Cahier du GRES 2008 - Cahiers du GRES

Groupement de Recherches Economiques et Sociales http://www.gres-so.org

Cahiers du GRES

Competitive intelligence and network mapping of interfirm alliances: strategic implications

Brigitte GAY LEREPS-GRES & Toulouse Business School [email protected]

Cahier n° 2008 - 05 Février 2008

Cahier du GRES 2008 - 05

Intelligence économique et cartographie des réseaux d’alliances Résumé L’intelligence économique doit aider les entreprises à comprendre leur environnement concurrentiel global, et à assurer la maîtrise de leur position dans de multiples secteurs. La cohérence de la stratégie d’une entreprise et la pertinence de sa position dans chacun des segments de ou des industries dans lesquels elle est impliquée, peuvent être analysées par cartographie des structures complexes de réseaux d’alliances qui caractérisent les industries aujourd’hui. Dans les secteurs où l’innovation est soutenue, souvent de rupture, et globale, ces outils permettent aussi de faire des audits technologiques des différents secteurs et d’analyser les interactions entre pays et les positions des nations elles-mêmes dans la sphère globale Mots-clé : Alliance, réseaux, cartographie, biotechnologie, compétition

Competitive intelligence and network mapping of interfirm alliances

Abstract In many industries, networks, rather than firms, have become the organizing level at which firms compete with each other. One role of competitive intelligence is to help firms understand their global environment as well as master their position in a number of industrial sectors. The pertinence of a firm strategy and position in the different segments it is involved in, can be analyzed by mapping the complex alliance networks that characterize industries today. These tools enable also a tech watch of individual, highly innovative, sectors as well as understand the links between countries and the positions of the nations themselves in the global environment. Keywords: Alliance, network, mapping, biotechnology, competition JEL : L14 ; O3


1. Introduction One of the most important trends in industrial organization of the past quarter century has been the growth of interfirm alliances. Indeed, since the early 1980s, the aims of most strategic alliances have been to gain access to new and complementary technologies, to speed up innovative and learning processes, and to improve the efficiency of particular activities such as R&D, supply-chain management, manufacturing, and marketing [Hagedoorn, 1993].It is therefore accepted today that alliances are not thought as trade-offs between perceived benefits of sharing risks and capital outlay on the one hand, and the costs of a loss of control associated with reduced or no ownership on the other [Dunning, 1995]. A recent survey found that alliances already account for anywhere from 6 percent to 15 percent of the market value of the typical company and that alliances are expected to account for 16 percent to 25 percent of median company value within five years and more than 40 percent of market value for almost one-quarter of companies. In current dollars, this means that for the advanced economies as a whole, alliances will represent somewhere between $25 trillion and $40 trillion in value within five years. Evidence of the growing emphasis on alliances and external collaborations as a route to success is highlighted by the fact that partnerships within the biopharmaceutical sector for example are currently being formed at the rate of $5 billion per year. In many industries, networks, rather than firms, have thus become the organizing level at which firms compete with each other. Strategy is therefore conceptualized today as a portfolio of links whereby position in wider networks is crucial to competitive advantage [Gulati and Zajack, 2000]. One role of competitive intelligence is to help firms understand their global environment as well as master their position in a number of industrial sectors. The pertinence of a firm strategy and position in the different segments it is involved in, can be analyzed by mapping the complex alliance networks that characterise industries today. In this article, we show how to quickly generate data and network maps essential to firms’ strategy on the basis of information available on Internet. Processing data as soon as they appear on the web is important, so as to be as close as possible to ‘real-time’ analysis of alliances and firm embeddedness in these complex meshes of interactions. By way of example, an analysis of the biotechnology industry is carried out from January 2004 to January 2006. Though often described as a single industry, biotechnology1 is a diverse field that impacts on several industries such as pharmaceutical research and manufacturing (including biopharmaceutical companies)2, tool developers (genomics, proteomics, transcriptomics, bioinformatics, combinatorial chemistry companies), companies developing advanced materials for human therapeutics, animal health, agribusiness (food, feed, fibres, transgenics), environmental bio-remediation and biodefense, industrial processes and efficiency. Biotechnology has also convergent applications with other technologies such as information technology, micro- and nanotechnologies, advanced materials and energy3. Therefore, within 1

Biotechnology is any technological application that uses biological systems, living organisms, or derivatives thereof, to make or modify products or processes for specific use (UN Convention on Biological Diversity). 2 Human therapeutics, diagnostics, drug delivery, cell and gene therapy devices and drug/device combinations 3 A survey of the use of biotechnology in U.S. industry, U.S. Department of Commerce Technology Administration and Bureau of Industry and Security publication, October 2003.

-3-


each industry, there are many sub-sectors defined by a complex mix of technology classes and application space, with continually changing opportunities for reconfiguration. Instead of only « creative destruction » (Schumpeter) of firms, old products, or industries, biotechnology is a case where ‘creative potential is exploited both in existing and in new sectors and firms”. It is thus more and more difficult to define industry boundaries or develop simple classification systems of firms engaged in biotechnology activities. Using visualization techniques as well as network metrics is thus interesting in the sense that they can determine that a network structure is not random, define different types of topologies, their dynamics, as well as address important questions such as: Have firms the capacity to manipulate the complex system, and hence the economic environment, into which they are situated? How does the macro network in turn influences their context and may provide benefits or constraints? In this paper we describe the use of mapping tools to follow, through network alliances, technical, financial, flows whether within or between sectors worldwide and thus begin to answer some of the above questions.

-4-


2. Background data The proliferation of alliances marks a shift in the conception of the intrinsic nature of competition, which is increasingly characterized by sustained multi-purpose technological change, the demands of innovation-led production, and fast entry in new and global markets. This has led to the notion that the key to success in coming years lies in the creation of collaborative advantage through strategic alliances4 [Das and Teng, 2000]. Importantly, as written by [Dunning J. H., 1995], “An asset-seeking alliance response does have implications for governance structures” … “the term alliance capitalism might be a more appropriate description of the features of innovation-led capitalism now spreading through the globalizing economy, than the term hierarchical capitalism”. Therefore, contrary to long-established views, contemporary organizations are increasingly built out of emergent linkages, linkages that are transient in that they are formed, maintained, broken, and reformed with considerable facility [Palmer, Friedland and Singh, 1986] [Monge and Contractor, 2003]. The firm has, within the network, the opportunity to pursue its idiosyncratic competencies and to complement others. Firms entering alliances become close in the network, affect their own specific governance structure but also the overall network structure and therefore other firms’ embeddedness and governance structure. Networks, rather than firms, become the organizing level at which firms compete with each other [GomesCasseres, 1994]. The broader network level structure establishes the extended resource endowments, whereas, at egonet5- and firm- levels, resource idiosyncrasy can be achieved. Firms are connected to each others in multiple networks of resources and influence or are influenced by information/knowledge flows derived from the structure to which they belong [Gulati, Nohria, and Zaheer, 2000]. Anand and Khanna [2000] provide compelling evidence for the existence of positive outcomes in managing many alliances. Firms forming many alliances extract more of the value created relative to their partners and are perceived by financial markets as more value creating. A firm position in the industry as well as its ego network will thus have a profound influence on its overall performance; hence the importance of networking capabilities. The network as a whole, therefore, permits elaborate constructions borne from complementary skills, that allow the handling of complex situations a firm can’t follow on its own as well as rapid adjustment to sustained change in highly competitive industries. Individual firm performance can be conceived only if the firm ‘fits into the network’ i.e. performs capably a missing, complementary, function. For Miles and Snow [1986], there is “symmetry between the characteristics and operations of the dynamic network and the features and behaviour of the firms within an industry (or major industry segment)”. Therefore, to the widespread agreement that most industries can contemporaneously support companies with different competitive strategies, is added another role firms have to play: that of implicit interdependence among competitors. Interdependence is needed not only for the firm to meet the dual objectives of innovation and performance, but also for the whole

4

An alliance is a formal agreement establishing an association or alliance between nations or other groups to achieve a particular aim. 5 In network analysis, a network is a set of actors connected by a set of ties. A single focal actor is called an “ego”. The set of actors ego has ties with are called “alters”. The ensemble of ego, his alters, and all ties among these (including those to ego) is called an ego network or egonet (Borgatti and Foster, 2003).

-5-


industry. As the industry, or major industry segment, evolves, so must the elaborated complementary constructions. Understanding the network dynamics that influence the formation of new alliances may help managers design alliances portfolio and therefore egonet structures that do not constrain firms’ future action. Managers that fathom out firm optimal positioning in a network may also derive possible control and information benefits. Who controls the bigger network and why, and possible limits and constraints of networks, are relevant issues. Historically, strategy and structure have evolved together. They need now to address the networking dimensions between organizations. The literature on industrial networks has rarely been approached from an economic perspective. Economists are still primarily concerned with the firm or the dyad as units of analysis, both being constitutive elements of a network. For example, many of neo-classical economists treat the market as their primary unit of analysis whereas evolutionary economists take the firm. Gulati and Zajac [2000] also underlined the fact that the approaches to studying alliances have paid little attention to the overarching networks in which firms are embedded. Most economists consider networks therefore as being simply the sum of interdependent dyadic relationships. In more recent years however, a significant new movement in network research has been witnessed, with the focus shifting from the study of rather small systems in the social sciences and methodology inherent to this field to the consideration of large scale statistical properties of graphs, such as path length and degree distributions, that characterize the structure and behaviour of large network systems. The body of theory in this field also aims to understand the meaning of these properties through models (see [Newman, 2003] for a review). We have already discussed some of these models and network metrics elsewhere [Gay B. and Dousset B., 2005]. In this paper, we mainly want to show the potential of data mining followed by simple visualisation techniques to decipher complex alliance networks. Therefore, in the following chapters, we describe the research methodology we used and analyse briefly the biotechnology industry, view some of its major sectors and firm egonets. The worldwide distribution of interfirm alliances is also looked at.

3. Research methodology A corpus on alliances between firms in the biotechnology industry was built quickly over period 01/01/2004-01/02/2006 starting from Internet sites (primarily Business Wire and PRNewswire). The sample thus formed covers approximately 5800 companies, public and private, but also leading american organizations (universities, government agencies) and capital investors. Network maps with actual nodes and links were drawn to address primary questions about network structure and dynamics by using TETRALOGIE network display program (IRIT, Institut de Recherche en Informatique, France) for large network visualization and analysis. A weighted spring embedder was employed to assign node locations, using an algorithm developed following the work of Fruchterman Reingold [1991], and Dousset [2003]. Spring embedders are based on the notion that the nodes may be thought of as pulling and pushing one another. The nodes that represent firms who are close will pull on each other, while those who are distant will push one another apart. The algorithm seeks to find an optimum in which there is minimal stress on the springs connecting the whole set of nodes. We are interested in both the process through which the structure unfolds, and thus a visualization technique that

-6-


displays the nature of change that leads to new states of an evolving network, and the cumulative structure of the network.

4. Results, analysis and discussion The complexity of graphs of alliances in the biotechnology industry (pharmaceutical sector), was already shown for period 1988-1999 using Bioscan, a directory of industries in the life sciences published by Bioworld Online [Powell W. W., White D. R., Koput K. W., and Owen-Smith J., 2005]. We have also quickly obtained from press releases on the Web a complex network map of interfirm alliances in the biotechnology industry for period 2004-january 2006 (Figure 1). Figure (1) Graph of business relationships in the biotechnology industry for the 2-year period. Each node in the network represents a company. Two companies are connected with a white line if they have announced a business partnership.

Starting from our data, we can look individually at each of the different sectors that form the whole industry and therefore try to reconstitute the full, complex, picture. For example, the analysis alone of three major segments (Antibody, Vaccines, Diagnostics) of the biotechnology industry explains approximately 30% of the alliances made over the period

-7-


studied. On the other hand, alliances in the nanotechnology sector applied to the bioindustry constitute 1.2% of the total activity. Figure (2) Illustration of firm-firm interaction network architecture in the antibody sector for the 2-year period, main component). To help visualisation, nodes are replaced on the graph by red bars, the size of which being proportional, for each node, to its number of business relationships.

The alliance activity in the Antibody and Diagnostics sectors is higher than in the vaccine sector: the size of the main component in the vaccine sector represents only 67% and 41% of that obtained for the Antibody and Diagnostics sectors respectively. As highlighted in Figures 2 and 3, each sub-network also possesses an individual structural signature that involves different firms with different types of linkage. We have already shown that innovation capacity of firms influences the growth and structure of the antibody sector; clear technological phases can be distinguished and their sequential importance to the field appraised, innovation being asymmetrically distributed in time [12]. It is possible as well to extract from the data base subnetworks such as business partnerships corresponding to different phases of a value chain. For example, in the pharmaceutical industry, 15% of linkages concern preclinical and clinical (phase I, II, and III) steps and hence describe all product flows. Manufacturing of products using advanced technologies concerns about 12% of the overall alliance activity. Important companies implicated in these different steps can be identified rapidly. For example, among companies involved in the development of novel production processes using biotechnologies over these two years, we find: Inverness Medical Innovations, Dow, Crucell et DSM Biologics, Lonza, Xoma, GTC Biotherapeutics et Laureate Pharma, Diosynth Biotechnology, and firms involved in cellular engineering like Morphotek or Chromos.

-8-


Figure (3) Illustration of firm-firm interaction network architecture in the vaccine sector (2-year period, main component).

Moreover, we are concerned by the alliance activity of individual firms in any given sector versus the whole industry. Indeed, central actors (in terms of alliance number) in a segment are not necessarily central actors in others. Network mapping of alliances for period 20042006, reveals that big pharmaceutical companies and other leading american organizations (universities, government agencies) dominate the overall network in terms of degree centrality (number of direct links), in agreement with the work of Powell et al (2005), while the analysis of three main sectors within this industry shows that, within sectors, major hubs display technical/scientific competences specific to the segment. Firms are also rarely central in more than one segment. Indeed, 8 of the 10 actors who signed more contracts in the whole industry are among the first 12 pharmaceutical companies in the world. Three large European companies, GlaxoSmithKline, Roche, and Astrazeneca -, and big American pharmaceutical firms such as Pfizer, Merck, Squibb Bristol-board-Myers, Abbott, and Wyeth, signed on average, by company and per annum, according to our web sources, 45 alliances over the whole period studied (Table 1). Conversely, when we look at individual sectors rather than the global industry, we find that the firms that sign more contracts are those which possess key intellectual property related to the sector (e.g. Genentech and Medarex in the Antibody sector) and/or technical/scientific competencies that are strictly limited to the field (e.g. Quest Diagnostics in the Diagnostic sector). We also observe that technologies with high added value in a given sector may move to other sectors that have reached maturity or face exponential growth. For example, Sequenom, which uses its Mass Array ® system for genotyping, or Affymetrix, with its DNA microarrays, both move to the Diagnostic sector.

-9-


Table (1) Main Actors forming alliances in the industry, or in different industry sectors, for the 2-year period Industry GlaxoSmithKline Pfizer Inc. Merck. National Institutes of Health (NIH) Roche Bristol-Myers Squibb Astrazeneca Abbott Genentech Wyeth

Antibody Genentech Medarex Morphosys Ag Biogen Idec

Vaccine NIH Merck GlaxoSmithKline Crucell

Medimmune Xoma Protein Design Labs

Chiron Aventis Medimmune

Seattle Genetics Centocor Abbott

Id Biomedical Dow Genvec

Diagnostics Roche Diagnostics Quest Diagnostics Abbott Laboratories Inverness Medical Innovations Qiagen Sequenom Inc. Imaging Diagnostic Systems Nanogen Celera Diagnostics Affymetrix

It is also useful to compare the business relationships of individual firms within industrial segments, the whole industry, on a yearly basis, or for the whole period analyzed. Figures 4 and 5, for example, show the egonets of two major pharmaceutical companies.

Figure (4) Pfizer (central node) alliance network in the pharmaceutical industry for the whole period.

- 10 -


Figure (5) Roche (central node) alliance network in the pharmaceutical industry for the whole period.

Global images are then simplified to allow comparison of firms’ respective investments in different sectors (Figures 6 and 7). Figure (6) Pfizer (central node) alliance network in the diagnostic sector

- 11 -


Figure (7) Roche (central node) alliance network in the diagnostic sector

To make a more rational analysis, it is important to also consider the equity investments a company holds in other firms. For example, Figure 5 underestimates the potential of Roche in the antibody sector, as demonstrated in Figure 8, if the alliance profile of at least one partner, Genentech, is not considered.

Figure (8) Genentech (central node) alliance network in the antibody sector

- 12 -


Likewise, how global a firm is can be rapidly assessed (Figure 9). Figure (9) World-wide alliance network of the Japanese pharmaceutical company Eisai in the whole industry for the 2-year period

Similarly, histograms and patterns of linkages between countries can be drawn (Figure 10 and 11). Figure (10) Differences in alliance numbers across countries in 2004 and 2005 25% 2004

2005

20%

% Alliances

15%

10%

5%

sweden

israel

austral

france

switzer

india

netherl

china

japan

uk

germany

canada

usa

0%

Pays

The United States, then Canada, Germany, the United Kingdom and Japan, head the group during the two consecutive years. China goes from the eleventh position to the sixth from

- 13 -


2004 to 2005. China is one of the rare countries to see its alliance number increase approximately two-fold over the period studied. Though China interacts with occidental countries, Figure 11 shows above all cohesive interactions with Asian countries, interactions being strongest with Japan, South Korea, and Taiwan. Figure (11) China main links to other countries

In a yearly analysis of the data (figures 12 and 13), the main connections between countries reflect the worldwide market, with a hard core formed by North America (strong bonds between the United States and Canada), Europe (strong bonds between two of the five most important European markets, Germany and the United Kingdom), Japan, and finally Switzerland. Figure (12) Main linkages between countries in 2004

- 14 -


American primacy (central position) during the two periods is not a surprise; it is still reinforced in 2005, in particular in Europe. Figure (13) Main linkages between countries in 2005

Notably also, Europe is well connected to Japan, but connections of European countries to Asia are made primarily through Japan or the United States. Linkages within the Triad USAEurope-Japan are more cohesive during the second period.

5. Conclusions We have made here a rapid analysis of alliances in the biotechnology industry over a two-year period. The analysis is not thorough and intends only to show the potential of mapping tools and simple visualisation techniques. We have therefore not gone into any details regarding the many different possible typologies of contracts or attributes of firms. Similarly, the data comes from 2 sources essentially, thus necessarily introducing bias in the analysis. We have shown chiefly that mapping of evolving networks starting from open sources found on the World Wide Web can bring important information such as: What are the different technical or product flows that occur through alliances, at which step in the value chain? In which sector have firms invested? How does that investment compare with that of competitors? With which countries does a firm interact? Are countries heavily involved in key sectors, how are they connected to other countries? Etc This analysis emphasizes the importance of continuous acquisition of technologies and products through alliances and how crucial worldwide interfirm but also inter-industry linkages are.

- 15 -


References Anand, B., N., Khanna, T., 2000, «Do firms learn to create value? The case of alliances», Strategic Management Journal, n° 213, p. 295-315. Das T.K., Teng B.S., 2000, «A resource-based theory of strategic alliances», Journal of Management, n° 26(1), p. 31-61. Dousset B., 2003, «Intégration de méthodes interactives de découverte de connaissances pour la veille stratégique», Toulouse, Habilitation à diriger des recherches, Université Paul Sabatier. Dunning J. H., 1995, «Reappraising the Eclectic Paradigm in an age of Alliance Capitalism», Journal of International Business Studies, n° 26(3), p. 461-491. Fruchterman T., Reingold E., 1991, «Graph Drawing by Force-Directed Placement», Software Practice and Experience, p. 1129-1164. Gay B., Dousset B., 2005, «Innovation and network structural dynamics: Study of the alliance network of a major sector of the biotechnology industry», Research Policy, n° 34(10), p. 1457-1475. Gomes-Casseres, B., 1994, «Group Versus Group: How Alliance Networks Compete», Harvard Business Review, Jul.-Aug., p. 62-74 Gulati, R., Zajack, E. J., 2000, Reflections on the study of strategic alliances, Cooperative Strategy: Economic, Business, and Organizational Issues, David Faulkner & Mare De Rond Eds., p. 365-374, Oxford Press: England. Gulati, R., Nohria, N., Zaheer, A., 2000, «Strategic networks», Strategic Management Journal, Special Issue, n° 213, p. 203-215. Hagedoorn J., 1993, «Understanding the rationale of strategic technology partnering: Interorganizational modes of cooperation and sectoral differences», Strategic Management Journal, n° 14, p. 371-385. Miles, R.E., Snow, C.C., 1986, «Network Organisations, New Concepts for New Forms», California Management Review, 28, p. 62-73. Monge, P.R. andContractor, N.S., 2003, Theories of communication networks, New York: Oxford University Press. Newman M. E. J., 2003, «The structure and function of complex networks», Society for Industrial and Applied Mathematics SIAM Review, n° 45, p. 167-256. Palmer, D., Friedland, R., Singh, J., 1986, «The Ties that Bind: Organizational and Class Bases of Stability in a Corporate Interlock Network», American Sociological Review, n° 51, p. 781-796. Powell W. W., White D. R., Koput K. W., Owen-Smith J., 2005, «Network Dynamics and Field Evolution: The Growth of Inter-Organizational Collaboration In the Life Sciences», American Journal of Sociology, n° 110(4), p. 1132-1205.

- 16 -

Cahiers du GRES Le GRES (Groupement de Recherche Economiques et Sociales) est un Groupement d’Intérêt Scientifique entre l’Université Montesquieu-Bordeaux IV et l’Université des Sciences Sociales Toulouse I. Il regroupe des chercheurs appartenant à plusieurs laboratoires : - GREThA - UMR CNRS 5113 (Groupe de Recherche en Economie Théorique et Appliquée), Université Montesquieu-Bordeaux IV ; - LEREPS - EA 790 (Laboratoire d'Etudes et de Recherche sur l'Economie, les Politiques et les Systèmes Sociaux), Université de Toulouse 1 Sciences Sociales; - L’UR 023 “Développement local urbain. Dynamiques et régulations”, IRD (Institut de Recherches pour le Développement) ; - Le laboratoire EGERIE (Economie et de Gestion des Espaces Ruraux, de l’Information et de l’Entreprise), ENITAB (Ecole Nationale des Ingénieurs des Travaux Agricoles de Bordeaux).

www.gres-so.org Cahiers du GRES (derniers numéros) 2007-17 : GALLIANO Danielle, SOULIE Nicolas, Organisational and spatial determinants of the multiunit firm: Evidence from the French industry 2007-18 : BROSSARD Olivier, VICENTE Jérôme, Cognitive & Relational Distance in Alliance Networks: Evidence on the Knowledge Value Chain in the European ICT Sector 2007-19 : OLTRA Vanessa, SAINT-JEAN Maïder, Incrementalism of environmental innovations versus paradigmatic change: a comparative study of the automotive and chemical industries 2007-20 : FRIGANT Vincent, Les fournisseurs automobiles après dix ans de modularité : une analyse de la hiérarchie mondiale et des performances individuelles 2007-21 : DOUAI Ali, Wealth, Well-being and Value(s): A Proposition of Structuring Concepts for a (real) Transdisciplinary Dialogue within Ecological Economics 2007-22 : THOMAS Olivier, L’inflation de la fiscalité locale en France : une conséquence inattendue de la promotion de l’intercommunalité ? 2008-01 : BERR Eric, Quel développement pour le 21ème siècle ? Réflexions autour du concept de soutenabilité du développement 2008-02 : NICET-CHENAF Dalila, Les accords de Barcelone permettent – ils une convergence de l’économie marocaine ? 2008-03 : CORIS Marie, Les délocalisations sous l’angle de la coordination : Une lecture par les catégories de la proximité appliquée au logiciel 2008-04 : DUPUY Claude, LAVIGNE Stéphanie, Investment behaviors of the key actors in capitalism: when geography matters 2008-05 : GAY Brigitte, Competitive intelligence and network mapping of interfirm alliances: strategic implications

La coordination scientifique des Cahiers du GRES est assurée par Alexandre MINDA (LEREPS) et Vincent FRIGANT (GREThA). La mise en page est assurée par Dominique REBOLLO.

- 17 -