists of character- standing of what leads to learning in alliances’ (p.
istics of the (mutual) relationship between the organ- 22). Inkpen and Crossan (1995, p. 598) make the same
isations involved. They for instance include the argument when they suggest that the different levels
strength of the tie or relation (Hansen, 1999) and rec- of learning (individual, group and organisational)
ognition and trust between the organisations should be embedded in another level. At each level
(Larsson et al., 1998). Further, the closer an activity is of learning, different learning processes are at work.
to a client, the less co-operative an organisation will Inkpen and Dinur (1998) additionally propose that if
be (Bengtsson and Kock, 2000; Zander and Kogut, the tacitness of transferred knowledge is high, such
1995). Other conditions include positive former as in firm-specific technological capabilities (Mowery
experiences, either with co-operation in alliances in et al., 1996), it is more likely that individuals are the
general, or with the specific partner, and an un- primary knowledge transfer agents. Thus, it seems
limited ‘shadow of the future’4 (Axelrod, 1984; Lars- necessary to study knowledge processes in the con-
son et al., 1998; Inkpen and Dinur, 1998; Kogut and text of co-opetitive organisations in alliances not only
Zander, 1992). Further, a favourable climate (Inkpen, from an inter-organisational, but also particularly
1998) positively influences knowledge sharing as from an inter-personal perspective. Our study looks
well. into this inter-personal level of co-operation and
interaction in the context of competitive alliances, in
A final cluster of conditions can be found in the order to see what conditions are important for learn-
characteristics of the knowledge shared. Tacit knowl- ing and knowledge processes. In the remainder of
edge is less easy to exchange (Larsson et al., 1998), this paper we will present results from two case stud-
and the more codifiable and teachable knowledge is, ies. The next section will first describe our research
the easier it is to transfer it (Zander and Kogut, 1995). method and approach.
All of the conditions mentioned above indicate how
organisations can successfully learn and share knowl-
edge.
Research Method and Approach
The alliance literature described above seems to be
characterised by a tendency to view and analyse Our empirical insights are based on two case studies
learning at the (inter-) organisational level. An organ- in knowledge-sharing groups within the alliance con-
isation or alliance can develop various kinds of text (respectively KITe and Delta). Both alliances con-
knowledge, capabilities and skills, and several con- sist of several (potentially) competing organisations
ditions are presented that influence these learning and consider knowledge sharing as one of their major
processes. These conditions mainly concern the goals. The data were collected through interviews
organisational or alliance level as well. However, with several participants in the knowledge-sharing
because individuals are the actual knowledge work- groups and through analysis of documents, com-
ers, the settings where most knowledge processes are bined with observations5. The cases were selected in
likely to take place are inter-personal groups such as order to provide illustrative descriptions and develop
(multidisciplinary) project teams, job-groups, net- insights into the analysis of learning in competitive
works of expertise, design groups, or communities of alliances. They particularly generate insights into the
practice (CoPs). Here, at a more interactive level, conditions that could be found within a single setting
people from the different organisations actually meet of KITe and Delta (see also Eisenhardt, 1989). The
each other and can share knowledge. CoPs and net- study is subject to the general limitations associated
works of expertise in particular are centred on with this type of field research, particularly the con-
knowledge sharing and learning and can therefore be straints on generalising from such a small number
considered as knowledge-sharing groups. In these of cases.
groups people tend to learn the essentials of their
daily work by participating in them, or develop new
knowledge in their practice. Here, particularly tacit Data Collection
knowledge is likely to be transferred or developed
(Inkpen and Crossan, 1995; Inkpen and Dinur, 1998). KITe was studied over a period of four months. Data
Therefore we propose to look specifically at these collection in this case consisted of interviews with
knowledge-sharing groups when trying to find con- eight representatives of different types of organis-
ditions for (successful) learning and knowledge pro- ations (varying in size and involvement), including
cesses in alliances. the facilitating organisation. Delta was approached in
a more longitudinal way, stretching out over a period
Knowledge-sharing groups might provide a neces- of 18 months in total. Here we did six face-to-face,

580 European Management Journal Vol. 21, No. 5, pp. 578–587, October 2003
 CONDITIONS FOR KNOWLEDGE SHARING IN COMPETITIVE ALLIANCES

and six telephone interviews in a first phase, and 18 ations had knowledge management as their core
interviews in a later phase. In Delta we also inter- business, although their actual daily business dif-
viewed (five6) people that were not directly involved fered from training or consultancy to the selling of
in the knowledge-sharing group itself, but who could products or services for knowledge management.
provide additional information on how groups in They all had their own particular view on and inter-
Delta functioned. est in knowledge management. The participating
organisations varied in their motivation for partici-
In both cases all of the interviews were semi- pation and also in size. Some organisations consisted
structured, taking approximately two hours. They of only one employee, while others had 30–50
were held in combination with a small structured employees, complemented with a few large multi-
questionnaire. Both interview and questionnaire nationals, who had more than 5000 employees world-
included at least the subjects: the co-operation pro- wide.
cess, learning dilemma, knowledge, trust, learning
structures, and supporting conditions. Other sources Organisations — contrary to individuals — could
of data in both cases included documents7, such as become KITe members by paying contributions.
(financial) plans, ideas, reports, minutes of meetings, However, not all participating organisations were
figures, contracts, correspondence, and policy docu- equally involved and active in KITe. Some (larger)
ments. And, as a third data source we used obser- organisations were represented by a sales manager,
vation. Meetings and other plenary activities were whereas others by a top-level manager. The most
visited in order to see for instance patterns of behav- active people appeared to be those from medium size
iour. This was also a reason for performing the inter- organisations who had a considerable power of
views at the sites of the people involved, enabling us decision. Not all organisations therefore benefited
to take a look behind the scenes. All of the contacts equally from their KITe membership.
and remarkable facts were noted in a small log file
or diary by the researcher, which formed an As a platform KITe was a clearly identifiable partner
important point of reference in the analysis. for knowledge management products or services.
This means that clients could ‘meet’ various KM-
specialists through the ‘portal’ of KITe. The alliance
of KITe also aimed at encouraging research and con-
Data Analysis
sultancy in the area of knowledge infrastructures and
In both cases we used the concepts (language) of the knowledge management in their own companies and
parties involved in the interviews, which were later in other Dutch organisations. Co-operation took
coded into our own structure. All of the interviews place in a knowledge-sharing group in which one to
were taped and transcribed, and were used to find four individuals from each organisation participated.
global patterns, concerning the themes of the co- This group regularly held meetings to discuss pro-
operation process, learning dilemma, knowledge, cess and progress matters. The other activities in the
trust, learning structures, and supporting conditions. group consisted of a few meetings where one partici-
The findings were also discussed with the respon- pant presented himself, the organisation and hosting
sible facilitators of the knowledge-sharing group in of a conference, they published a book, and they had
the end. a small number of orienting sub-group meetings with
potential customers.

Knowledge Sharing Processes

Results from Case Studies People in KITe mainly appeared to share knowledge
for individual learning purposes. Each participant
could use the knowledge for personal goals and
This section presents insights from the two case stud- development, such as their personal career. Most
ies of KITe and Delta. Both examples start with an respondents agreed that the co-operation had not
introduction that is followed by insights in the (yet) reached the stage of real co-operative learning
knowledge-sharing processes, and conclude with for the organisations involved. Even knowledge shar-
conditions for learning and knowledge processes. ing between individuals seemed to occur less fre-
quently than anticipated and expected. People
Case Study 1: KITe merely exchanged experiences and wanted to assess
their own ideas by comparing them with others.
Introduction Thus, learning at a different level than the individual
KITe, as an alliance on Knowledge Infrastructures did not occur much. Reasons for this lack of learning
and Technology, was founded in 1997 as a Dutch can be found in the fact that parties had little idea of
knowledge platform in which participants aimed to what the other party knew and what they could learn
share experiences and do business together. This from each other. Therefore, many parties tended to
alliance was organised as a formal foundation and turn to their already known ‘own’ contacts, in their
consisted of 12 organisations at the time of the study, own existing networks when co-operation was
which was in 1999. All of the participating organis- necessary. Other reasons were that the goals for par-

European Management Journal Vol. 21, No. 5, pp. 578–587, October 2003 581
CONDITIONS FOR KNOWLEDGE SHARING IN COMPETITIVE ALLIANCES

ticipation in the alliance varied between the ation, learning never reached the organisational level,
organisations. This caused a lack of clarity about as colleagues were often not aware of KITe partici-
what people could expect from the knowledge-shar- pation and/or activities. Thus, without organisation-
ing group and what was expected from them. For wide commitment (organisational) learning appeared
instance, when KITe was organising a conference, it to be less likely.
appeared that for some organisations commercial
interests were more important than learning goals. Inter-personal Trust and Group Cohesion. Most
These organisations mainly wanted to use the confer- respondents in KITe felt that participation in the
ence as a ‘notice board’ (as a marketing tool). knowledge-sharing group was and could only be vol-
untary for individuals. If too many regulations were
The fact that partner organisations in KITe could be necessary, the group, and thus in this case the
considered as (potential) competitors seemed to alliance as a whole, was doomed to fail. One person
influence what subjects could said that certain dynamics had
be discussed. Certain subjects %at the inter-personal to be developed among the
were excluded from discussion, people involved, in the sense
such as information about that a bond could grow. It was
(potential) clients, opport-
level people decided for further said that this ‘bonding’
unities for new projects or
methodological issues (the themselves who to trust in (cohesion) could only be based
on personal goodwill. Trust for
development of instruments, example, could only be based
e.g.). One director of a small sharing their knowledge on the person, not on his/her
organisation said its method- organisation. One respondent
ology was strategically very important for his organ- illustrated an incident of decreased trust: ‘Although
isation. Therefore it could not be shared with others agreements were made to do something together, one
in KITe. He added that it had taken much time to person went behind our backs to that potential client.
develop, and he did not want to help others saving This diminished our trust in him/her’. This caused
time, neither to reassess nor to discuss the method- the people involved to further ignore the untrust-
ology again, for that would take even more time. worthy person as much as possible. And indeed, no
Another respondent referred to the (strategic) impor- further knowledge was shared with that person.
tance of a large client: ‘If this is our most important Thus, at the inter-personal level people decided for
client, I do not feel the need to introduce others to it themselves who to trust in sharing their knowledge.
(…) this could be a large threat to us’. Representa-
tives, especially from the smaller organisations, were Learning by Doing. Attending meetings, especially
worried about sharing knowledge, because their those in which presentations were held about a meth-
organisations were ‘niche-players’ compared to the odology or instrument sometimes resulted in indi-
larger, less specialised organisations. They were how- vidual learning. However, respondents stated they
ever highly interested in learning from others (also could learn most from the other participants during
in combination of benchmarking). co-operative activities, such as organising a confer-
ence or publishing a book. Then discussions became
Conditions for Learning and Knowledge Processes concrete and people needed to know and understand
Although participants stated that they learned less what the other meant. Thus, learning could most eas-
than anticipated considering the overall goal of KITe, ily be achieved by co-operative doing, which is by
they did learn about how other organisations in the definition at the inter-personal level.
alliance ‘dealt with knowledge management issues’,
for instance. Several factors were mentioned and are
listed below, which respondents considered being of Case Study 2: Delta
utmost importance to the success of these learning
and knowledge processes in KITe. Introduction
Delta is a consortium founded by five Dutch know-
Degree of Power of Decision. KITe was based on ledge institutes, which focuses on research into
voluntary participation without formal regulations or prospects for sustainable development of densely
agreements on intellectual property, revenues or populated river-delta areas. This alliance, in which
leads to potential projects. Therefore, the organis- consultancy and engineering firms, and govern-
ations differed in their degree of participation. The mental organisations also participate, was formally
people who were most active in the group had a for- established in 1999. It is partly sponsored by the
mal commitment within their organisations for par- Dutch government for a certain period of time. All
ticipation and spending time on KITe: their power of of the participants come from the sector of Civil and
decision was considerable. In these organisations Hydraulic Engineering, as well as the topics dis-
more people than the participating individuals in the cussed. The sector, as well as the alliance consists of
knowledge-sharing network alone were aware of investors and managers of both public and private
organisational participation in KITe. In cases where infrastructures, and organisations in construction and
there was a lack of commitment from the organis- engineering. Delta is structured around seven

582 European Management Journal Vol. 21, No. 5, pp. 578–587, October 2003
 CONDITIONS FOR KNOWLEDGE SHARING IN COMPETITIVE ALLIANCES

‘themes’ (content-related topics of expertise in Delta, ders or quotations. The effect of this competition,
such as the Coast and River Theme). Each theme has both in the commercial and the academic setting, is
defined several projects in which interested sector that participants in Delta co-operate in pre-
organisations can participate. The projects contribute competitive contexts, and do not discuss or even
to the overall Delta goal to strengthen its knowledge mention competitive activities. One academic said,
and position in the field of sustainable river-delta they would never discuss the proposals they have
development. Delta has the ambition of developing submitted or won (with a ‘losing party’).
into an internationally-renowned knowledge centre
in the sector of Civil and Hydraulic Engineering. In There are differences between the academic and com-
this paper we discuss a knowledge-sharing group in mercial worlds, each considering different types of
Delta consisting of researchers and engineers in the competitors and each having different types of topics
practice of hydrology. The members are spread and ideas that can be shared. In general, many
among three themes and are highly expert people in respondents considered only very few organisations
their field. as their real competitors in the alliance.

Knowledge Sharing Processes Conditions for Learning and Knowledge Processes

Many people believed they learned in the context of In this sub-section, we list the conditions that people
Delta. However, it should be mentioned that most of in Delta mentioned which influenced their learning
them are experts and it is their job to develop knowl- and knowledge processes.
edge (e.g. researchers). The type of activities also
mostly consisted of knowledge development projects. The Need for Co-operation. A clear necessity for co-
Where these activities were part of an organisation’s operation appears to influence the success of the co-
core business, it can be said that with the transfer of operation positively. This necessity can be caused by
the results to the organisation, the organisation a high complexity of the ‘problem’ or issue8, or
learned. However, most people stated that the les- because a third party (such as a client) asks the par-
sons learned were often merely applicable at the indi- ties to co-operate. The latter was illustrated by a
vidual level. Often, new knowledge was considered respondent who said: ‘If you are certain to get the
inextricably bound up with a particular individual. job by co-operating, you will co-operate’. But, many
respondents agreed that very large and monopolistic
A remarkable issue in Delta was the fact that large partners are less dependent on a small party (niche
differences in goals and degree of competition player) than the other way around.
existed between the academic world and the more
commercial world. In the latter, the organisations and Shared (Organisational) Interests. Apart from com-
institutes were more focused on competition and sur- parable knowledge levels, parties involved should
vival that stems from products, services and have shared interests, or as some respondents said:
methods. One issue that a com- shared goals. It can involve the
mercially-oriented respondent feeling of ‘us against a third
would never share with
Often, new knowledge was party’, or together for another
his/her ‘competitive party, e.g. a client. In this latter
colleagues’ was vision, i.e. how considered inextricably bound case, the interest of the client is
he/she saw market develop- primary. A respondent said
ments and his/her plans. up with a particular that knowledge could be
He/she would not show their shared only ‘as long as there is
cards, except in a situation in individual a mutual commercial goal that
which the parties had a third, you agree on… If it serves, for
mutual interest (usually a commercial goal). As it instance, a coalition for a project or business that you
was stated: ‘I would not disclose my plans, or show share’. If there are no common interests, the interest
how I assessed a certain development in the market’. of one party should at least not harm the interest of
the other. This might be enhanced by the fact that
In the same way as commercial competitors try to many people in Delta worked for more than one
enhance their competitive advantage through co- employer. It can cause people to be creative with
operation, academics need each other as well. An aca- organisations’ interests, as one person already unifies
demic said: ‘A dredger who has developed a new several of these interests. Thus, conflicting interests
method for dredging must try even harder to protect between organisations can be a barrier for co-oper-
it, because it is competitive advantage. Our competi- ation and knowledge sharing.
tive position will be particularly strengthened when
we have as much up-to-date knowledge as possible, Inter-personal Trust and Group Cohesion. ‘You
and the fastest way to have that is through co- don’t trust an organisation. You only trust people’.
operation’. Nevertheless, competition also exists in Many people in Delta, as in KITe, stated that if it did
the academic world. Here the various institutions are not work at a personal level, it would not work at
fighting for the same assignments and money, by all. Co-operation can be accomplished when people
submitting competitive proposals, international ten- have social activities together, and when it includes

European Management Journal Vol. 21, No. 5, pp. 578–587, October 2003 583
CONDITIONS FOR KNOWLEDGE SHARING IN COMPETITIVE ALLIANCES

affective trust, i.e. whether people like each other. expertise in a certain knowledge area a personal
Some respondents said that co-operation only (often tacit) form of knowledge of which only they
worked if personalities (or characters) matched and could appreciate its value. Their personal reputation
if there was sufficient mutual respect (for instance, made them feel at ease sharing their knowledge.
with regard to the competency of a person). Then, They even felt ‘obliged’ and proud to do so, as it con-
knowledge sharing can occur almost automatically. tributed even more to their reputation. In that sense
they valued their reputation higher than merely pro-
Comparable as Peers: Recognition. Several respon- tecting it.
dents interviewed said they were (more) willing to
co-operate and share knowledge with peers or
specialists in a certain, very small area. One reason
mentioned for this is shared language or understand- Conclusions from the Two Case Studies
ing: ‘The thing that binds us with them is that they
are complete hydrology-freaks’. They feel a certain Both cases discussed above are examples of knowl-
level of recognition and they are happy to share their edge-sharing groups in competitive alliances where
thoughts and ideas with other specialists. And knowledge sharing is an important goal. The group
because they have a comparable level of knowledge in KITe consists only of employees from commercial
in the same practice, this can even create a certain organisations. This is different from the group in
level of energy to improve each other. As a respon- Delta, where co-operation took place with academics.
dent explained: ‘Because you are “co-opetitors” you In general, KITe was much smaller than Delta and
want to show that you are at least as good. So getting consisted only of one knowledge-sharing group. In
together encourages us to try harder’. It also implies Delta several groups existed, of which we discussed
that parties need to have a similar level of knowl- only one. In comparison to Delta, the KITe group
edge. If one party is consistently superior in this showed less mutual trust, less active participation
respect, while the other can never reach that level, and interaction of individuals, and less insight in the
they are not considered peers. An additional issue other parties (concerning expectations and
that respondents in Delta mentioned was that people competencies) both at the organisational and per-
are required to have high integrity standards and sonal level. In KITe, not every organisation appeared
must be loyal, since it is only a small world of peers to aim at learning evenly, there existed large differ-
and specialists (or highly expert people), who always ences in knowledge level and goals for participation,
seem to be able to find each other. There is a high and learning mainly occurred individually in the
level of social control in such groups encouraging KITe group. Learning needs to be transferred from
‘correct’ behaviour. the individual level to the organisational level as
well, in order to be fully successful (Huysman & de
Diversity in Personal Skills. Peer comparability and Wit, 2000).
shared interest does not automatically imply simi-
larity. Several respondents felt the group should be The conditions found in this section appear to be
composed of people who complement each other. both determinants of group formation as well as of
Individuals need to be different and should add the success of the knowledge sharing. In the two
value to each other. Some respondents referred to dif- alliances studied at least two success conditions fit
ferences in personal character as useful, while others the clusters of factors mentioned in the third section,
preferred a different knowledge area, experience and i.e. characteristic of the organisation, of the relation
personal network. Some respondents stated that between the organisations, and of the knowledge
people find each other because of their shared back- itself. The need for co-operation because of com-
ground, humour or other interests in the first place. plexity of the problems is an illustration of a con-
When they have a shared identity or understanding dition related to the characteristic of knowledge.
they can easily form a group, but in order to stay Whereas the condition that a third party makes
in a group in which learning occurs, inter-personal organisational co-operation necessary, is related to
diversity proved to be a requirement as well. This the relationship between the organisations. And, the
pattern was also evident in the development of the need for a common or shared interest for each organ-
knowledge-sharing groups as a whole. isation can be related to the organisational character-
istics, as well as their mutual relation.
Personal Reputation. In Delta, many specialists are
invited to participate (by clients) for their personal Further, the additional success conditions that were
reputation and knowledge level. Most of the respon- found in the studies appeared to address the inter-
dents stated they were not afraid openly to discuss personal, rather than the inter-organisational level.
their knowledge, because they had invested many Some conditions can be related to characteristics of
years in building that particular knowledge. They an individual, such as: personal reputation, degree of
believed it could be transferred easily. However, this power of decision, and the language and skills (and
was only true when the demand for these specialists reputation) of a person. The other conditions in parti-
was larger than supply. Many respondents — in cular refer to the inter-personal relation of the parti-
particular academics — considered their personal cipants in the group:

584 European Management Journal Vol. 21, No. 5, pp. 578–587, October 2003
 CONDITIONS FOR KNOWLEDGE SHARING IN COMPETITIVE ALLIANCES

❖ Inter-personal trust and group cohesion, not be shared in knowledge-sharing groups in com-
❖ Active interaction or learning by doing, petitive alliances (see Table 2). Most topics that were
❖ Comparable knowledge levels (as peers), not shared in our cases were perceived to be of stra-
❖ Diversity in skills and expertise. tegic importance, were too expensive or of little bene-
fit.

We propose a two-level model for the analysis and

Concluding Remarks and Discussion management of knowledge sharing in alliances,
combining the inter-organisational level and inter-
If one of the most important reasons for organisations personal level. Various conditions at the inter-
to co-operate in alliances is the acquisition, sharing organisational level may result in the establishment
and development of knowledge, the success of an of an alliance between organisations and in the
alliance can also be valued in these terms. So, if much decision to co-operate in certain groups. Once organ-
learning and knowledge sharing occurs, an alliance isations have formed these groups, major success fac-
might be considered successful. In this paper we tors have to be found at the inter-personal level, i.e.
aimed to find conditions for learning and knowledge in characteristics of the individuals, of the group, and
sharing in the particular context of a competitive of the interaction process. When interaction in such
alliance. Although our insights are only based on a group breaks down, the inter-organisational level
material from two cases, we do believe they are may ‘take over’ and change the composition or rules
worthwhile for alliance literature and practice. for the groups (or stop co-operation at all). Or, the
Whereas the existing literature on alliances provides inter-personal conditions can first have their
insight into reasons for organisational participation determining influence on the success of the knowl-
in alliances and mainly addresses conditions at the edge sharing. In this case, the success of the groups
inter-organisational level, we have shown, from case may initiate decision-making at the organisational
studies in two alliances in The Netherlands, that level to institutionalise or authorise the functioning
inter-personal level conditions are of great impor- of these groups.
tance as well. Table 1 provides an overview of every
cluster of conditions that is discussed in this paper, The cases show that two trajectories for co-operation
linked to the level concerned. For each cluster, some in alliances can be found. In the first i.e. a top-down
examples from the text are provided. trajectory, learning in alliances is considered an
important strategic factor to improve the competitive
The conditions derived from our case studies showed advantage of an organisation. This causes top
the inter-personal level has its own dynamic on management in an organisation to ‘construct’ certain
whether or not to share knowledge. The result is sup- settings for co-operation. However, these do not
ported by work by Von Hippel (1987) and Kreiner automatically guarantee successful learning and
and Schultz (1993). The latter found that informal knowledge processes in the alliance. Success depends
contacts, such as between friends, can overrule the on the inter-personal dynamics as well. On the other
power of competition between organisations. Across hand, a bottom up trajectory can be seen in ‘spon-
formal structures of competitive organisations, these taneous’ co-operation among individuals who feel
contacts appear to share knowledge. Von Hippel also the need for knowledge sharing with people from
showed that people at rival organisations, where they other, competing organisations. They can do this in
are in contact, trade their know-how. They mainly do knowledge-sharing groups. In this case, it might
this for personal motives. When an executive thinks become necessary to obtain commitment from the
sharing certain knowledge can be valuable for organisations. These two trajectories (or elements of
him/her, now or in the future, it will be shared. them) can respectively be found in the knowledge-
sharing groups of KITe (top down alliance) and Delta
The cases also provide several suggestions on what (bottom up alliance). The following implications are
type of information is relatively easy to share or can- important for both trajectories.

Table 1 Overview of Conditions for Learning in Alliances

Level Clusters of Conditions Examples of Conditions

Inter-organisational Alliance Relation between organisations Strength of the relationship, mutual trust, need for
co-operation
Organisation Organisational characteristics Motivation, (relative) absorptive capacity,
accessibility
Inter-personal Group Relation between individuals Inter-personal trust and group cohesion, group
composition (comparable peers, diversity in skills
and expertise)
Individual Individual characteristics Degree of power of decision, language and skills
of a person

European Management Journal Vol. 21, No. 5, pp. 578–587, October 2003 585
CONDITIONS FOR KNOWLEDGE SHARING IN COMPETITIVE ALLIANCES

Table 2 Overview of Learning and Knowledge Sharing in Alliances

Information Not Shared Information Shared

Plans and market developments Experiences were exchanged (interpersonal level), on:
— processes (e.g. co-operation, learning ABOUT alliances)
— expertise (e.g. hydrological theories)
— projects that were finished
Proposals etc. for research (e.g., academics) Insights into the other organisation (e.g. their KM products)
New projects (or prospects) Information on publications about a certain knowledge domain
Knowledge of strategically important clients
Models, methods or instruments (e.g. dredgers)

Implications Notes

Whereas Larsson et al. (1998) provided many reasons 1. According to their own perception at least.
for an inter-organisational learning dilemma to occur 2. In the Delta case study presented in this paper for example,
all of the services are related to hydrological research and
in competing alliances, we illustrated different every organisation in the alliance delivers knowledge, tech-
dynamics at the inter-personal level. In our cases we niques, methods or instruments in the area of hydrolog-
found evidence that organisations use several stra- ical research.
tegies to solve this dilemma. Often, respondents did 3. Nonaka and Takeuchi (1995) contrary to us, presume that
not consider the others with whom they co-operated learning will automatically lead to knowledge creation.
4. Repetitive interaction that is expected in future is said to
(at various levels) as ‘real’ competitors. Bengtsson enhance a co-operative attitude (Axelrod, 1984).
and Kock (2000) found a similar inter-personal effect 5. M. Soekijad performed interviews, as well as observations.
in their case studies on co-opetition in business net- 6. It included five formal interviews in the first phase, but
works. Their work suggests individuals might find it many additional conversations (less formal, and not taped
and transcribed) throughout the process.
hard to simultaneously co-operate and compete with 7. This material was abundantly present. We used all docu-
each other, although organisations as a whole can ments that could be found, from the start of the alliance to
benefit. They conclude individuals and groups are the time of research.
better off specialising either in competition or in co- 8. The other party has complementary knowledge you can
operation in a particular relationship. This might never develop yourself and which needs co-operation to
develop.
imply that individuals perceive a learning dilemma
to be less ‘problematic’ than it appears at the inter-
organisational level, because they have their own
strategies for dealing with it. An issue for further References
research is whether an inter-organisational learning
dilemma does occur and how it is perceived at vari- Axelrod, R. (1984) The Evolution of Co-operation. Basic Books,
ous levels. Another issue is that although our paper New York.
Bengtsson, M. and Kock, S. (2000) ‘Coopetition’ in business
explains how learning within groups can occur suc- networks — to cooperate and compete simultaneously.
cessfully, we have not explained why some con- Industrial Marketing Management 29, 411–426.
ditions in some situations are actually successful. Brandenburger, A.M. and Nalebuff, B.J. (1996) Co-opetition.
Currency-Doubleday, New York.
Cohen, W.M. and Levinthal, D.A. (1990) Absorptive capacity:
In a managerial sense, it is important for co-operation a new perspective on learning and innovation. Adminis-
and knowledge sharing in particular, to take place in trative Science Quarterly 35, 128–152.
groups where there is a possibility of group cohesion, Doz, Y.L. (1996) The evolution of co-operation in strategic
active interaction, comparable knowledge levels, and alliances: initial conditions or learning processes? Stra-
diversity in skills and expertise. In order to enable tegic Management Journal 17, 55–83.
Eisenhardt, K.M. (1989) Building theories from case study
learning, group members need to be aware what the research. Academy of Management Review 14(4), 532–550.
others do, expect, and how others can be of value to Eisenhardt, K.M. and Martin, J.A. (2000) Dynamic capabilities:
them. Other conditions that should be managed in what are they? Strategic Management Journal 21, 1105–
order to make knowledge sharing more likely are 1121.
mentioned in Table 1. However, not all knowledge Gulati, R. (1999) Network location and learning: the influence
of network resources and firm capabilities on alliance
will automatically be shared, as Table 2 shows. formation. Strategic Management Journal 20, 397–420.
Gulati, R., Nohria, N. and Zaheer, A. (2000) Strategic net-
works. Strategic Management Journal 21, 203–215.
Hamel, G. (1991) Competition for competence and interpart-
Acknowledgements ner learning within international strategic alliances. Stra-
tegic Management Journal 12, 83–103.
This publication is part of the Commshare project, which is Hamel, G., Doz, Y. and Prahalad, C.K. (1989) Collaborate with
co-sponsored by the Telematica Instituut, The Netherlands. your competitors and win. Harvard Business Review 67,
We would like to thank participants in the case studies and 133–139.
reviewers of this paper for their comments. Hansen, M.T. (1999) The search-transfer problem: the role of

586 European Management Journal Vol. 21, No. 5, pp. 578–587, October 2003
 CONDITIONS FOR KNOWLEDGE SHARING IN COMPETITIVE ALLIANCES

weak ties in sharing knowledge across organization sub- Lane, P.J. and Lubatkin, M. (1998) Relative absorptive
units. Administrative Science Quarterly 44, 82–111. capacity and interorganizational learning. Strategic Man-
Harding, R. (2002) Competition and collaboration in German agement Journal 19(5), 461–478.
technology transfer. European Management Journal 20(5), Lane, Salk and Lyles (2001) Absorptive capacity, learning, and
470–485. performance in international joint ventures. Strategic
Huang, J. C., Newell, S. and Galliers R.D. (2002) Inter-organis- Management Journal 22, 1139–1161.
ational Communities of Practice. In OKLC 2002 Confer- Larsson, R., Bengtsson, L., Henriksson, K. and Sparks, J. (1998)
ence Proceedings, Athens, Greece. The interorganizational learning dilemma: collective
Huber, G.P. (1991) Organizational learning: the contributing knowledge development in strategic alliances. Organiza-
processes and the literatures. Organization Science 2(1), tion Science 9(3), 285–305.
88–114. Levinson, N.S. and Asahi, M. (1995) Cross-national alliances
Huysman, M. and de Wit, D. (2000) Kennis delen in de praktijk, and interorganizational learning. Organizational Dynam-
vergaren, uitwisselen en ontwikkelen van kennis met ICT. ics Autumn, 50–64.
Van Gorcum, Assen. Lofstrom, S.M. (2000) Absorptive capacity in strategic
Inkpen, A.C. (1998) Learning and knowledge acquisition alliances: investigating the effects of individuals’ social
through international strategic alliances. Academy of Man- and human capital on inter-firm learning. In Organization
agement Executive 12(4), 69–80. Science Winter Conference 2000.
Inkpen, A.C. and Dinur, A. (1998) Knowledge management McEvily, B. and Zaheer, A. (1999) Bridging ties: a source of
processes and international joint ventures. Organization firm heterogeneity in competitive capabilities. Strategic
Science 9(4), 454–468. Management Journal 20, 1133–1156.
Inkpen, A.C. and Crossan, M.M. (1995) Believing is seeing: Mowery, D.C., Oxley, J.E. and Silverman, B.S. (1996) Strategic
joint ventures and organization learning. Journal of Man- alliances and interfirm knowledge transfer. Strategic
agement Studies 32(5), 595–618. Management Journal 17, 77–91.
Kale, P., Dyer, J. and Singh, H. (2001) Value creation and suc- Nonaka, I. and Takeuchi, H. (1995) The Knowledge-creating
cess in strategic alliances: alliancing skills and the role Company: How Japanese Companies Create the Dynamics of
of alliance structure and systems. European Management Innovation. Oxford University Press, New York.
Journal 19(5), 463–471. Powell, W.W. (1990) Neither market nor hierarchy: network
Khanna, T., Gulati, R. and Nohira, N. (1998) The dynamics of forms of organization. Research in Organizational Behavior
learning alliances: competition, co-operation, and rela- 12, 295–336.
tive scope. Strategic Management Journal 19, 193–210. Powell, W.W. (1998) Learning from collaboration: knowledge
Kogut, B. and Zander, U. (1992) Knowledge of the firm, com- and networks in the biotechnology and pharmaceutical
binative capabilities, and the replication of technology. industries. California Management Review 40(3), 228–240.
Organization Science 3(3), 383–397. Teece, D.J. (1994) The dynamic capabilities of firms: an intro-
Kraatz, M.S. (1998) Learning by association? Interorganiza- duction. Industrial and Corporate Change 3(3), 537–556.
tional networks and adaptation to environmental Von Hippel, E. (1987) Co-operation between rivals: informal
change. Academy of Management Journal 41(6), 621–643. know-how trading. Research Policy 16, 291–302.
Kreiner, K. and Schultz, M. (1993) Informal collaboration in Zander, U. and Kogut, B. (1995) Knowledge and the speed of
R&D: the formation of networks across organizations. the transfer and imitation of organizational capabilities:
Organization studies 14(2), 189–209. an empirical test. Organization Science 6(1), 76–92.

MAURA SOEKIJAD, ERIK ANDRIESSEN,

Delft University of Tech- Delft University of Tech-
nology, Department of Work nology, Department of Work
and Organizational Psy- and Organizational Psy-
chology, P.O. Box 5015, chology, P.O. Box 5015,
2600 GA, Delft, The Nether- 2600 GA, Delft, The
lands. E-mail: [Link] Netherlands. E-mail:
@[Link] erika.@[Link]

Maura Soekijad is a doctoral J.H. Erik Andriessen is Pro-

researcher at Delft Univer- fessor of Organizational
sity of Technology, working Psychology researching into
on knowledge sharing groups in competitive alliances. new forms of work with ICT, virtual groups, communi-
ties of practice, and knowledge management.

European Management Journal Vol. 21, No. 5, pp. 578–587, October 2003 587