Strategic Management Journal

Strat. Mgmt. J., 28: 981–1000 (2007)

Published online 16 March 2007 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/smj.616
Received 5 October 2000; Final revision received 15 December 2006

BUILDING FIRM CAPABILITIES THROUGH LEARNING:

THE ROLE OF THE ALLIANCE LEARNING PROCESS IN
ALLIANCE CAPABILITY AND FIRM-LEVEL ALLIANCE
SUCCESS
PRASHANT KALE1 * and HARBIR SINGH2
1
Stephen M. Ross School of Business, University of Michigan, Ann Arbor, Michigan,
U.S.A.
2
The Wharton School of Business, Philadelphia, Pennsylvania, U.S.A.

In recent years, academics and managers have been very interested in understanding how firms
develop alliance capability and have greater alliance success. In this paper, we show that an
alliance learning process that involves articulation, codification, sharing, and internalization of
alliance management know-how is positively related to a firm’s overall alliance success. Prior
research has found that firms with a dedicated alliance function, which oversees and coordinates
a firm’s overall alliance activity, have greater alliance success. In this paper we suggest that
such an alliance function is also positively related to a firm’s alliance learning process, and
that process partly mediates the relationship between the alliance function and alliance success
observed in prior work. This implies that the alliance learning process acts as one of the main
mechanisms through which the alliance function leads to greater alliance success. Our paper
extends prior alliance research by taking a first step in opening up the ‘black box’ between
the alliance function and a firm’s alliance success. We use survey data from a large sample of
U.S.-based firms and their alliances to test our theoretical arguments. Although we only examine
the alliance learning process and its relationship with firm-level alliance success, we also make
an important contribution to research on the knowledge-based view of the firm and dynamic
capabilities of firms in general by conceptualizing this learning process and its key aspects, and
by empirically validating its impact on performance. Copyright  2007 John Wiley & Sons,
Ltd.

INTRODUCTION manage, and that firms generally fail with roughly

half the alliances they form (Bleeke and Ernst,
Organizations have substantially increased their 1993; Kogut, 1988; Alliance Analyst, 1996). On
use of alliances in recent years (Gulati, 1995; the other hand, recent work also shows that not
Zajac, 1998; Dyer and Singh, 1998; Doz and all firms suffer from low alliance success rates;
Hamel, 1998; Gulati, Nohria, and Zaheer, 2000; firms exhibit significant heterogeneity in terms of
Das and Teng, 2000). Extant research has made their overall alliance success, and some firms are
two important observations in the context of this much more successful at managing alliances or
steady uptrend in alliance activity. On the one creating value from them (Anand and Khanna,
hand, most research shows that although alliances 2000; Kale, Dyer, and Singh, 2002) than other
are increasing in popularity, they are difficult to firms. Firms such as Corning, Hewlett-Packard,
or Eli Lilly are examples of firms that belong to
Keywords: alliance learning process; alliance function; the former category (Alliance Analyst, 1996; Sims,
alliance capability; learning and knowledge accumulation Harrison, and Gueth, 2001). Firms with greater
*Correspondence to: Prashant Kale, Stephen M. Ross School of
Business, University of Michigan, 701 Tappan Street, D3602, alliance success are presumed to have alliance
Ann Arbor, MI 48109, U.S.A. E-mail: [email protected] capability.

Copyright  2007 John Wiley & Sons, Ltd.

982 P. Kale and H. Singh

In light of the growing prevalence of alliances (Nonaka, 1994; Grant, 1996) offer useful insights
and the generally low success that firms usually into what some of these processes may be. We
achieve with them, a firm can enjoy a significant build on this literature to develop the notion of
competitive advantage over its peers or rivals if it ‘alliance learning process,’ which is a process
can achieve greater overall alliance success (Dyer that involves articulation, codification, sharing, and
and Singh, 1998; Gulati, 1999). Hence, academics internalization of alliance management know-how.
and managers have become extremely interested in This process is directed toward learning, accumu-
understanding factors that explain how firms have lating, and leveraging alliance management know-
alliance capability and greater alliance success. how to develop a firm’s alliance management
Earlier work on this topic suggested that having skills. Hence, the more developed a firm’s alliance
greater alliance experience helped firms develop learning process, the greater its overall alliance
alliance capability and have greater overall alliance success. We also suggest that the alliance function
success (Lyles, 1988; Simonin, 1997; Anand and is positively related to the alliance learning process
Khanna, 2000). But later work showed that having in firms, and that the process acts as an impor-
a dedicated function to oversee and coordinate a tant mechanism through which the alliance func-
firm’s overall alliance activity perhaps plays a far tion influences a firm’s overall alliance success. In
more important role in explaining firms’ overall other words, the alliance learning process partly
alliance success (Kale et al., 2002). Other research mediates the direct, positive relationship between
and case-based studies (Alliance Analyst, 1996; alliance function and alliance success observed in
Doz and Hamel, 1998; Dyer, Kale, and Singh, prior research (Kale et al., 2002). We use large
2001; Sims et al., 2001; Draulans, deMan, and sample survey and archival data to test our argu-
Volberda, 2003; Bamford, Gomes-Casseres, and ments.
Robinson, 2002) also have demonstrated that firms Our research contributes to extant alliance and
with a dedicated alliance function enjoy greater strategy research in important ways. We contribute
alliance success. This work highlighting the impor- to the alliance literature by conceptualizing and
tance of structural aspects as the alliance function testing the importance of the alliance learning pro-
in explaining alliance capability and success is cess in firms and examining how it might lead to
certainly important. But researchers (Gulati, 1999, greater alliance success; we also show how this
Kale et al., 2002; Hoang and Rothaermel, 2005) learning process partially mediates the impact of
have suggested that we need to extend it further other factors, namely the alliance function, in influ-
by understanding other factors that may also play encing a firm’s overall alliance performance. In
an important role in this regard. Do firms that are doing so, we open up the ‘black box’ between the
more successful with their alliances also follow alliance function and overall alliance success in
other processes or practices that lead to greater firms (Gulati, 1999; Mitchell, 1999; Kale et al.,
alliance success? If yes, what is the relationship 2002). We also contribute to the strategy litera-
between these processes and the dedicated alliance ture on dynamic capabilities in general. Dynamic
function, which is seen to have a direct impact on a capability refers to the capacity of an organiza-
firm’s overall alliance success? This paper adopts tion to purposefully create, extend, or modify its
a ‘learning or a knowledge-based perspective’ to resources or skills (Helfat, 2007). The alliance
address these unanswered questions. learning process that we conceptualize and empir-
Prior strategy research tells us that firms can ically test in this paper is akin to a higher-order
develop skills to successfully manage any given dynamic capability that is relevant in the con-
task by following deliberate, firm-level processes text of alliances. This is because it enables firms
to learn and accumulate knowledge relevant to to achieve greater alliance success by helping
managing that task (Grant, 1996). The resource- them develop or improve their lower-order part-
based view of the firm (Barney, 1991; Dierickx and nering skills to manage different phases or aspects
Cool, 1989), evolutionary economics (Nelson and in alliances more successfully. Finally, by con-
Winter, 1982), research on dynamic capabilities ceptualizing the alliance learning process as one
(Teece, Pisano, and Shuen, 1997; Zollo and Winter, that involves articulation, codification, sharing, and
2002; Eisenhardt and Martin, 2000; Helfat, 2007), internalization of (alliance management) knowl-
organizational learning theory (Huber, 1991), and edge, and testing its relevance and importance in
work on the knowledge-based view of the firm firms, we also make an empirical contribution to
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
 Building Capabilities through Learning 983

the theoretical literature on the knowledge-based provides several benefits to firms (Dyer, Kale and
view of the firm. Singh, 2001). First, it facilitates strategic and oper-
The rest of the paper is organized as follows. ational coordination between the firm’s numerous
In next section, ‘Theory and hypotheses,’ we first alliances. Second, it becomes a focal point for
briefly review prior research that has examined attracting, screening, and identifying appropriate
factors influencing overall firm-level alliance suc- alliance opportunities. Third, it guides individ-
cess. We then conceptualize the notion of alliance ual business units on a variety of alliance-related
learning process in firms and develop hypothe- issues such as searching and selecting appropriate
ses regarding its relationship with a firm’s over- partners, drafting legal agreements, etc. (Mitchell,
all alliance success. We also discuss the relation- 1999; Reuer, 1999). Fourth, it can serve as a focal
ship between the alliance process and the alliance point for initiating organization-wide efforts to
function, and how it explains overall firm-level learn and accumulate alliance management lessons
alliance success. In the following sections we pro- and best practices within a firm. Collectively, these
vide information about the data used to test our actions not only enable better integration across
hypotheses, and present their analyses. We elabo- all alliances in a firm, but also help improve its
rate on our results in the ‘Discussion’ section, and alliance management skills. Hewlett-Packard, Ora-
we review the limitations and contribution of this cle, Siebel, Citibank, and Eli Lilly are some of
work in ‘Future research.’ the companies that have created an alliance func-
tion and achieved greater alliance success (Kale,
THEORY AND HYPOTHESES Dyer and Singh, 2001; Alliance Analyst, 1996;
Sull, 2001).
Academic research that investigates how firms The above-mentioned research on the alliance
have greater alliance success and alliance capa- function has played an important role in explain-
bility is fairly recent. Some of the earlier work in ing a firm’s alliance capability and greater overall
this area (Simonin, 1997; Barkema et al., 1997; alliance success. But in doing so, it has mainly
Anand and Khanna, 2000) suggested that firms paid attention to some of the structural aspects
with greater alliance experience had higher alliance of alliance capability, and perhaps relatively less
success. They also observed significant fixed firm attention to the organizational processes that may
effects in explaining firms’ alliance success; they be involved in this endeavor. Hence, in this paper
interpreted them as a measure of firms’ alliance we attempt to do that, and believe we can under-
capability (Anand and Khanna, 2000: 311) and stand the nature of these processes by turning to
emphasized the need for future research to explore strategy literature that has emphasized the rele-
the organizational determinants of such a capa- vance of learning and knowledge accumulation
bility. Kale et al. (2002) investigated this aspect processes within firms to explain firm capabilities
and found that having a dedicated alliance func- (Conner and Prahalad, 1996; Grant, 1996). This
tion, which was responsible for overseeing and literature includes research on the ‘knowledge-
coordinating a firm’s alliance activity, was posi- based view of the firm’ (Kogut and Zander, 1992;
tively linked to greater alliance success. In other Nonaka, 1994; Grant, 1996), organizational learn-
words, they showed that one way of having ing (Argyris and Schon, 1978; Huber, 1991), and
alliance capability and greater alliance success ‘dynamic capabilities’ (Teece et al., 1997; Eisen-
was to create a dedicated alliance function (Kale hardt and Martin, 2000; Zollo and Winter, 2002;
et al., 2002: 765). This research along with other Winter, 2003).
work (Draulans et al., 2003; Gueth, 2001) comple- In proposing the ‘knowledge-based view of
mented case-based studies (Alliance Analyst, 1996; the firm,’ Grant (1996) argues that organizations
Harbison and Pekar, 1998; Sull, 2001), which also get better at managing any given task (i.e., they
were done at that time to understand the role of develop capabilities to undertake that task) by
the alliance function in explaining alliance capa- accumulating and applying knowledge (Collis,
bility and a firm’s overall alliance success. An 1996) relevant to the execution of that task; firms
alliance function is essentially a structural mech- learn and accumulate such knowledge by making
anism, in the form of a separate organizational deliberate associations between past actions, the
unit or team of managers, responsible for manag- effectiveness of those actions, and future actions
ing and coordinating a firm’s alliance activity, that (Fiol and Lyles, 1985). In their work on dynamic
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
984 P. Kale and H. Singh

capabilities, Zollo and Winter (2002) propose that Articulation of alliance know-how
deliberate learning efforts to articulate and cod-
ify collective knowledge relevant to undertake Individuals in firms are the repositories of know-
complex organizational tasks act as a basis for how and skills related to managing critical tasks
improving a firm’s skills to manage those tasks (Senge, 1997), including the task of managing
more effectively. In other words, these learning alliances. Such knowledge is often ‘tacit or per-
and knowledge accumulation processes reflect a sonal’ in nature (Polanyi, 1966; Badaracco, 1991).
higher-order dynamic capability through which a Nonetheless, companies can learn much if they
firm systematically generates and modifies its oper- can get individual managers to externalize their
ating routines or skills in pursuit of improved personally held, tacit knowledge. The efforts of
effectiveness with the task at hand. According to
accessing and externalizing individually held
Zollo and Winter (2002) such learning processes
knowledge into explicit knowledge, to the extent
may be especially important in the context of
that it is possible, is referred to as ‘articulation.’
building skills to manage tasks or activities that
Knowledge can be articulated in many different
occur repetitively.
ways in firms, such as through spoken or written
We use these ideas to conceptualize the notion
of an ‘alliance learning process’ in firms. In this words, or through the use of metaphors, analogies,
process, firms undertake deliberate efforts to learn, or models. Articulated knowledge is easy to access
accumulate, and leverage alliance management and store, and hence it facilitates learning (Nonaka,
know-how. By doing this, firms are able to develop 1994; Winter, 1987).
or improve their first-order partnering skills and In the context of managing alliances, articulation
thereby achieve better alliance performance. These of alliance management knowledge possessed by
partnering skills are routines or practices to better individual managers can be extremely beneficial.
manage different phases in the life cycle of any First, it helps a firm create a record of its prior
alliance they engage in, i.e., alliance formation and alliance history, which otherwise lapses over time
partner selection, alliance negotiation, formulation due to personnel turnover. Second, the articulation
of alliance design, post-formation alliance man- process itself can facilitate ex post sense-making
agement, etc. In this manner, the alliance learning of actions and decisions in prior alliances as man-
process can help firms develop alliance capabil- agers talk about or reflect on them. This helps
ity and have greater alliance success. Next, we a firm (and its managers) better understand the
describe the various aspects of the alliance learn- causal relationships that might exist between those
ing process and explain how it relates to greater actions and their associated outcomes (Zollo and
alliance success in firms. Winter, 2002). Consequently, articulation can help
managers identify both effective and suboptimal
execution of particular tasks during alliance forma-
The alliance learning process and alliance tion and management, and the practices associated
success with them. In our fieldwork, we have observed that
We see the alliance learning process as a pro- companies follow numerous practices of articulat-
cess that is directed toward helping a firm (and its ing alliance management know-how. Some firms
managers) learn, accumulate, and leverage alliance formally, and regularly, debrief their alliance man-
management know-how and best practices. We agers. In other firms, managers prepare internal
draw motivation from prior research on dynamic reports/presentations pertaining to their alliances.
capabilities (Zollo and Winter, 2002) and the In yet others, managers keep a simple logbook
knowledge-based view of the firm (Grant, 1996; of alliance-related events, decisions, actions, and
Nonaka, 1994) to suggest that such a process outcomes. Collectively, these articulation activi-
involves deliberate efforts to articulate, codify, ties facilitate the externalization of individually
share, and internalize alliance management know- held alliance know-how and extraction of valu-
how in firms. In the following paragraphs we able lessons associated with same. Consequently,
explain each of these aspects of the alliance pro- this aspect of the alliance learning process leads to
cess in greater detail and show how they might more effective management of future or ongoing
relate to a firm’s overall alliance success. alliances, and thereby to greater alliance success.
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
 Building Capabilities through Learning 985

Codification of alliance know-how and frameworks for managing alliances. Hewlett-

Packard has developed ‘40 decision-making tem-
Some scholars have suggested that companies can plates’ to help managers understand and manage
build their skills or expertise to manage any given key activities at every stage of the life cycle of any
task by codifying the task-related knowledge that alliances (Harbison and Pekar, 1998; Dyer et al.,
exists within the firm and its managers (Zollo and 2001). Eli Lilly, which is considered a ‘premier
Singh, 2004; Hansen, Nohria, and Tierney, 1999). partner’ in the pharmaceutical industry, also has
Earlier researchers viewed codification as mere developed several such codified tools and tem-
documentation of existing knowledge (Kogut and plates to improve its managers’ partnering skills
Zander, 1992; Nonaka, 1994) within firms. But (Draulans et al., 2003). Overall, such codification
Zollo and Winter (2002) go one step further and is expected to enhance a firm’s decision making
see it as a more proactive and deliberate effort and actions in its alliances, and consequently lead
that involves creating and using codified resources to greater alliance success over time.
to guide action, based on critical analysis and
abstraction of experience associated with a specific
Sharing of alliance know-how
activity or task.
We conceptualize codification in the alliance According to the knowledge-based view of the
learning process in a similar manner. Codification firm, the development of organizational skills to
involves creating and using knowledge objects or manage any particular task also rests upon a firm’s
resources such as alliance guidelines, checklists, ability to share knowledge associated with manag-
or manuals to assist action or decision making in ing or executing that task with all relevant parts
future alliance situations. We also see it as being within the organization (Grant, 1996). This is not
distinct from the aspect of articulation described only true for knowledge that is articulated and cod-
earlier. Articulation primarily emphasizes external- ified, but also for ‘tacit’ knowledge that is less
izing the content residing within individuals. Cod- amenable to easy articulation or codification (Win-
ification, on the other hand, focuses on providing ter, 1987).
the content (know-what), the methodology (know- Knowledge sharing plays an important role in
how ), and even the rationale (know-why) for this regard. In the context of the alliance learning
executing and managing various alliance-related process, knowledge sharing involves exchanging
tasks. Its ‘people-to-documents’ approach empha- and disseminating individually and organization-
sizes ‘reuse economics,’ by which a firm reuses ally held alliance management knowledge, which
the alliance management knowledge that exists is both tacit and/or codified, through interpersonal
within the firm itself, or that resides with firms interaction within the organization. ‘Communities
or people outside the firm (Hansen et al., 1999), of personal interaction’ are a central element of
to manage future alliances. Although the principal such knowledge sharing within firms (Seely Brown
benefits of codification arise from the use of the and Duguid, 1991; March, Sproull, and Tamuz,
codified alliance management manuals or tools, it 1991). They provide a means for regularly and sys-
also potentially provides more subtle benefits to tematically sharing alliance management knowl-
managers in a firm. By involving themselves in the edge that has already been articulated or codi-
effort to codify alliance management knowledge, fied by the firm. More important, however, they
managers emerge with a crisper understanding of provide a forum to share individually held tacit
what works, or what does not work and why, in knowledge through direct person-to-person inter-
the context of managing certain tasks in alliances. action between managers since tacit knowledge is
Hence codification not only helps firms replicate more easily shared through dialogue between indi-
and transfer alliance best practices, but also iden- viduals than through knowledge objects (Hansen
tify or select what those best practices are. In our et al., 1999). Third, they also play a role in helping
fieldwork, we observed firms adopt several prac- managers better conceptualize the alliance knowl-
tices of codifying alliance management know-how. edge that is being shared or disseminated through-
One company has created ‘35 rules of thumb’ for out the firm. Dialogue in the form of face-to-
managing alliances. Another company has devel- face communication between managers provides
oped an in-house ‘power of partnerships’ program them an opportunity to test their hypotheses and
that provides its managers with detailed guidelines assumptions regarding best practices to carry out
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
986 P. Kale and H. Singh

alliance-related tasks at hand; consequently it cre- wherein the recipient individual focuses on acquir-
ates a platform for constructing a better, shared ing a ‘recipe’ of ‘how to do it’ and not just
understanding of the data and know-how being why it works. Internalization also enhances the
shared. Also, the redundancy or overlap of infor- absorptive capacity (Cohen and Levinthal, 1990)
mation that exists in such sharing processes also of individuals with respect to the task at hand—in
helps managers get a better sense of the alliance this case, the knowledge of managing alliances.
knowledge that is being shared (Glynn, Lant, and By virtue of internalizing existing alliance man-
Milliken, 1994). Hence such knowledge-sharing agement knowledge and best practices, managers
aspects of the alliance learning process play a criti- possess a knowledge base that consequently helps
cal role in the context of learning and accumulating them better absorb any new know-how underly-
alliance know-how within firms. ing relevant skills to better manage alliances in
In our fieldwork we observed companies using the future. Generally, training programs and men-
several practices of alliance knowledge sharing. toring are important internalization mechanisms
They ranged from using informal mechanisms, that firms traditionally use (Davenport and Prusak,
such as casual conversations and discussions 1998) to help individuals to absorb specific lessons
between alliance managers, to having formal mech- and best practices.
anisms such as alliance committees and task forces In the case of alliances, our fieldwork showed
that meet periodically to review and exchange that the alliance learning process in companies
alliance management experiences and best prac- includes a variety of practices to help managers
tices. Some firms also rotate experienced alliance internalize alliance management know-how. For
managers across different alliances within their example, some companies use ‘in-house’ or ‘exter-
company so that the ‘tacit’ alliance wisdom of nal’ alliance training programs to help their man-
these veterans is shared through their interpersonal agers learn and absorb relevant alliance know-how
interaction with others who work with them. One and best practices (Harbinson and Pekar, 1998;
manager we interviewed said: Draulans et al., 2003). Alliance training programs
help managers acquire some basic know-how on
We learned the importance of connecting people alliance management. This knowledge base then
who had the expertise and knowledge of handling
alliances with managers who required it.
provides them a foundation to identify or recognize
new and valuable alliance management know-how
Hence, overall we suggest that the knowledge- or best practices that might exist within their own
sharing aspect of the alliance learning process company or externally, assimilate it, and apply it in
helps firms build their alliancing skills and thereby managing specific alliances they are involved with.
manage alliances more successfully. One company we studied has also created alliance
mentorships for this purpose. Reflecting the use-
Internalization of alliance know-how fulness of this practice, one alliance manager we
interviewed said:
Finally, for firms to enjoy repeated success with
given tasks, it is important that persons responsible At first I was unsure about the usefulness of attend-
for managing those tasks individually possess the ing a training program on alliances. But having
relevant know-how of managing them. Individuals attended one, I gained a lot. Listening to some
can possess this know-how in the form of per- of the senior managers talk about our company’s
alliances made me aware of the specific challenges
sonal skills, heuristics, or mental models. Nonaka and pitfalls that we had encountered in our own
(1994) has suggested that knowledge internaliza- alliances and how we had overcome them. I can
tion, which involves efforts to facilitate absorp- certainly use this knowledge in managing my own
tion of accumulated organizational level know- alliances in the future.
how by individuals, plays an important role in
achieving this. In contrast to knowledge sharing, Overall, knowledge internalization facilitates indi-
which focuses more on dissemination of know- viduals’ absorption of alliance management know-
how between its source and its receiver, inter- how and thereby helps them manage alliances
nalization places more emphasis on the absorp- more effectively.
tion of relevant knowledge by individual receivers. To sum up the discussion thus far, we suggest
The internalization process stresses ‘learning how,’ that a firm’s alliance learning process plays an
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
 Building Capabilities through Learning 987

important role in explaining its overall alliance alliance success. Understanding the direct relation-
success. The alliance learning process is directed ship between the alliance learning process and
toward having alliance capability and greater overall alliance success is certainly important. We
alliance success by helping firms learn, accumu- feel, however, that the alliance learning process
late, and leverage alliance management know-how. also provides a means for partially explaining the
This process involves deliberate efforts to articu- direct positive relationship between the alliance
late, codify, share, and internalize alliance manage- function and overall alliance success. That is, the
ment know-how as described above. Conceptually, alliance learning process potentially represents one
each of these aspects of the alliance learning pro- mechanism through which the alliance function
cess is somewhat distinct in terms of the manner influences a firm’s overall alliance success.
in which it facilitates learning and accumulation of As we noted earlier, prior alliance research
alliance know-how. Articulation helps in external- shows that firms with an alliance function achieve
izing individually held alliance management know- greater overall alliance success because of the
how of managers and making it more explicit. several benefits the function provides. Per this
Codification helps in creating codified and usable research, one of the main benefits of the alliance
tools, templates, or guidelines to help managers function is to act as a focal point for initiat-
when undertaking specific alliance-related tasks. ing or coordinating the learning and accumulation
Sharing helps in disseminating alliance manage- of alliance management know-how within a firm
ment know-how, both explicit and tacit, throughout (Kale et al., 2002; Draulans et al., 2003); firms can
the firm. Finally, internalization helps individual learn this knowledge from sources within or out-
managers absorb or retain the alliance manage- side the firm. Although earlier research has briefly
ment knowledge derived internally from their own highlighted this important point, it does not pro-
firm/colleagues or from external sources. At the vide a detailed theoretical explanation of how the
same time, each of these aspects is uniformly learning actually might occur within firms, and
centered around the learning and accumulation of how that might lead to greater alliance success.
alliance management knowledge within the firm. Our conceptualization of the alliance learning pro-
In other words, they represent an important facet of cess in firms, and how it leads to greater alliance
the alliance learning process that commonly under- success, offers that account.
lies all of them. The theoretical account of the alliance learning
Based on the arguments made in this section, process suggests that establishing and implement-
we feel that a firm’s alliance learning process ing the alliance learning process within firms can
that involves articulation, codification, sharing, and be a fairly complex task. Given the complexity of
internalization of alliance management know-how this task, it seems very likely that firms would be
plays an important role in explaining a firm’s better served by having an entity that is clearly
overall alliance success. Firms that have a stronger responsible and capable of initiating and coordi-
alliance learning process will enjoy greater alliance nating this organization-wide activity on a regular
success. This implies that firms that have higher basis. The alliance function, which is a dedicated
alliance success, and hence alliance capability, do structural entity explicitly responsible for man-
so by having a stronger alliance learning process aging and coordinating firm-wide alliance-related
in place. Therefore: initiatives and activities, seems clearly best posi-
tioned to lead and manage this important task in
firms. By doing so, the function becomes the focal
Hypothesis 1: A firm’s alliance learning process
point for initiating and managing the learning and
is positively linked to its overall alliance suc-
accumulation of alliance management know-how
cess.
within firms. Given the critical role of the alliance
The alliance learning process and the alliance function in coordinating this important task, it
function seems firms that have a dedicated alliance func-
tion are likely to have a stronger alliance learning
Thus far we have highlighted how the alliance process in place; that learning process, in turn,
learning process that involves articulation, codifi- leads to greater alliance success. On the other
cation, sharing, and internalization of alliance man- hand, firms that lack such a function may exhibit
agement know-how is positively linked to greater relatively poor or lesser developed processes to
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
988 P. Kale and H. Singh

learn and accumulate alliance management know- and hence greater overall alliance success. We fur-
how, and hence enjoy relatively lower success ther suggest that the alliance learning process acts
with alliances. This implies that firms with alliance as an important mechanism through which a dedi-
capability enjoy greater alliance success not only cated alliance function influences a firm’s alliance
because of having a dedicated alliance function to success. By doing so, we provide a partial expla-
coordinate their overall alliance activity (as prior nation for the direct positive relationship between
research showed), but also because the function the alliance function and alliance success observed
enables them to have a stronger alliance learn- in prior alliance research. We present the data and
ing process to improve their partnering skills by analyses to test our hypotheses in the next section.
articulating, codifying, sharing, and internalizing
relevant alliance management know-how and best RESEARCH DESIGN AND
practices within the company. METHODOLOGY
Our fieldwork with companies such as Hewlett- First, we did a field study with several companies
Packard, Pfizer Warner-Lambert, and Oracle sup- to understand the nature of the alliance learning
ports this view. Many of the firms we studied process in firms. Based on the fieldwork and a thor-
created a dedicated alliance function to manage ough study of extant literature, we developed our
their alliance activity. This function, as part of its theoretical arguments and model. This approach
core responsibilities, led the implementation and provided rich contextual detail, and enabled us to
institutionalization of their firm’s alliance learning develop grounded specification of the framework
process to articulate, codify, share, or internalize and constructs that used the language of the phe-
alliance management know-how and best practices nomenon. We then collected large sample survey
in these firms. Consequently, this learning pro- data to validate the theoretical constructs and test
cess enabled these firms to achieve greater overall the proposed relationships.
alliance success by helping their managers acquire
or improve their alliance management skills. This Sample selection
suggests that the alliance function is positively
linked to the alliance learning process that, in turn, To select the sample, we first identified firms
influences a firm’s overall alliance success. Thus: whose annual sales for 1998 were greater than
$100 million; we selected these firms from indus-
Hypothesis 2a: The alliance function is posi- tries where alliances are considered an impor-
tively linked to a firm’s alliance learning pro- tant part of firm strategy.1 We then identified
cess. appropriate respondents in each firm using two
criteria: (a) a person’s familiarity with his/her
Hypothesis 2b: The alliance learning process firm’s alliances and alliance management prac-
partially mediates the relationship between the tices; and (b) a person’s ability to respond compre-
alliance function and a firm’s overall alliance hensively to the survey questionnaire. Our field-
success. work had shown that executives in firms’ cor-
porate development or planning departments, or
Prior alliance research has shown that firms that those that belonged to its dedicated team to coordi-
enjoy greater alliance success do so by having nate/manage alliances, if such a unit existed in the
a dedicated alliance function to develop their company, usually met these criteria.2 They were
alliance capability. In this paper, we extend that the primary respondents for our study, but they
work by proposing the following theoretical argu- also helped us in coordinating some of the other
ments. According to the learning and knowledge- data collection in their respective companies. We
based view of the firm (Grant, 1996), a firm can
develop its abilities to successfully manage any 1
Prior research shows that firms in the computer, telecommuni-
given activity or task, by learning or accumulating cations, pharmaceutical, chemical and electronics industries fall
within this category (Culpan and Kostelac, 1993; Alliance Ana-
the knowledge or the know-how relevant to man- lyst, 1996). Industry categories were defined using three-digit
aging that task. In the context of alliances, this SIC codes.
2
means that firms that have a strong alliance learn- We identified such individuals through secondary databases
such as Standard & Poor’s digest on company executives and
ing process to learn and accumulate alliance man- through mailings and telephone calls to the companies (Standard
agement know-how will have alliance capability & Poor’s, 1998).

Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
 Building Capabilities through Learning 989

were able to find the contact information for such Overall, firm-level alliance success
respondents in 692 out of the 932 companies in
In this study, we are interested in a firm’s success
our original set, and hence we mailed the sur-
across its entire portfolio or set of alliances over
vey to them. We received 175 complete responses
a given time period, rather than the performance
(response rate 25%). We observed no significant
of any single alliance.4 To develop this measure,
differences with respect to annual sales, employee
we first assessed the performance of every alliance
size, or alliance experience between companies
belonging to each firm in our sample.5 We then
that received the survey and those that did not.
aggregated these assessments for each firm to cre-
Similarly, we observed no differences between ate an overall, firm-level measure of alliance suc-
the respondent and nonrespondent groups. For the cess or performance for that firm in the sample.
dependent variable, alliance success, managers in Over time, alliance researchers have agreed that
each firm evaluated every alliance their firm had using managerial assessments to measure alliance
formed from 1994 to 1998. Managers in our sam- performance may be one of the most useful ways
ple firms were able to provide these data, even to assess alliance success, especially in large sam-
though 16 percent of the alliances they had formed ple settings, regardless of some of the limitations
had ceased to exist when we collected the data. of this approach. Such a measure has two advan-
Exclusion of data on terminated alliances did not tages. First, it provides a consistent or uniform
change our main results.3 way to measure performance across a large sam-
ple of alliances (Gulati, 1999). Second, the use
Operationalization of key variables of this measure has gained acceptance in alliance
research after Geringer and Hebert (1991) demon-
Alliance learning process strated a positive correlation between alliance per-
formance assessments based on this measure, with
According to our theory, the alliance learning pro- assessments based on other objective measures
cess involves articulation, codification, sharing, that use accounting or financial data. Researchers
and internalization of alliance management know- have measured alliance performance by seeking
how. We used survey-based, multi-item scales to managers’ assessments on a number of differ-
measure each of these aspects of the alliance ent dimensions, such as the extent to which an
learning process. In such scales, individual item alliance has achieved its overall objectives, the
idiosyncrasies cancel one another, making the extent of a parent’s satisfaction with the alliance
measures more reliable (Nunnally, 1978; Marsden, (Killing, 1983; Beamish, 1985; Lyles and Baird,
1990). Since we had little empirical precedent in 1994; Mohr and Spekman, 1994; Saxton, 1997;
developing these measures, we selected the scale
items through fieldwork and through a study of 4
Our independent variable measures, especially those for learn-
relevant academic literature. We also pretested the ing processes, are also at the firm level and not at the level
survey instrument with 30 executives and modi- of each alliance. For this reason, we need an overall firm-level
fied items as necessary (refer to Appendix 1A for measure of alliance performance.
5
details.) For each firm in our sample, we calculated Measuring alliance performance, even at the level of an individ-
ual alliance, has traditionally been subject to considerable debate
the score for each individual aspect of the alliance regarding the appropriate yardsticks to be used, due to the obsta-
learning process by taking a mean of the scores cles of assessing performance in a consistent manner across a
it achieved on the different items representing that large sample of alliances (Gulati, 1999). Some early researchers
used alliance survival as a measure of performance (Harrigan,
particular aspect. We then aggregated these mean 1988). But this measure fails to distinguish between alliances that
scores across the four different aspects to get an fail and therefore die and those that accomplish their objectives
overall score for that firm’s alliance learning pro- and thus are terminated (Kogut, 1988). Scholars have also used
other objective measures, including accounting measures such
cess. as ROI, sales growth, etc. (Kurokawa, 1994; Hagedoorn and
Schakenraad, 1994) or financial measures like abnormal stock
gains (Koh and Venkatraman, 1991; Anand and Khanna, 2000).
3
The 175 companies in our sample evaluated the 3647 alliances Others, however, have argued that these measures too may fail
they formed during 1994–98. The industry-wide breakup of to adequately reflect the extent to which an alliance has achieved
firms was as follows: computer (34%), pharmaceutical (20%), its aims (Geringer and Hebert, 1991). Recently, therefore, many
chemical (16%), electronics (16%), and communications (14%). alliance researchers have reached a consensus that using man-
Sixteen percent of the alliances assessed were terminated as agerial assessments to evaluate alliance performance might be
of the end of 1997, and 37 percent of these were rated as a useful approach, especially since studies reveal that such a
‘successful.’ measure correlates well with other measures of performance.

Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
990 P. Kale and H. Singh

Child and Yan, 1999), the extent to which the across that firm’s entire set of alliances to create
alliance has contributed to the competitive advan- a firm-level measure of average, alliance perfor-
tage and core competency of the parent firms (Sax- mance (we termed this measure AVGPER). The
ton, 1997), and the level of trust and/or harmony two measures of overall, firm-level alliance suc-
between alliance partners (Anderson, 1990; Das cess or performance correlate strongly. We also
and Teng, 2000). However, given the multipurpose conducted some sensitivity analysis by replicating
nature of alliances, Anderson (1990) suggested that the analysis with the performance variable calcu-
instead of using single-item, managerial assess- lated in several other ways, but the results did not
ments of performance, it would be more useful change substantially.7
to have a multidimensional scale for performance
that included several of these dimensions in it.
Therefore, we developed a multidimensional Other variables
scale to measure alliance performance using man- Alliance function
agerial assessments. Performance dimensions were
selected on the basis of prior research (see Prior research has shown that having a dedicated
Appendix 1B). A manager who was most closely alliance function to manage alliances is positively
associated with each specific alliance provided the linked to a firm’s overall alliance success. We first
assessment for that alliance. To avoid common used a dummy variable to measure this variable.
response bias, we ensured that these individuals Firms that had an alliance function were coded as
were different from the primary respondents who ‘1’ and others were coded as ‘0.’ We also measured
provided ratings on the measures for independent this variable using a seven-point scale to assess the
variables representing their firms’ alliance learning extent to which a firm’s overall alliance activity
process. Respondents used a simple seven-point was coordinated by a dedicated function or a team
Likert scale to give their performance assessments. (1 = Very Low and 7 = Very High).
For each firm, we first used the multidimen-
sional scale to evaluate the performance of every Alliance experience
alliance it had formed during 1994–98. We then
used these alliance-level evaluations for that firm Prior research has suggested that a firm’s alliance
to create two measures of its overall, firm-level experience has a positive relationship with its over-
alliance success. One: we calculated an ‘alliance all alliance success, because there is presumably
success rate’ for each firm by using the following an implicit flow of feedback from prior experi-
methodology. We first identified each alliance that ence that enables either an improvement in a firm’s
scored four or more (above-average performance) existing alliance practices or development of new
on any three of the four performance dimensions ones (Anand and Khanna, 2000). Recent work has
as ‘successful or satisfactory.’ We then aggregated also proposed and found a positive relationship
these data to calculate an ‘alliance success rate’ between alliance experience and the alliance func-
for each firm; it is basically a ratio of a firm’s suc- tion. Hence, we controlled for the impact of this
cessful or satisfactory alliances to all its alliances variable in our analyses. We measured a firm’s
during the given time period.6 (For example, if six cumulative alliance experience by taking a count
out of ten alliances of Company A were identi- of all its alliances formed during 1989–98. The
fied as ‘successful’ using the above methodology, Alliances database of the Securities Data Company
its success rate was 60% or 0.60). Two: for each (SDC, 1998) was the primary source for these data.
alliance of a firm we also aggregated the perfor- We verified these data with the primary respondent
mance scores across different dimensions and then in each company and made corrections wherever
took a simple average of these alliance-level scores necessary.

6 7
Besides identifying successful alliances using the methodol- For example, since not all alliances might necessarily involve
ogy described above, we had also asked each respondent to ‘learning opportunities,’ we dropped the ‘learning item’ while
give his/her overall assessment (satisfactory/successful vs. unsat- assessing alliance and firm-level alliance performance. We also
isfactory/unsuccessful) for each alliance based on the various used just the first two dimensions of the performance scale to
performance dimensions. We found that their assessment usu- calculate our performance measures. Our results did not alter
ally correlated strongly (0.91) with our own categorization of substantively when we used these alternate measures of firm-
success/failure for each alliance. level alliance performance in our analyses.

Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
 Building Capabilities through Learning 991

ANALYSES AND RESULTS four, first-order factors; and the alliance learning
process represents the broader, second-order factor
Assessment of scale validity and reliability that commonly underlies all of them. Figure 1 pro-
vides an overview of the second-order, four-factor
Before testing the hypotheses, we assessed the model.
validity and reliability of the survey scales used to Confirmatory factor analysis (Joreskog, 1969)
measure the alliance learning process by following showed that a second-order, four-factor model cor-
methodologies similar to that in prior research. We responding to Figure 1 provides the best fit (χ 2 =
have conceptualized the alliance learning process 143.1 with p > 0.05, GFI = 0.948,
as a process that involves articulation, codification, AGFI = 0.912, and NFI = 0.957) relative to other
sharing, and internalization of alliance manage- models, namely, a null model that assumes no rela-
ment know-how within a firm, and we have used tionship between any of the scale items, a first
survey-based, multi-item scales to measure each order, one-factor model that assumes all survey
of these aspects. We selected individual scale items items to be clubbed directly under just one fac-
for each of these aspects on the basis of systematic tor/construct (χ 2 = 394.3, GFI = 0.734, AGFI =
literature review, fieldwork and pretesting of the 0.658 and NFI = 0.749), or a first-order, four-
survey instrument (refer to Appendix 1A for details factor model where each of the items loads respec-
of the scale items). This approach ensured content tively on the respective factors of articulation,
validity of the measurement scales. The signifi- codification, sharing, and internalization as pro-
cant loadings of individual survey items on their posed, but there is no second-order, common factor
underlying factor established convergent validity such as the alliance learning process underlying all
of these scales (refer to Figure 1 for details). We of them (χ 2 = 153.1, GFI = 0.905, AGFI = 0.886
assessed scale reliability by computing Cronbach and NFI = 0.902). Further, in the second-order,
alpha coefficients for each of the underlying fac- four-factor model correlations between the first-
tors. They were well above the acceptable thresh- order factors are significant (p < 0.01), and each
old level of 0.70 (Nunnally, 1978). of the first-order factors also shows a high fac-
As we explained earlier, each aspect of the tor loading on the second-order factor. Overall,
alliance learning process is somewhat distinct in these results confirm that the ‘alliance learning pro-
terms of how it facilitates the learning and lever- cess’ is indeed a second-order, common factor that
aging of alliance management know-how within underlies the different aspects or factors of artic-
a firm, and leads to greater alliance success; at ulation, codification, sharing, and internalization
the same time, each aspect is commonly directed that comprise it.8
toward the learning and accumulation of alliance
management know-how. Hence they are likely to
be correlated with each other, and represent differ- 8
We also compared the proposed second-order, four-factor
ent facets of the alliance learning process that com- model of the alliance learning process with other theoreti-
cally plausible models: a second-order, two-factor model where
monly underlies all of them. Thus, we used confir- alliance learning process involves just two aspects—one aspect
matory factor analysis to estimate a second-order is Codification and the other aspect is a combination of Articula-
factor model that best represents these relation- tion, Sharing, and Internalization (Zollo and Winter, 2002), and
another second-order, three-factor model where alliance learning
ships. The four aspects of knowledge articulation, process involves three aspects: Codification and Internalization
codification, sharing, and internalization represent are two of them and the third is a combination of Articulation
0.72
0.76
0.69
0.79
0.73

0.73
0.81
0.72
0.69

0.75
0.78
0.75
0.73
0.59
0.67

0.67
0.63

0.67
0.62

X1 X2 X X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 X16 X17 X18 X19
3

Articulation Codification Sharing Internalization

0.861 0.823
0.748 0.623
Alliance
Learning Process

Figure 1. A second-order four-factor model: a confirmatory factory analysis

Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
992 P. Kale and H. Singh

Hypothesis testing Bold lines in the figure represent the main rela-
Table 1 provides the correlation matrix for the key tionships proposed in this paper: (a) the direct
variables. We observe that the alliance learning relationship between the alliance learning process
process and the alliance function are both pos- and alliance success proposed in Hypothesis 1;
itively and significantly correlated with a firm’s (b) the positive relationship between the alliance
overall alliance success. We also note the alliance function and alliance learning process proposed in
function is significantly and positively correlated Hypothesis 2a; and (c) the mediating effect of the
with the alliance learning process. We then used alliance learning process proposed in Hypothesis
structural modeling to test our hypotheses since 2b. Dashed lines in the figure represent the rela-
it enabled us to test all the proposed hypothe- tionships that we control for in our study since
ses at the same time by simultaneously estimating they have been observed in prior research. These
multiple, dependent relationships between the con- include the direct relationship between the alliance
structs of interest. In structural modeling, the mea- function and alliance success (Kale et al., 2002),
surement and structural sub-models are estimated as well as the relationship between alliance expe-
simultaneously. The measurement model uses con- rience and the alliance function, and that between
firmatory factor analysis to assess the validity and alliance experience and alliance success (Anand
reliability of the scales used to measure the con- and Khanna, 2000; Kale et al., 2002). Earlier in the
structs, whereas the structural model estimates the paper, we reported the results of the measurement
strength and direction of relationships between model used to confirm the validity and reliability
them (Anderson and Gerbing, 1988; Hair et al., of the scales used to measure the alliance learning
1998). process. Therefore, we do not report it again here.
The structural model that we tested is presented We only report results for overall model fit and
in Figure 2; we refer to it as the ‘Proposed Model.’ structural parameters for the structural sub-model
presented in Figure 2.
and Sharing (Nonaka, 1994). These models, however, exhibited
The dependent variable in our model is a firm’s
an inferior fit. overall alliance success rate. Our analysis showed

Table 1. Descriptive statistics and correlation matrix

Mean S.D. 1 2 3 4 5 6 7

1 Alliance learning process 13.68 4.60 1.00

2 Alliance function 0.59 0.46 0.44 1.00
3 Alliance experience 20.84 16.87 0.17 0.37 1.00
4 Alliance success rate 0.56 0.15 0.422 0.38 0.31 1.00
5 Avg. alliance performance 17.06 4.19 0.438 0.39 0.35 0.72 1.00
6 Firm size (sales in $ billion) 4.84 7.05 0.074 0.18 0.21 0.102 0.132 1.00
7 Firm size (employees ‘000) 18.26 26.15 0.085 0.17 0.22 0.065 0.082 0.834 1.00

Figures in italics are significant at the 0.05 level.

X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13X14 X15 X16 X17 X18 X19

Articulation Codification Sharing Internalization

H2a Alliance H1 Alliance

Alliance
Function Learning Process Success Rate
0.315 ** 0.374 **

0.398 ** 0.104

Alliance
Experience
0.114

Figure 2. Structural model. The bold lines represent the relationships proposed in this paper; the dashed lines represent
relationships observed in prior research; ∗∗ p < 0.05
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
 Building Capabilities through Learning 993

our proposed model fits the data very satisfacto- other models that may be theoretically possible
rily. Both the absolute and incremental fit indices (Gerbing and Anderson, 1988; Hair et al., 1998).
are above the generally acceptable level of 0.90 We briefly discuss those models and their results
(χ 2 = 191.05, GFI = 0.921, AGFI = 0.892, and below.
NFI = 0.908), and the chi-square is nonsignificant
at p > 0.01. As far as the structural parameters Baseline model (Model a)
are concerned, we observe the following: (a) the
This model includes only those variables that
alliance learning process is positively and sig-
were deemed important to alliance success in prior
nificantly related to firm-level alliance success;
alliance research, i.e., a model with only alliance
(b) the alliance function is also positively related to
experience and alliance function having a direct
the alliance learning process; and (c) the alliance
relationship with alliance success. The chi-square
function does not have a very significant, direct
values and goodness-of-fit indices for this model
relationship with firm-level alliance success when
indicate an inferior fit as compared to our proposed
its relationship with success is mediated by the
model (GFI = 0.577, AGFI = 0.481).
alliance learning process. Overall, the results pro-
vide support for Hypotheses 1, 2a and 2b.9
Model b
We also note some interesting relationships as
far as alliance experience is concerned. Unlike This model includes all the relationships from
prior research, we see that alliance experience ‘Model a’ described above, as well as an additional
ceases to have a significant, direct relationship with relationship between alliance experience and the
a firm’s overall alliance success if we include the alliance function that has been observed in some
other variables that might influence alliance suc- earlier work. However, this model exhibits an
cess in our model. We note, however, that alliance inferior fit as compared to our proposed model
experience is positively and significantly linked to (GFI = 0.643, AGFI = 0.571).
the alliance function; this result is consistent with
that observed in prior research (Kale et al., 2002). Model c
We discuss the implications of these results in the
This model not only has all the relationships from
next section.10
‘Model b’ described above, but also includes an
To ensure the acceptability of the proposed the-
additional relationship between the alliance learn-
oretical model, we also compared it with several
ing process and alliance success, proposed in
Hypothesis 1. In this model, however, we do not
9
Prior alliance research (Kale et al., 2002) found that the alliance include the relationship between the alliance func-
function has a very significant, direct relationship with a firm’s
overall alliance success (p < 0.01). But in our model we see that tion and alliance learning process that we have
the direct relationship between the alliance function and alliance proposed in Hypothesis 2a. Although the fit of
success ceases to be very significant (p < 0.10) if we also this model improves considerably (GFI = 0.881,
account for the mediating effect of the alliance learning process
in between. We also tested for the mediating effects of the AGFI = 0.848) compared to ‘Model b,’ it is still
alliance learning process using tests based on OLS regressions, inferior compared to our proposed model.
as suggested by Baron and Kenny (1986), and we found strong
support for it.
10
Model d
After controlling for the relationship between alliance experi-
ence and the alliance function, which has been studied in prior This is a ‘full’ model that is similar to the one in
research, our figure implies that the alliance function probably
mediates the relationship between alliance experience and the Figure 2, but it also includes an additional direct
alliance learning process. However, one of the reviewers sug- path between alliance experience and alliance
gested that it might be useful to examine whether the alliance learning process. This direct relationship tests
function moderates the relationship between alliance experience
and the alliance learning process, instead of mediating it. We whether firms’ alliance experience directly helps
tested for such a moderating or interaction effect using structural them have a stronger alliance learning process
modeling techniques as well as tests based on OLS regressions without having an alliance function in between.
as suggested by Baron and Kenny (1986), and we did not find
strong support for it. However, when we tested for the medi- This model was less significant than our ‘proposed
ating effects of the alliance function using OLS regression as model’ and was also less parsimonious; the direct
suggested by Baron and Kenny (1986), we found strong support relationship between alliance experience and the
for it. This suggests that the alliance function mediates the rela-
tionship between alliance experience and the alliance learning alliance learning process was also not very signif-
process. icant.
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
994 P. Kale and H. Singh

Overall, we see that model fit improves consid- Prior research has indicated that a dedicated
erably (reduction in chi-square and improvement alliance function has a direct influence on firms’
in goodness-of-fit values) when we go from any of alliance capability and alliance success. However,
the above alternate models to our proposed model in our paper we find that the alliance function
presented in Figure 2; this provides strong support is also positively linked to the alliance learning
for the theoretical relationships proposed in this process that we have examined here. We further
paper. We noticed that model fit improves even observe that including the relationship between the
more if we drop the nonsignificant direct rela- alliance function and the alliance learning process
tionships of alliance experience and the alliance in a theoretical model that explains firms’ alliance
function with alliance success, respectively, from success reduces the direct influence of the alliance
the proposed model. We also replicated all the function on alliance success. Our arguments and
analyses using the alternate measure of firm-level findings suggest that the alliance learning pro-
alliance success, namely a firm’s average alliance cess, which we conceptualize and describe in this
performance, and obtained similar results. paper, partially mediates the relationship between
the alliance function and alliance success observed
in prior research; in other words, the alliance learn-
DISCUSSION ing process seems to act as one of the main mech-
anisms through which the alliance function influ-
Why do some firms have greater, overall alliance ences a firm’s alliance success. By uncovering this
success than other firms? In this paper, we provide relationship, we extend prior research on this sub-
insight into this issue by investigating the role of ject in important ways because we actually open
some critical organizational learning processes that the ‘black box’ between the alliance function and
may play an important role in influencing a firm’s alliance success. Prior case studies (Alliance Ana-
overall alliance success. Our work builds upon, and lyst, 1996; Gueth, 2001; Sull, 2001) and academic
extends, earlier alliance research that has examined research (Kale et al., 2002; Draulans et al., 2003)
factors explaining firms’ alliance capability and have suggested that the alliance function leads to
overall alliance success. greater alliance success because of the many ben-
We provide a theoretical account of the alliance efits it provides, one of them being its potential
learning processes that potentially explain firms’ to act as a focal point for learning and leverag-
greater overall alliance success. Drawing on the lit- ing alliance management lessons and know-how
erature on the ‘knowledge-based view of the firm’ to improve a firm’s alliancing skills. Although
(Nonaka, 1994; Grant, 1996) and ‘dynamic capa- prior research highlighted this point, it failed to
bilities’ (Teece et al., 1997; Eisenhardt and Martin, provide a detailed account of how that learning
2000; Zollo and Winter, 2002), we suggest that or knowledge accumulation might occur in firms.
the ‘alliance learning process’ is directed toward By conceptualizing the alliance learning process
learning and accumulating alliance management and its different aspects, and by showing not only
know-how and best practices in firms. We con- how this learning process has a direct impact on
ceptualize it as a process that involves articulation, firms’ alliance success but also how it mediates
codification, sharing, and internalization of alliance the relationship between the alliance function and
management know-how within firms. We describe alliance success, we address this important gap in
how each of these aspects of the alliance learning the alliance literature.
process plays a very important role in helping a In our analysis, we have also controlled for
firm and its managers develop their alliance man- the influence of alliance experience on the key
agement know-how and/or improve their partner- variables of interest. Earlier scholars had argued
ing skills. As a consequence, the alliance learning that greater alliance experience is directly linked
process has a positive and significant relationship to greater alliance success of firms (Lyles, 1988;
with firms’ overall alliance success. Our empirical Anand and Khanna, 2000). Subsequent work had
analysis findings support this argument, implying also found that firms with greater alliance experi-
that firms with a stronger alliance learning pro- ence are more likely to have a dedicated alliance
cess to learn and accumulate alliance management function, implying that the effects of alliance
know-how and practices have greater alliance suc- experience on firms’ alliance success may occur
cess. through the choice of having such a function.
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
 Building Capabilities through Learning 995

Studying the role of alliance experience on firms’ alliance success, we address this gap in the lit-
overall alliance success was not the primary focus erature. We develop and validate the measure of
of this paper. However, we feel that it may the alliance learning process in firms that involves
be worthwhile to briefly discuss the results we articulation, codification, sharing, and internaliza-
observe in this paper after controlling for alliance tion of alliance management know-how, and we
experience in our model. In our analyses we see illustrate its importance in explaining differential
that alliance experience is positively and signifi- alliance success. Therefore, by using alliances as
cantly linked to the alliance function, but it has no a context, we empirically assess the existence and
significant direct relationship with either alliance importance of some of the learning and knowledge
success or the alliance learning process when we accumulation processes outlined in the knowledge-
also account for the other proposed relationships based view of the firm (Nonaka, 1994; Grant,
in this paper. We also find strong support for our 1996). Our work also contributes to dynamic capa-
hypotheses even after we account for the relation- bilities research (Teece et al., 1997, Zollo and Win-
ship between alliance experience and the alliance ter, 2002; Eisenhardt and Martin, 2000; Helfat and
function; namely, we see that the alliance learning Peteraf, 2003). Dynamic capabilities are seen as
process is positively linked to alliance success, and ‘higher-order capabilities that help a firm extend,
it seems to mediate the relationship between the modify, or improve its ordinary or operational
alliance function and a firm’s overall alliance suc- capabilities that are relevant to managing any given
cess. Taken together, these findings suggest some task.’ In the context of alliances, a firm’s skills to
interesting relationships. It seems that firms with manage different aspects of any alliance (Gulati,
more alliance experience are likely to have a ded- 1999) represent relevant operational skills neces-
icated function to manage their alliances; in turn, sary to manage alliances. But the alliance learning
the alliance function leads to greater alliance suc- process seems like a higher-order dynamic ability
that helps a firm learn, accumulate, and leverage
cess through a strong alliance learning process
alliance know-how so as to modify or improve its
that is directed toward learning and accumulating
operational alliance management skills and achieve
alliance management know-how and best practices
greater overall alliance success. Thus, we con-
in firms. This may suggest that alliance experience
tribute to dynamic capabilities research by concep-
may not directly affect alliance success, as previ-
tualizing and validating the relevance of a potential
ously believed, but rather its effect occurs through
dynamic capability in the context of alliances. We
the intervening factors of the alliance function and believe that a similar learning process may also
the alliance learning process respectively. Future be relevant in helping firms develop their opera-
research needs to investigate these relationships by tional skills to manage other important organiza-
undertaking in-depth case-based research to fully tional tasks such as acquisitions, restructurings, etc
understand the interdependent nature of these rela- (Zollo and Winter, 2002).
tionships and how they evolve over time in firms. This research also contributes to a deeper under-
Broadly speaking, this research also contributes standing of the development of competitive advan-
to the knowledge-based view of the firm (Grant, tage by earning relational rents (Dyer and Singh,
1996; Nonaka, 1994). A central theme of this 1998). From a relational perspective, interfirm
research is that organizational processes to learn knowledge sharing routines are argued to be one
and leverage individually and organizationally held of the four major drivers of competitive advantage
knowledge act as a basis for improving a firm’s in individual alliances (along with having com-
ability to manage a given task or activity more plementary capabilities, investing in relationship-
effectively (Conner and Prahalad, 1996; Grant, specific assets, and choosing appropriate gover-
1996; Eisenhardt and Martin, 2000). There is very nance mechanisms). Interfirm knowledge sharing
little research, however, that tries to conceptualize routines are important to establish a repeatable pro-
and empirically analyze some of the specific orga- cess for engagement across organizations. In this
nizational processes through which knowledge is paper, by studying the alliance learning process in
accessed, learned, shared, or leveraged. By using firms, we focus on some of the intrafirm knowl-
firms and their alliances as a context, and by con- edge processes or routines that involve codification
ceptualizing the alliance learning process that helps and sharing of alliance management know-how.
a firm have alliance capability and greater overall Using Helfat’s definition of capabilities (Helfat,
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
996 P. Kale and H. Singh

2007), relational capability can be viewed as the at any stage? Our preliminary analysis showed
capacity of the organization to create, extend, that the square term of alliance learning process
or modify a firm’s resource base, augmented to has a negative, but insignificant relationship with
include preferred access to the resources of its a firm’s overall alliance success. Third, there are
alliance partners. In the context of any specific some limitations in how we have measured alliance
alliance, a firm’s relational capability would get success in firms. In this paper we have used man-
reflected in its ability to implement the four drivers agerial assessments on a multidimensional scale
of competitive advantage mentioned above. But to measure alliance performance, and we have
the creation of such relational capability in a firm aggregated these scores to get an overall firm-level
would envision a purposeful process that is more measure of alliance success. Future work could
likely to succeed if particular structural and process conceptualize alliance success at the firm level
decisions within the firm are made on the lines we in different ways, such as using measures based
have suggested in this paper, and if there is sup- on financial or accounting data, and replicate our
port for creation of such capabilities at the senior research to examine the robustness of our measures
management level. and findings.
The managerial implications of ‘opening the
black box’ between the alliance function and over- CONCLUSION
all alliance success are obviously important. Com-
This paper makes several contributions to alliance
panies that desire to have alliance capabilities and
and strategy research. It contributes to alliance
greater overall alliance success certainly need to
research by showing how the alliance learning pro-
have a dedicated alliance function to manage their
cess that involves articulation, codification, shar-
alliances. But, more important, they also need to
ing, and internalization of alliance management
have a strong alliance learning process to articu-
know-how leads to greater, overall alliance suc-
late, codify, share, and internalize alliance man-
cess of firms. It also suggests that the alliance
agement know-how since it has a direct impact on
learning process may, in fact, mediate the impact
firms’ alliance capability and overall alliance suc-
of the alliance function on firms’ alliance success.
cess, and it acts as one of the main mechanisms
That is, the alliance learning process acts as one of
through which the alliance function leads to greater
the main mechanisms through which the alliance
alliance success. function leads to alliance success. Earlier research
(Kale et al., 2002) opened up the ‘black box’
FUTURE RESEARCH between alliance experience and alliance success
by highlighting the role of the alliance function
This research, like any other, has some limitations
in explaining alliance capability and alliance suc-
that future work can address. First, in this paper
cess. In this paper, by conceptualizing the alliance
we suggest that a firm’s alliance learning process
learning process and its effect on a firm’s alliance
leads to greater overall alliance success by presum-
success, we carry that work forward by further
ably improving its first-order alliance management
opening up the ‘black box’ between the alliance
skills. But we do not directly measure a firm’s
function and alliance success. Finally, by using
alliance management skills (i.e., skills required to
a knowledge-based approach to conceptualize the
manage any particular aspect of the alliance life alliance learning process, and by empirically val-
cycle). It would have been very difficult and time idating and testing its impact on firms’ alliance
consuming to get these data for the large num- performance, this research also contributes to lit-
ber of firms in our sample. In the future, however, erature on the knowledge-based view of the firm
scholars could attempt to do that either through and dynamic capabilities in general.
case-based research or by collecting detailed data
on these practices for a small subset of firms and ACKNOWLEDGEMENTS
their alliances. Second, it would be useful to exam-
ine whether the alliance learning process has any We acknowledge the financial assistance provided
adverse or declining effects in firms. Given the by the William and Phyllis Mack Center for Tech-
high levels of investments that may be involved in nological Innovation at the Wharton School, Uni-
having a strong alliance learning process, would versity of Pennsylvania, for this research. Sid-
the costs of this endeavor outweigh the benefits ney Winter, Maurizio Zollo, Jeffrey Dyer, Phanish
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
 Building Capabilities through Learning 997

Puranam, Sendil Ethiraj, and Will Mitchell pro- In Multinational Strategic Alliances, Culpen R (ed).
vided valuable suggestions and comments in the Haworth Press: Binghamton, NY; 103–122.
Das TK, Teng B. 2000. Instabilities of strategic alliances:
course of several conversations on this paper.
an internal tensions perspective. Organization Science
11(1): 77–101.
Davenport T, Prusak L. 1998. Working Knowledge: How
REFERENCES Organizations Manage What They Know . Harvard
Business School Press: Boston, MA.
Alliance Analyst. 1996. Managing alliances: skills for the Dierickx I, Cool K. 1989. Asset stock accumulation and
modern era. March. sustainability of competitive advantage. Management
Anand B, Khanna T. 2000. Do firms learn to create Science 35(12): 1504–1511.
value? The case of alliances. Strategic Management Doz Y, Hamel G. 1998. Alliance Advantage: The Act of
Journal 21(3): 295–316. Creating Value through Partnering. Harvard Business
Anderson E. 1990. Two firms, one frontier: on assessing School Press: Boston, MA.
joint venture performance. Sloan Management Review Draulans J, deMan A-P, Volberda HW. 2003. Building
31(2): 19–30. alliance capability: management techniques for
Anderson J, Gerbing D. 1988. Structural equation superior alliance performance. Long Range Planning
modeling in practice: a review and recommended 36(2): 151–166.
two-step approach. Psychological Bulletin 103(3): Dyer J, Singh H. 1998. The relational view: cooperative
411–423. strategy and sources of international competitive
Argyris C, Schon D. 1978. Organizational Learning. advantage. Academy of Management Review 23(4):
Addison-Wesley: Reading, MA. 660–679.
Badaracco JL. 1991. The Knowledge Link: How Firms Dyer J, Kale P, Singh H. 2001. How to make strategic
Compete through Strategic Alliances. Harvard Busi- alliances work. Sloan Management Review 42(4):
ness School Press: Boston, MA. 37–43.
Bamford J, Gomes-Casseres B, Robinson M. 2002. Mas- Eisenhardt K, Martin J. 2000. Dynamic capabilities: what
tering Alliance Strategy: A Comprehensive Guide to are they? Strategic Management Journal , Special
Design, Management, and Organization. Jossey-Bass: Issue 21(10–11): 1105–1121.
San Francisco, CA. Fiol CM, Lyles MA. 1985. Organizational learning.
Barkema HG, Sheker O, Vermeulen F, Bell JH. 1997. Academy of Management Review 10(4): 803–813.
Working abroad, working with others: how firms learn Gerbing DW, Anderson JC. 1988. An updated paradigm
to operate international joint ventures. Academy of for scale development incorporating unidimensionality
Management Journal 40(2): 426–442. and its assessment. Journal of Marketing Research
Barney J. 1991. Firm resources and sustained competitive 25(2): 186–192.
advantage. Journal of Management 17(1): 99–120. Geringer JM, Hebert L. 1991. Measuring performance of
Baron RM, Kenny DA. 1986. The moderator–mediator international joint ventures. Journal of International
variable distinction in social psychological research: Business Studies 22(2): 249–263.
conceptual, strategic, and statistical considerations. Glynn M, Lant T, Milliken F. 1994. Mapping learning
Journal of Personality and Social Psychology 51(6): processes in organizations. In Advances in Managerial
1173–1182. Cognition and Organizational Information Processing,
Beamish PW. 1985. Joint Venture Performance in Vol. 5, Stubbert C, Meindl JR, Porac JFA (eds). JAI
Developing Countries. University of Western Ontario: Press: Oxford, UK; 43–83.
London, Canada. Grant RM. 1996. Toward a knowledge-based theory
Bleeke J, Ernst D (eds). 1993. Collaborating to Compete: of the firm. Strategic Management Journal , Winter
Using Strategic Alliances and Acquisitions in the Special Issue 17: 109–122.
Global Marketplace. Wiley: New York. Gueth A. 2001. Entering into an alliance with big
Child J, Yan Y. 1999. Predicting the performance of pharma: benchmarks for drug delivery contract
international alliances: an investigation in China. service providers. Pharmaceutical Technology 25(10):
Working paper, Chinese Management Centre, Univer- 132–135.
sity of Hong Kong. Gulati R. 1995. Does familiarity breed trust? The
Cohen W, Levinthal D. 1990. Absorptive capacity: a new implications of repeated ties for contractual choice
perspective on learning and innovation. Administrative in alliances. Academy of Management Journal 38(1):
Science Quarterly 35(1): 128–152. 85–112.
Collis D. 1996. Organizational capability as a source of Gulati R. 1999. Network location and learning: the
profit. In Organizational Learning and Competitive influence of network resources and firm capabilities
Advantage, Moingeon B, Edmondson A (eds). Sage: on alliance formation. Strategic Management Journal
London; 139–163. 20(5): 397–420.
Conner KR, Prahalad CK. 1996. A resource-based Gulati R, Nohria N, Zaheer A. 2000. Strategic networks.
theory of the firm: knowledge versus opportunism. Strategic Management Journal 21(3): 203–215.
Organization Science 7(5): 477–501. Hagedoorn J, Schakenraad J. 1994. The effect of strategic
Culpan R, Kostelac EA Jr. 1993. Cross-national cor- technology alliances on company performance.
porate partnerships: trends in alliance formation. Strategic Management Journal 15(4): 291–309.
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
998 P. Kale and H. Singh

Hair J, Anderson R, Tatham R, Black W. 1998. Multi- Kurokawa S. 1994. Technological alliance strategies in
variate Data Analysis. Prentice-Hall: Upper Saddle Japanese manufacturing firms. Paper presented at the
River, NJ. annual meeting of the Academy of Management,
Hansen M, Nohria N, Tierney T. 1999. What’s your Dallas, TX.
strategy for managing knowledge? Harvard Business Lyles MA. 1988. Learning among joint venture sophisti-
Review March–April: 106–116. cated firms. Management International Review , Spe-
Harbison J, Pekar P Jr. 1998. Smart Alliance: A Practical cial Issue 28: 85–98.
Guide to Repeatable Success. Jossey-Bass: San Lyles MA, Baird IS. 1994. Performance of international
Francisco, CA. joint ventures. International Management Review
Harrigan KR. 1988. Strategic alliances and partner 34(4): 313–329.
asymmetries. Management International Review 28: March JG, Sproull LS, Tamuz M. 1991. Learning from
53–72. samples of one and fewer. Organization Science 2(1):
Helfat C. 2007. Relational capabilities: drivers and impli- 1–13.
cations. In Dynamic Capabilities: Strategic Change in Marsden PV. 1990. Network data and measurement.
Organizations, Helfat CE, Finkelstein S, Mitchell W, Annual Review of Sociology 16: 435–463.
Peteraf M, Singh H, Teece DJ, Winter SG (eds). Mitchell W. 1999. Alliances: Achieving Long-term Value
Blackwell: Oxford, UK; 65–80. and Short-term Goals. Mastering Strategy Series.
Helfat C, Peteraf M. 2003. The dynamic resource-based Financial Times: London.
view: capability lifecycles. Strategic Management Mohr J, Spekman R. 1994. Characteristics of partnership
Journal 24(10): 997–1010. success: partnership attributes, communication behav-
Hoang H, Rothaermel FT. 2005. The effect of general ior, and conflict resolution techniques. Strategic Man-
and partner-specific alliance experience on joint R&D agement Journal 15(2): 135–152.
project performance. Academy of Management Journal Nelson R, Winter S. 1982. An Evolutionary Theory of
48(2): 332–345. Economic Change. Belknap Press: Cambridge, MA.
Huber GP. 1991. Organizational learning: the contribut- Nonaka I. 1994. A dynamic theory of organizational
ing processes and the literatures. Organization Science knowledge creation. Organization Science 5: 14–37.
2(1): 88–115. Nunnally JC. 1978. Psychometric Theory (2nd edn).
Inkpen A, Crossan M. 1995. Believing is seeing: joint McGraw-Hill: New York.
ventures and organization learning. Journal of Polanyi M. 1966. The Tacit Dimension. Routledge &
Management Studies 32(5): 595–619. Kegan Paul: London.
Joreskog KG. 1969. A general approach to confirmatory Reuer J. 1999. Collaborative Strategy: The Logic of
maximum likelihood factor analysis. Psychometrika Alliances. Mastering Strategy Series. Financial Times:
34: 183–202. London.
Kale P, Singh H, Perlmutter H. 2000. Learning and Saxton T. 1997. The effects of partner and relationship
protection of proprietary assets in strategic alliances: characteristics on alliance outcomes. Academy of
building relational capital. Strategic Management Management Journal 40(2): 443–461.
Journal 21(3): 217–237. SDC. 1998. Securities Data Company, Thomson
Kale P, Dyer J, Singh H. 2001. Value creation and Financial. New York.
success in strategic alliances. European Management Seely Brown J, Duguid P. 1991. Organizational learning
Journal 19(5): 463–471. and communities-of-practice: toward a unified view
Kale P, Dyer J, Singh H. 2002. Alliance capability, stock of working, learning, and innovation. Organization
market response, and long-term alliance success: the Science 2(1): 40–56.
role of the alliance function. Strategic Management Senge P. 1997. Sharing knowledge: the leader’s role is
Journal 23(8): 747–768. key to a learning culture. Executive Excellence 4(11):
Khanna T, Gulati R, Nohria N. 1998. The dynamics 17–18.
of learning alliances: competition, cooperation and Simonin BL. 1997. The importance of collaborative
relative scope. Strategic Management Journal 19(3): know-how: an empirical test of the learning
193–210. organization. Academy of Management Journal 40(5):
Killing JP. 1983. Strategies for Joint Venture Success. 1150–1174.
Routledge & Kegan Paul: New York. Sims N, Harrison R, Gueth A. 2001. Managing alliances
Kogut B. 1988. A study of the life cycle of joint ventures. at Lilly. In Vivo 19(6): 71–77.
In Cooperative Strategies in International Business, Standard & Poor’s Register of Corporations, Directors
Contractor FK, Lorange P (eds). Lexington Books: and Executives. 1998. McGraw-Hill: New York.
Lexington, MA; 169–186. Sull D. 2001. Siebel Systems: partnering to scale.
Kogut B, Zander U. 1992. Knowledge of the firm, Harvard Business School Press Case 9-802-029 ;
combinative capabilities and the replication of 1–24.
technology. Organization Science 3: 383–397. Teece DJ, Pisano G, Shuen A. 1997. Dynamic capabili-
Koh J, Venkatraman N. 1991. Joint venture formations ties and strategic management. Strategic Management
and stock market reactions: an assessment in Journal 18(7): 509–533.
the information technology sector. Academy of Winter SG. 1987. Knowledge and competence as
Management Journal 34: 869–892. strategic assets. In The Competitive Challenge:
Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
 Building Capabilities through Learning 999

Strategies for Industrial Innovation and Renewal , APPENDIX 1A: LIST OF ITEMS USED
Teece DJ (ed). HarperCollins: New York; 159–184. TO MEASURE DIFFERENT ASPECTS
Winter SG. 2003. Understanding dynamic capabilities. OF THE ALLIANCE LEARNING
Strategic Management Journal 24(10): 991–995.
PROCESS
Zajac EJ. 1998. Commentary on ‘alliances and networks’
by R. Gulati. Strategic Management Journal , Special
Issue 9(4): 319–321. The following items were used in the survey ques-
Zollo M, Singh S. 2004. Deliberate learning in corporate tionnaire to measure each aspect of the alliance
acquisitions: post-acquisition strategies and integration
learning process. Respondents used a seven-point
capability in U.S. bank mergers. Strategic Manage-
ment Journal 25(12): 1233–1257.
Likert-type scale to indicate his/her disagreement
Zollo M, Winter S. 2002. Deliberate learning and or agreement with the statement that represented
the evolution of dynamic capabilities. Organization each item, such that ‘1 = Strongly Disagree’ and
Science 13(3): 339–351. ‘7 = Strongly Agree.’

Item no. Construct items Theoretical reference

Knowledge articulation
X1. Managers involved with the company’s alliances are regularly Alliance Analyst (1996); Harbison
debriefed about their prior and/or current alliance experience. and Pekar (1998); Nonaka (1994);
X2. Managers responsible for the company’s alliances maintain a Davenport and Prusak (1998);
record (in the form of a memo, note, report, or presentation) Winter (1987)
of all major incidents, decisions, or actions associated with
their respective alliance(s).
X3. Alliance managers regularly report on the progress and
performance of their respective alliance(s).
X4. The company maintains a ‘repository’ or database containing
factual information of each of its alliances (e.g., date and
purpose of alliance formation, name of the alliance partner,
names of managers/executives who manage that alliance,
etc.).
X5. The company maintains a directory or ‘contact list’ of
individuals from within the company or outside who can
potentially provide inputs or assistance on alliance
management.
Knowledge codification
X6. Company managers follow a well-defined ‘process’ to guide Alliance Analyst (1996); Badaracco
the formation or management of any alliance. (1991); Winter (1987); Zollo and
X7. Resources such as checklists or guidelines are developed and Singh (2004)
used to assist managerial decision making and actions while
forming or managing strategic alliances.
X8. Resources such as alliance manuals (containing tools,
templates, or frameworks) are developed and used to assist
managerial decision making and/or actions while forming or
managing alliances.
X9. The company updates the alliance checklists, guidelines or
manuals that have been developed and are in use.
Knowledge sharing
X10. Company management conducts a ‘collective review’ to assess Glynn, Lant, and Milliken (1994);
the progress and performance of its strategic alliances. Huber (1991); Nonaka (1994);
X11. Alliance managers participate in forums such as committees or Seely Brown and Duguid (1991);
task forces to take stock of their alliance management Senge (1997)
experience and practices.
X12. Company managers participate in forums such as meetings,
seminars, or retreats to exchange alliance-related
information, experiences, war stories, etc.

Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj
1000 P. Kale and H. Singh

Item no. Construct items Theoretical reference

X13. Company managers engage in informal sharing and exchange

of alliance-related information and know-how with peers or
colleagues within the organization.
X14. Company managers with substantial prior experience in
managing alliances are usually rotated across some of the
company’s key alliances.
X15. Managerial incentives are used to encourage individual
managers to share their personal alliance management
experience and know-how with other managers within the
company.
Knowledge internalization
X16. Company managers attend ‘in-house’ training programs on Davenport and Prusak (1998); Nelson
‘alliance management’ whenever they are assigned to and Winter (1982); Nonaka (1994);
manage or work with any alliance. Teece et al. (1997)
X17. Company managers attend externally conducted training
programs on ‘alliance management’ whenever they are
assigned to manage or work with any alliance.
X18. The company provides opportunities for ‘on-the-job’ alliance
training to individuals who are relatively new to managing
alliances. Here, individuals are assigned to work in existing
alliances, especially with managers who have substantial
experience in managing such relationships.
X19. The company provides managers access to documented and
codified information and know-how on its prior and ongoing
alliance experience.

APPENDIX 1B: LIST OF ITEMS USED These individuals have been referred to as sec-
TO ASSESS ALLIANCE PERFORMANCE ondary respondents. Respondents used a seven-
point Likert-type scale to give their assessment on
items 1–4 and item 5, such that ‘1 = Strongly Dis-
The following items were used to assess the perfor- agree’ and ‘7 = Strongly Agree.’
mance of an individual alliance within each com-
pany. The alliance was assessed by the manager/
individual who was responsible for managing
the alliance or was associated closely with it.

Item Items for assessing the performance of each alliance Theoretical reference
no.

1. The alliance is characterized by a strong and harmonious Anderson (1990); Child and Yan (1999);
relationship between the alliance partners. Geringer and Hebert (1991); Inkpen and
2. The company has achieved its primary objective(s) in forming Crossan (1995); Saxton (1997); Khanna,
this alliance. Gulati and Nohria (1998); Kale, Singh, and
3. The company’s competitive position has been greatly enhanced Perlmutter (2000)
due to the alliance.
4. The company has been successful in learning some critical
skill(s) or capabilities from its alliance partner (s).
5. An overall assessment of this alliance, based on all the above
dimensions. Check (a) or (b):
(a) Satisfactory/Successful OR
(b) Unsatisfactory/Failure

Copyright  2007 John Wiley & Sons, Ltd. Strat. Mgmt. J., 28: 981–1000 (2007)
DOI: 10.1002/smj