The Cyclic Innovation Model - CiteSeerX

European Planning Studies Vol. 15, No. 2, February 2007

The Cyclic Innovation Model: A New Challenge for a Regional Approach to Innovation Systems? PATRICK VAN DER DUIN , ROLAND ORTT & MATTHIJS KOK Delft University of Technology, Section Technology, Strategy and Entrepreneurship, Delft, The Netherlands, Accenture, Amsterdam, The Netherlands

[Received April 2006; accepted September 2006]

ABSTRACT Innovation processes have changed significantly in the last four decades. Organizations no longer innovate on their own, aware that they need to decentralize their innovation activities and have to cooperate closely with other organizations in innovation systems. In this paper we discuss the spatial consequences of these developments, introducing the Cyclic Innovation Model (CIM) as a framework to analyse system innovation and applying it to the case of Thixomoulding, i.e. the development and exploitation of a revolutionary new material in the region of Flevoland, a province of the Netherlands.

Introduction Spatial or regional arrangements of innovation systems can be approached from two perspectives (Doloreux & Parto, 2005): innovation systems or regional science. In this article we focus on the perspective of systems of innovation by addressing the following research question: what does the Cyclic Innovation Model (CIM) mean for regional approaches to innovation systems? We illustrate the operation and mechanisms of CIM in a case study by applying it to the Thixomoulding innovation system. For this research we applied three research methods: (1) conducting interviews, (2) studying documents (desk research), and (3) carrying out participant observation by attending meetings and workshops. First, we provide a short historical overview of how innovation processes evolved to innovation systems and how this evolution influenced the spatial arrangement of innovative activities (second section). In short, we describe a historical shift from innovation in individual companies towards innovation systems consisting of different companies. Then,

Correspondence Address: Patrick van der Duin, Delft University of Technology, Section Technology, Strategy and Entrepreneurship, Jaffalaan 5, 2628 BX, Delft, The Netherlands. Email: [email protected] ISSN 0965-4313 print=ISSN 1469-5944 online=07=020195–21 DOI: 10.1080=09654310601078689

# 2007 Taylor & Francis

196

P. van der Duin et al.

we describe the CIM, an example of a so-called fourth-generation model (third section). In the fourth section we present a specific application of the CIM by describing the Thixomoulding case. We close this article in the final section by discussing how the CIM can be applied to analyse the spatial aspects of innovation systems and argue that regional innovation systems are a specific and temporal version of a spatial innovation system. Historical Development and Spatial Aspects of Innovation Generations In this section we provide a historical description of the development of innovation processes. It shows that the concept of the innovation system is the most modern one and that innovation systems (and innovation processes) should be adapted to the context in which they operate. Generations of Innovation Management After World War II several generations of innovation management can be distinguished. The so-called generations are descriptions “(. . .) of what constitutes the dominant model of best practice (. . .)” (Rothwell, 1994, p. 23). We will distinguish four generations. Table 1 provides an overview of the generations. The first column shows when a specific generation is thought to prevail. The second column describes the philosophy and main characteristics of each generation as well as its main disadvantages. The last column briefly describes the structure of the innovation process that prevailed during each generation. The first generation. After World War II innovation was primarily seen as science driven. In fact, corporate R&D labs are based on the notion that invention and innovation processes have to be integrated. These labs are organized like traditional universities, in mono-disciplinary departments (Roussel et al., 1991). They are regarded as large staff departments tied to and funded by the headquarters of organizations to grant them the freedom that is thought to be necessary to develop breakthrough technologies. In general, the structure of innovation processes is linear sequential and of a technology-push nature. In the course of the innovation process an innovation is managed by various departments that each contribute to the end product. An effect of the fact the demand exceeds supply is that innovation management could focus on technology rather than market needs. Both generous government funding (for industry and military purposes) and corporate growth allowed companies to focus on the kind of long-term research that was required for technological breakthroughs. Compared to earlier ideas about innovation management, the first generation of innovation management had several advantages. Although science-based corporate labs such as Bell Labs and the General Electric lab yielded many breakthrough technologies (Buderi, 2000; Chesbrough, 2003), in hindsight their approach also has significant disadvantages. Because an innovation would tend to move from department to department— and project management had yet to be introduced—it was not always clear who was responsible. This also implies that little attention is paid to the overall transformation process from idea to innovation. Scientific freedom of professionals seems more important than commercial relevance for the company. Innovation processes sometimes lack a strategic goal. Market needs and commercial aspects are incorporated late in the process, as a result of which commercial failures are discovered relatively late and much effort is put into unsuccessful innovation processes.

The Cyclic Innovation Model

197

Table 1. Short description of four generations of R&D managementa Period of the R&D management generations First generation (1950s –mid 1960s)

Second generation (mid-1960s– early 1970s)

Description of the generations Technology (science) push: The process of commercialization of technological change, i.e. the industrial innovation process, was generally perceived as a linear progression from scientific discovery, through technological developments in firms, to the marketplace. Because science is considered the starting point, R&D-institutes are structured like scientific institutions. Disadvantages: †Little attention is paid to the transformation process itself, or the role of the market place. †Scientific freedom is considered more important than the relevance and accountability of the research itself. †Innovation projects have no strategic goals, maybe short-term goals on the level of the project. There is no direct relationship with general management. †Commercial aspects are incorporated quite late in the innovation process. †Responsibility for the research is handed over from manager to manager (no project leader is appointed) and therefore final responsibility is not clear. †Professional project management practices are not applied. Market pull (need-pull): The market role is the source of innovations and the R&D organization merely has a reactive role. Because innovation processes are managed as projects, R&D-institutes are organized in a matrix. Disadvantages: †Neglects long-term R&D programmes and therefore leads to “incrementalism” only. †Projects are individual units, strategic relationships between these projects and corporate goals were not yet established. It was impossible to serve company goals that superseded the interests of separate internal clients.

Structure of the innovation processes Linear sequential process from department to department, starting with scientific discovery.

Linear sequential process in a project, starting with market need.

(Table continued)

198

P. van der Duin et al. Table 1 Continued

Period of the R&D management generations Third generation (early 1970s– mid-1980s)

Fourth generation (mid-1980s– early 2000s)

a

Description of the generations Market pull and technology push combined: Innovation is a process that, at each stage, enables interaction between technological capabilities and market needs. This interaction is generally facilitated by intra- and extra-organizational communication networks. These networks link R&D to in-house functions and link the firm to scientific and technological communities as well as to the marketplace. Innovation projects are considered to be part of a portfolio of projects. The goals of this portfolio are aligned with the corporate strategy. Disadvantages: †Focuses on product and process innovations rather than market and organizational innovations. †Focuses on the creation of innovations rather than the exploitation. †Focuses on evolutionary improvements rather than breakthroughs. R&D in alliances; Parallel and Integrated R&D, from R&D to New Business Development (NBD): R&D departments belongs to a network of internal departments and external organizations. R&D management means managing research links, networks and external research environments. Because of the number of actors involved, development processes are scheduled in parallel. Parallel development can significantly increase product development speed. The fourth generation is broader than the third since it includes business and market models that encompass the management of knowledge, technology, and market/ industry infrastructure. Future possibilities for improvement: †Increased networking and integration with internal and external partners. †Increased use of information technology to cooperate and communicate. †Increased flexibility of the structure of innovation processes.

Structure of the innovation processes Model of an essentially sequential process with feedback loops and interaction with market needs and state of the art technology at each stage.

Coordinated process of innovation in a network of partners. The required coordination is often attained by system integration (with key suppliers and customers) and parallel development (of components or modules of the innovation).

The description is based on Liyanage et al. (1999), Miller (2001), Niosi (1999), Rothwell (1994), and Roussel et al. (1991).


199

The second generation. The mid-1960s and the late 1970s required a different innovation management approach. Firstly, new market and societal conditions, such as increased levels of competition and increased awareness of ethical, social and environmental issues, requires consumer aspects to be taken into account earlier in the innovation process (Levitt, 1960; Sissors, 1960; Day, 1981). Consumer research rather than scientific research becomes the basis for new product ideas (Fornell & Menko, 1981). As a result, the interface between R&D and marketing becomes more important (Souder, 1988). Secondly, the scientific freedom and the accompanying lack of direct market results that characterized the previous generation are considered unacceptable in a situation of increased competition. Divisions start funding R&D, become internal clients and therefore demand that R&D efforts be more closely related to their business (Corcoran, 1994; Gupta & Wilemon, 1996). The focus on market results and efficiency also requires a stricter governance of innovation processes. A project management approach is adopted and a project leader, rather than subsequent managers of departments, is given final responsibility. Thirdly, R&D is separated because divisions are funding development while corporate headquarters continue to fund research. Fourthly, both the consumer orientation and the influence of divisions call for multidisciplinary cooperation in innovation processes (Liyanage et al., 1999). The multidisciplinary projects stimulate R&D institutes to abandon the functional structure and to adopt a matrix-based organizational structure. The innovation process remains essentially linear-sequential, albeit of a market-pull nature. In light of the context, the second generation approach of innovation management had many advantages: the market orientation reduced the risk associated with innovation and, together with the adoption of a project management approach, reduced the time-to-market of innovation processes. With hindsight we can also see that the consumer-oriented approach that prevailed between the mid-1960s and the late 1970s was not perfect, because potential consumers could not express their needs beyond those solved by familiar products (Bennet & Cooper, 1982; Tauber, 1974). As a result there was a focus on small improvements of existing products rather than on more radical innovations (Bennet & Cooper, 1982; Corcoran, 1994; Szakonyi, 1998). Another drawback of the second generation is that innovation projects are treated separately. As each project serves the goals of different internal company clients, relationships among the innovation projects and the strategic goals of company are hardly established.

The third generation. In the societal and organizational context between the late 1970s and the early 1990s innovation is increasingly seen as an investment that should make a relevant contribution to the business units’ strategies (Roussel et al., 1991). Innovation projects are also organized in programmes, to form a balanced portfolio of projects aimed at strengthening the strategic goals of these business units. Empirical studies indicate that technology-push and need-pull models of innovation were extreme, atypical examples of a more general process of interaction between technological capabilities and market needs (Mowery & Rosenberg, 1979). The structure of innovation processes remains essentially linear, but with feedback loops and constant interaction with market-related and technological factors (Rothwell, 1994). The concept of Quality Function Deployment, proposed by Hauser and Clausing (1988) is an example of an approach that combines technology and market knowledge. Companies

200


adopting third generation approaches try to find partners with essential technological and market knowledge. Together they form communication networks. The third generation of innovation management combines the strong points and remedies the weak points of the two previous generations. The main disadvantage of the third generation is a focus on the (end) product that is the outcome of the innovation process rather than the organizational and market changes that are required to introduce this product successfully in the market (Miller, 2001). Traditionally, R&D labs have no experience with organizational, market-related renewal, which used to be the domains of top managers and marketers, respectively. Involvement of top managers in R&D and close links between innovation processes and strategic company goals are important, since they facilitate the transfer of innovations from the R&D labs to the parent company. The fourth generation. During the fourth generation both invention and innovation processes increasingly involve the management of international alliances (Hagedoorn & Schakenraad, 1990). Gassmann and von Zedwitz (1999), for example, indicate that, from 1986 onwards, Swiss and Dutch companies had more labs abroad than at home because of the relatively small sizes of their domestic markets and a lack of R&D resources in their own country. Between the early 1990s and the start of the twenty-first century, innovation projects were no longer carried out in the isolation of R&D departments, but they were embedded in large networks with internal (other company departments) and external partners (universities, suppliers, customers, etc.). More specifically, the degree of integration of innovating companies with their suppliers and customers increases (Scott, 2001). In many markets, suppliers have become more capable and can therefore participate as partners in innovation processes, rather than merely produce pre-specified parts (Chesbrough, 2003). Developments in communication and information technology facilitate intra- and interorganizational cooperation in innovation processes (Howells, 1990). Exploitation (development?) has always been a bottleneck in innovation. The fourth generation pays more attention to market-related and organizational changes required for the successful introduction of product innovations in the market. The term innovation extends beyond product innovation to include process, organizational and market-related innovation (Trott, 2002) Business development becomes an integral part of innovation management (Chesbrough, 2003). In the fourth generation, innovation processes have become innovation systems. Spatial Aspects of the Innovation Management Generations Starting in the late nineteenth century, many companies set up R&D institutes that integrated basic research, applied research and development efforts (Bassala, 2001; Niosi, 1999; Chiesa, 2001). These institutes were usually built in the proximity of the headquarters and the production facilities of a company, resulting in a kind of central innovation system within a company. From the mid-1960s onwards, attention in society has focused on the demand-side of the market. Competition intensified because the growth targets of companies could no longer be achieved on the basis of the intrinsic growth of the market. Consequently, company strategies focused on growth and diversification to attain economies of scale and reduce


201

financial risks. Corporations developed divisional structures to meet the requirements of their diversified companies (Channon, 1973). In this context, many companies outsourced their basic research activities to universities or specialized research companies in the vicinity of their headquarters. Although the applied research activities and development efforts have remained an important element in innovation, in many companies they have become more separated. Research usually has remained a central corporate activity and development has re-allocated to different business units (Chiesa, 2001). As a result, within large multi-divisional corporations, central research units and multiple, dispersed innovation systems have emerged. Between the late 1970s and the early 1990s there was a period of decline, because of two oil crises, inflation and demand saturation. Companies started to focus on controlling and reducing costs rather than on realizing growth (Rothwell, 1994). The hierarchically-organized companies were transformed into more flexible companies, for example, by forming relatively independent business units and by outsourcing activities to suppliers or market actors providing complementary products and services. In this context, innovation efforts were closely coordinated in a network of companies. As a result, innovation systems of related organizations emerged around the business units of a company. Between the early 1990s and the early 2000s, technological developments in communication and information technology stimulated further globalization. Both globalization and the changes in communication and information technology influenced the organization and management of design, manufacturing, distribution and marketing processes. Globalization forces companies to focus on their core competences (Prahalad & Hamel, 1994) and has further increased the level of competition both in domestic and global markets (Gupta & Wilemon, 1996). In this context, companies transfer their production facilities to low cost countries. Communication and information technology enable more dispersed networks of organizations involved in innovation because of the ability to coordinate various activities at a distance. The innovation process has changed from a somewhat closed, singular innovation project to a continuous active and “open” innovation system (Chesbrough, 2003). In this system, the supply and demand for innovation sub-results, licenses, and so on even further disperse the network of companies involved in innovation. Dispersed international innovation systems emerge, in which the specific configuration depends on the type of innovation. Developing new software has become an almost global activity, as have some high-tech projects such as the development of aircraft (such as the Joint Strike fighter) or consumer electronics. Contextual Innovation The history of innovation processes and management is driven by overcoming the disadvantages of each former generation and by adjusting to changing societal and business contexts. It also shows that principles of innovation management of former generations are still applied (Griffin, 1997; Nessim et al., 1995). One can find different types of innovation management, even within a single firm (Van Gunsteren, 2003; Van den Elst et al., 2006; Verloop, 2006). So, management of innovation processes is often carried out in a contextual way; that is, it is adjusted to factors such as the type of industry, country or region in which a firm operates (external) or the type of product it produces or its organizational structure (internal). This means that innovation systems should also be

202


established in a contextual manner, in which the region and factors such as (specialized) the labour market force, subcontractor and supplier systems, supporting agencies, the presence of customers and users or local traditions for cooperative and entrepreneurial attitudes are contextually important (Doloreux & Parto, 2005, p. 136). In the following sections we discuss the CIM, an innovation model at system level, and show how this model is used in a specific way. Innovation Systems and the Cyclic Innovation Model (CIM) In this section we present the CIM (Berkhout, 2000), a fourth generation innovation model in which a distinction is made between four related innovation mechanisms: scientific exploration, technological research, product development and market transitions. These mechanisms form the basic building blocks of innovation systems. Since CIM can be considered an innovation system, we make some introductory comments on this subject in the next section. Innovation Systems As stated in the former section, the fourth generation of innovation management is the era of the innovation system. The essence of this generation is that firms no longer innovate on their own but open up their innovation processes for other firms (Chesbrough, 2003). Opening up the firm’s doors for innovation with other firms almost automatically means that innovation in general and innovation systems in particular become spatial or regional in character. According to Carlsson et al. (2002) there are many different types of innovation systems that can be viewed in several dimensions, such as a physical or geographical dimension, or can be characterized according to the amount of dynamics of the innovation system. The traditional input – output systems were static, while the more modern ones contain dynamic relationships. Nevertheless, they state that the function of every innovation system is “to generate, diffuse, and utilize technology” (Carlsson et al., 2002, p. 235). Also, according to Carlsson et al. (2002), each innovation system is made up of components (i.e. the operating parts of a system), relationships (i.e. the links between the components) and attributes (i.e. the properties of the components and the relationships between them). Regional innovation systems can be considered a subset of innovation systems. They received much academic interest in the 1990s.1 Considering the notion of contextual innovation (see previous section), To¨dtling and Trippl (2005) state that with regard to regional innovation policy there is no optimum regional policy approach. Every region should be approached differently. Best practices for regional policies are hard to find. The same holds true for innovation practices and management in general (i.e. the former generations of innovation management) (see, for instance, Van der Panne et al. 2004). Also, given the dynamic nature of most innovation systems, these systems are not constant but change over time (Carlsson et al., 2002). This might also be the reason why it is so difficult to find best practices that are valid for each type of innovation system (see also: Bessant et al., 2005). Furthermore, the fourth generation of innovation management (i.e. the innovation system approach) lacks the sophisticated models of first and second generation innovation management (for instance, product development and stage gate). Not surprisingly, nowadays many companies still apply intuitive and informal ways of innovating (Griffin, 1997; Nessim et al.,


203

1995). We consider the CIM (see previous section) as a serious attempt to rectify this omission.

Principle of Cyclical Interaction In this section and in the following two sections we give a summary of the CIM. We emphasize that this model is meant to visualize the complexity of an innovation system and to show which actors are important for enhancing the performance of an innovation system. CIM is a process model that can be used by actors within the system to establish an innovation system and to keep it running. An important element of the CIM is the principle of cyclic interaction. By adding feedback paths, models of dynamic systems—often referred to as regimes—are represented by interconnected cycles and, as a result, (work) processes become cyclical.2 It presents the basis for modern control and is the precondition for operating in a flexible manner; it also presents the inspiration for human creativity and is a necessary condition for sustainability. Thanks to feedback, (human) actors are constantly reminded of the consequences of their actions, preferably through built-in “early signals”, allowing companies to adjust quickly to unforeseen developments. The cyclical architecture also makes it possible to learn from mistakes, a very important property of innovation. Figure 1 illustrates the basic principle. A represents a subsystem that maintains a cyclical interaction with subsystem B. Examples are interactions between technical products and their users, commercial organizations and their customers, hospitals and their patients, etc.

Dynamics Around Technological Development Figure 2 shows two linked basic units (the double loop) in which technological research plays a central role. The cyclical interaction processes for new technological developments take place in the so-called technology-oriented sciences cycle (the left-hand side of Figure 2), with the help of a wide range of disciplines from the hard sciences. Technological research in this cycle is a multidisciplinary activity: a variety of disciplines from

Figure 1. Cyclical interaction is the basis for modern control and a precondition for operational flexibility. It is also the inspiration for creativity and a necessary condition for sustainability

204


Figure 2. The dynamics surrounding technological research—changes in the demand and supply of new methods and tools—is driven by the cyclical interaction between new scientific insights into technical processes (left-hand side) and new functional requirements for process –product combinations (right-hand side)

the hard sciences is needed to develop a new technology (many-to-one-relationship). Similarly, the cyclical interaction processes for new product development take place in the integrated engineering cycle (the right-hand side of Figure 2). Modern product development is a multi-technology activity: a variety of—often patented—technologies is needed to develop a new product (many-to-one relationship). As is the case in multidisciplinary science, a wide range of skills is needed to be successful. Experience shows that the skills of specialized suppliers play an important role in making the engineering process successful. Figure 2 shows that, in the sciences cycle, technological research is driven by new scientific insights: “science push”. It also shows that, in the engineering cycle, technological research is driven by new functional requirements in product development: “function pull”. The dynamics in technological research are therefore driven both by new scientific insights and by new functional requirements. In technological research, scientists and engineers must inspire one another constantly, which means that research has to be organized differently. We should emphasize that we use the term “product” in its broadest possible sense: everything we design and make. Hence, it includes intangible things such as databases, computer software, financial instruments, governmental regulations, governance models, etc. In addition, we also use the term “technology” in its broadest sense: knowledge—both implicit and explicit—on how to design and make products. This broad use of the concepts of technology and product is characteristic of the fourth generation innovation concept (Van der Duin et al., 2005). Dynamics Around Market Transitions Figure 3 also shows two linked cycles, although here the world of human needs rather than the central role played by technology. The cyclical interaction processes for the development of new insights into emerging and receding socio-economic trends (market transitions) take place in the so-called social-oriented sciences cycle (left-hand side of Figure 3), with the help of a wide range of different disciplines from the social sciences. These insights make it possible to develop new socio-technical solutions more quickly and with less


205

Figure 3. The dynamics surrounding market transitions (changes in demand for and supply of sociotechnical solutions) is driven by the cyclical interaction between new scientific insights in socioeconomic processes (left-hand side) and new investments in product – service combinations (right-hand side)

economic risk. The successful anticipation of market transitions is very much a multidisciplinary activity: a variety of disciplines from the social sciences is needed to explain and predict market transitions (futures research) in a scientific manner and to establish new solutions based on the underlying socio-economic forces (many-to-one-relationship). Similarly, the cyclical interaction processes needed to provide a changing society with new product – service combinations take place in the differentiated service cycle (right-hand side of Figure 3). Experience shows that in this cycle early users play an important role in making the innovation process successful (Von Hippel, 1986): using the creativity of customers. It is interesting to note that in recent years the services sector has expanded considerably, not only because consumer demand has grown but also because companies have outsourced many of their support processes. In the sciences cycle, market transitions are seen as a dynamic socio-economic process in which changing demand for product –service combinations is determined by the dynamics in the needs and concerns of society. However, in the service cycle market, transitions are seen as dynamic commercial processes in which changes in the supply of product –service combinations are determined by the innovative capacity of the business community. In an innovation economy both scientific insight into changing demand (lefthand side of Figure 3) and commercial investment in changing supply (right-hand side of Figure 3) should be constantly inspiring and reinforcing one another. Model of the Innovation Arena When we compare Figures 2 and 3, the dual nature of scientific exploration and product development becomes clear: science has both hard and soft aspects and product development has both technical and social aspects. Figure 4 is a combination of Figures 2 and 3. The result is a systems view of the cyclical change processes—and their interactions—as they take place in a successful innovation arena: hard and social sciences as well as engineering and commercialization are brought together in a cohesive system of creative processes. Entrepreneurship plays a central role. Without entrepreneurship there is no innovation. The first thing that stands out in Figure 4 is that the architecture is not that of a chain but of a circle: innovations build on innovations. Ideas create new developments, successes create

206


Figure 4. System model showing the foundations of the innovation economy, a circle of mutually influencing dynamic processes: “Circle of Change”. In the model, changes in science (left) and industry (right) and changes in technology (top) and markets (bottom) are cyclically interconnected, and entrepreneurship plays a central role

new challenges and failures create new insights, resulting in a constantly accumulating creation of value. New macro-economic instruments are needed to preserve the strength of the dynamics in the circle. Large-scale failures, such as the recent dotcom debacle, undermine confidence in the innovation economy, making people reluctant to invest. In terms of CIM, the processes of change are disconnected. The economy enters a phase of stagnation in which companies focus on producing more of the same at lower cost (“lifecycle management”) until confidence returns and investment capital once more becomes available to spur innovation, causing the dynamics in the circle to accelerate again. The situation outlined earlier leads to a system of linked cycles that in turn influence each other. The result is a more or less synchronized regime of interconnected dynamic processes that spark a creative interaction between changes in science (left-hand side) and industry (right-hand side) and between changes in technology (top) and market (bottom). In a successful innovation economy there will be few barriers between nodes and cycles: institutions and organizational structures facilitate the change processes (Volberda, 1998). Throughout the circle there is a continuous exchange of ideas and concepts, of knowledge and information, of capital and labour, of products and services and of technical and socio-economic values. To sum up, CIM views innovation processes as continuous interactions between developments and changes in markets, product and services, technology, and science. Innovative companies manage to integrate all of these four aspects and “play chess on four levels at the same time” (Berkhout et al., 2006). CIM portrays innovation as a system of dynamic processes—Circle of Change—with four creative “nodes of change”: scientific exploration, technological research, product development and market transitions. More importantly however, between these nodes there are “cycles of change” that ensure that the


207

dynamic processes in the nodes influence each other; in other words, they inspire, correct and supplement each other. CIM and Regional Innovation The CIM shows that an innovation system is complex. It combines science with business and technology with markets. This means that four different cultural environments need be interconnected. For regions to be innovative, it is not sufficient that all environments are present and act well, it is essential that they function as an integrated multicultural system. Until today, we have observed two types of system errors. Type one refers to a region that may be excellent in scientific research, but still underperforms economically because of a valorization barrier between the science and industry communities (the left- and righthand sides in Figure 4 make their own choices and plans). Type two refers to a region that may be excellent in designing and developing technical functions, but still underperforms economically because of a valorization barrier between the technical-oriented and marketoriented communities (the upper and lower parts in Figure 4 make their own choices and plans). The failure of the European Union’s (EU’s) Lisbon strategy has to be seen above all as a consequence of the existence of these systems errors in the European innovation system. The huge emphasis on more research and more technology is a far too one-sided, simplistic approach. The cyclic process model shows that the real problem lies elsewhere. The Thixomoulding Case In this section we apply the CIM to the development and exploitation of thixomoulding (see Table 2). CIM has been used by the different actors involved with thixomoulding to build up an innovation system. This case study is meant to illustrate the workings and principles of CIM. As such, CIM has been used both to prescribe and describe the workings of the Thixomoulding innovation system: the actors, their positions and the interconnections. Although more than 100 thixomoulding machines are in operation world-wide, thus far none of them are in Europe. The reason for this is that until now thixomoulding was used Table 2. The properties of Thixomoulding Magnesium is one of the most abundantly available materials on earth. Almost any country in the world and the Netherlands in particular possess enormous stocks of magnesium salts. Magnesium is the lightest construction material that keeps its rigidity characteristics. The specific mass of magnesium is about 30% lower than aluminium. A well-known disadvantage of magnesium is its sensitivity to oxidation. That is why in the thixomoulding process magnesium is being alloyed with aluminium and zinc. At certain temperatures this alloy becomes doughy, which makes it suitable for injection moulding. Thixomoulding is a combination of various existing technologies, among others injection moulding, mould production and finishing. The name Thixo is derived from the chemical term thixotrope. Thixotropy involves the behaviour of liquids whose viscosity decreases as a result of shift tensions. It becomes liquid when it is moved. A well-known example is ketchup. With thixomoulding the result is a unique, tough molecular structure with a high density, without inclosing air. Because there is little shrinkage when it cools off, the discharge angles (the angles between the matrix wall and the direction of the discharge of the product) are made redundant. As a result, less finishing is needed and, unlike plastics, the product is easier to shape.

208


mainly in assembling products in the electronics industry, an industry that is predominantly present in the US and Asia. In 2000, a chemical engineer and specialist on injection moulding production wrote a business plan for establishing a thixomoulding plant in the Netherlands. He contacted an advisor at Syntens in the Dutch province of Flevoland. Syntens is an independently operating foundation subsidized by the Dutch Ministry of Economic Affairs, whose mission is to stimulate and promote innovation among small and medium-sized enterprises (SMEs) in the country. Because of the complexity and the long time horizon of this project, Syntens advised cooperating with other organizations in developing innovations on the basis of thixomoulding. As a result, Flevobike Technology en Artiplast (both located in Flevoland) became involved in the project. Flevobike has a great deal of knowledge about bicycle technology, a specialist in developing prototypes in this field. For instance, the company was one of the originators of the recumbent bike. For Flevobike, thixomoulding means new opportunities for designing and building new bikes and parts of bikes. In light of the company’s huge experience with innovation, Flevobike is a logical addition to this project. Artiplast is an injection moulding company that also produces moulds. Since, at a conceptual level, thixomoulding is not that different from injection moulding, Artiplast’s knowledge and experience was a welcome contribution to the project. These three parties formed a strategic alliance called Thixotech Europe. Thixotech Europe is the first holder of a license for thixomoulding. Its goal is to establish a thixomoulding plant. They realize that, if they are to achieve their goal, they will need to work together with other organizations, government institutions and research institutes. They wanted to form an innovation system (or cluster) with these organizations. To realize this they founded MagTech Flevoland. Together with Syntens, the Flevoland Development Organization (de Ontwikkelingsmaatschappij Flevoland (OMFL)) and the province of Flevoland, MagTech will have a leading role in developing the thixomoulding innovation system. After the initial contacts were established, it became clear that what was needed to take the project to a higher level was a type of structural framework. Flevobike suggested using CIM. It was decided to invite scientists from Delft University of Technology to apply this model to the thixomoulding innovation system and to offer advice based on the findings of the application of this model. Consequently, CIM was used both in a descriptive and prescriptive role in the thixomoulding innovation system. The innovation system was built and analysed in three phases with the support of scientists of Delft University of Technology, according to the main principles of CIM. Phase 1: Analysis of the Starting Position First, all actors involved were placed in CIM, in close contact with the original group of organizations involved. In doing so, the following questions were addressed: Are one or more cycles underused (or even empty)? Are one or more cycles over-used? Are the cycles interrelated to each other? This is shown in Figure 5, which indicates that the relationships between most of the actors have not (yet) been made explicit and formalized, which does not promote the transfer of knowledge. The actors are unaware of their positions in the innovation system and their mutual relationships. By establishing these relationships it becomes clear what their knowledge output is, what knowledge input they need and which organizations are involved. In the current situation, the innovation system is (too) technologically oriented, which is hardly surprising considering the fact


209

Figure 5. The thixomoulding innovation system structured according to CIM

that thixomoulding technology was the starting point for the development of the thixomoulding innovation system. Phase 2: Analysis of the Desired Position What should the thixomoulding innovation system look like? To answer this question we look at the dynamics surrounding technology development, which consists of the technology-oriented sciences cycle and the integrated engineering cycle (Figure 6). Because showing all relationships in this part of CIM go too far, we focus on one of them, prototyping. In an ideal situation, the designers (together with the engineers) combine the design and the technical possibilities. Technical design is then connected to industrial design, after which prototyping becomes highly dependent on matrix construction, with which it is difficult to experiment, due to the high investment costs involved. Building a mould is an integral part of the design, because there is no room to make design mistakes. However, since building moulds is not a very innovative industry in the Netherlands, it is necessary to incorporate knowledge about mechanical engineering. Next, we look at the dynamics around market transitions, which consists of the socialoriented sciences cycle and the differentiated service cycle (not included here). The goal is to realize product innovation with high added value. Added value is related to the value a customer attaches to the product. This emotional value is, furthermore, related to the image and design of the product. Image is mainly the territory of public relations and marketing. As far as the thixomoulding innovation system is concerned, it is important that the first products do not get a bad image as a result of technical failure. Design is

210


Figure 6. The thixomoulding innovation system structured according to CIM

related to the aesthetic aspects, the ergonomics and the usability of the product. Together with the technical design, this is the scientific domain of industrial design. The thixomoulding innovation system can apply future research to identify potential markets for their products. This will become more important when (in the longer term) European competitors enter the market, which will make it necessary to develop more radical innovations. In short, phase 2 illustrates that the thixomoulding innovation system consists of four different networks that need to work together closely. Phase 3: The Transition Path To get from the current situation to the desired (future) one, the thixomoulding innovation system has to follow a transition path. This transition path has three aspects. First, during the process it will become apparent that certain types of knowledge will be lacking, which means that it is difficult to decide in advance which organizations need to be involved in the thixomoulding innovation system. Secondly, relationships within a network or system do not form overnight but need some time to develop. Long-term relationships are the result of a mutual trust that is reinforced by repeated innovation successes. Network management is something with which the organization and leadership of the innovation system has to become familiar. Thirdly, from a business perspective, limits to capital and investment in knowledge development slow down the growth of the innovation system. In a successful innovation cycle, new products arrive at the right time in the life cycle. A successful life cycle will finance the innovation cycle. The innovation system needs to make sure that they feed the life cycle quickly enough with new and


211

good products. Because this cannot be done currently with radical innovations, which would simply take too long, a strategy has been formulated for the immediate future as well as the mid-term and long-term future. A Regional Perspective on Thixomoulding The Thixomoulding innovation system comprises a small number of complementary organizations, all of which are regionally based. This spatial arrangement was not the result of a deliberate decision but emerged quite coincidentally. Most of the originators of the thixomoulding innovation system simply happen to live in this part of the Netherlands. However, when the province of Flevoland and Syntens became involved, they aligned the Thixomoulding innovation system by providing funds and other kinds of support. Syntens was suffering from institutional barriers, since its advisors were not allowed to spend more than 12 hours consulting on this innovation system. The question remains at what spatial level the Thixomoulding innovation system would operate in the future. We believe that the current regional scale (i.e. the province of Flevoland) is a temporal one. In searching for new applications, new partners and new expertise, the innovation system is forced to look outside its immediate geographical region. The use of CIM as a support to its innovation system is in itself an indication that its future will not be entirely in Flevoland. Thixomoulding can be considered a basic technology that is useful to many different industries that may be located elsewhere in the Netherlands or abroad. So, the evolution of the thixomoulding innovation system will not only touch the type of organization involved or the various applications in which thixomoulding can be used, but their geographical aspects as well. As far as using CIM as a supporting framework this has the following consequences: . The type of leadership will be different. If the Thixomoulding innovation system is reaching a higher conceptual and geographical level, its leadership should move in the same direction, which means that the Flevoland-based leadership will be replaced by a higher authority, or at least one that is capable of operating on a national rather than provincial level. . Cyclical interactions between various actors become less face-to-face and more (long) distance and anonymous. Although this can be facilitated by modern information and communication technology, it also demands a more formalized approach to knowledge-exchange flows, knowledge-protection agreements within the system and arrangements with regard to (possible) revenues, investments and operational costs. . A greater emphasis on the development of potential applications of thixomoulding, e.g. new types of bike, means that interaction with customers becomes much more important. If, for instance, thixomoulding-based bikes were to be developed for bike couriers, there would have to be some market research and testing in large cities (e.g. Amsterdam, London, Paris). A consequence of adopting the lead-user concept by Von Hippel (1986) in the development of thixomoulding in general is that there will be an emphasis on the market transitions node. Taken to its extreme, this would mean that the thixomoulding factory would be merely a production facility producing the main parts, while the assembly and incorporation of new parts of the bike would be decentralized, based on the geographical (or spatial) needs and characteristics of users. The thixomoulding plant might regain some of its position by focusing not only on the mass production

212


of the large parts of the bike (e.g. the frame), but by playing the role of a knowledge centre as well. As such, the thixomoulding plant could combine an operational and a strategic task in the (national or maybe European) thixomoulding innovation system. As a result, the thixomoulding innovation system would have more spatial levels. There would not only be a regional level (i.e. the province of Flevoland), but also a local (i.e. local contacts with users to gain information about the needs and uses of thixomoulding-based products) and a national level (i.e. national innovation policies to promote the development and application of thixomoulding-based products, leading even to specific thixomoulding-innovation systems related to existing industries). Conclusions In this final section we first formulate conclusions with regard to CIM, based on the case study of the Thixomoulding innovation system. We then look at it from a regional or spatial perspective.

Conclusions and Discussion about CIM CIM turned out to be a useful descriptive instrument for analysing knowledge relationships within the Thixomoulding innovation system, as expressed by interviewees. CIM gives the researcher insight into the complex relationships between science, technology, product development and market transitions, without having the need for much knowledge about the technology involved. CIM is a model that enables a learning process and is an innovation system model at the same time. A normative application of CIM confirms this: by gathering and mapping all actors and relationships, the missing actors and knowledge become visible. The process of filling in all the white spots in CIM is ongoing and is dependent on the chosen strategy (for example, the choice of which product – market combinations to develop), which is decisive for the direction of the innovation system. The strategy of the Thixomoulding innovation system is currently rather broad and aimed at the development of the organizational elements of the innovation system than on the content of the business it is developing. However, if CIM is used in a normative way, it means that the cluster has to be filled as completely as possible. From a practical viewpoint this is not very wise. First, the rotation (or circulation) speed hampers this. Actors within the Thixomoulding innovation system should not immediately expect a high rotation speed. This requires a mature and complete organization, as well as experienced managers. These elements are not currently present in the Thixomoulding innovation system. Therefore, the actors have opted initially to having a technology bias. Successful cooperation in the engineering cycle might lead to the system becoming broader. Secondly, making the dynamics of the system more manageable is an important reason for having strategic alliances. Changes in the market are less dynamic in some industries than changes in technology. In that case, an alliance with market actors is not necessary. An analysis by CIM would mean emphasizing the engineering cycle. In this case, this would not result in a system failure. Especially the omission of a cycle could make the organization less vulnerable for management failures. However, this does not mean that, in the long term, this cycle should and will be excluded. From a more theoretical view, CIM would benefit from linking it with other business theories to enable CIM to be used in a more operational way. None of the actors were


213

happy with the structure CIM gave to me, but for their daily work as innovators CIM was more difficult to apply. So, more practical guidelines for using and implementing CIM would be very welcome.

Discussion of CIM in Relation to Regional Aspects We close this article with three general comments on CIM, innovation systems in general and their spatial and regional aspects in relationship to CIM, by comparing some of the findings of the thixomoulding case with literature on (regional) innovation systems: . Carlsson et al. (2002) discuss the various aspects of innovation systems in terms of their specific structures, levels of detail and types of internal interaction. For our analysis, it is important to see the physical or geographical dimension as an important dimension of an innovation system. Furthermore, innovation systems are not static entities; they evolve over time. The situation we described with regard to the thixomoulding innovation system reflects an innovation system that at the moment is focused on the Flevoland-region. However, when we take into account the innovation system’s ambitions, we expect that in the near future it will start looking beyond its current provincial borders. According to Carlsson et al., this is a natural development because, “the function of an innovation system is to generate, diffuse, and utilize technology” (p. 235). In this case, it is especially the diffusing and utilization of the technology that will drive the geographical expansion. . CIM places much emphasis on developing the capability to exchange information and knowledge, whereas “traditional” innovation models pay more attention to how individual organizations can improve their innovativeness. More precisely, CIM claims that the viability of an innovation system is determined by the flexibility and speed by which actors in the four nodes exchange information and knowledge. This is confirmed by an analysis conducted by Ronde and Hussler (2005) into the determinants of regional innovative levels in French manufacturing industries: “. . . relational competences are an important channel for knowledge spillovers. Without those specific interactions the impact of geographical spillovers is reduced” (p. 1163). . Fromhold-Eisebith and Eisebith (2005) distinguish between innovative innovation systems that are implemented top-down and bottom-up. They argue that the best way to implement an innovation depends on dimensions such as the geographical scale of an innovation system initiative, regional structural preconditions, life cycle stages and sector orientation (p. 1265). As stated in an earlier section, the thixomoulding innovation system was initiated bottom-up, after which regional governments provided topdown support. As far as the regional dimension is concerned, we feel that the top-down support was motivated mainly by a desire to keep the economic activity around the thixomoulding technology within Flevoland, a desire that is decidedly less prominent among the originators of the innovation system. Their ambition is to make what they are doing into a success, even if that means having to do business with organizations outside their own region. Based on this, a top-down approach to supporting innovation systems tends to be more regionally based than initiatives that adopt a more bottom-up approach.

214


Acknowledgement The authors would like to express their gratitude to Professor A. J. (Guus) Berkhout for his valuable comments and ideas. Notes 1. For an overview of research issues around regional innovation systems, see Doloreux and Parto (2005). 2. A cycle consists of a succession of connected processes that take place repeatedly, each time with new starting conditions and a shifting context. The dynamics in a cycle are determined by “cycle time” and “change per cycle”.

References Bassala, G. (2001) The Evolution of Technology (Cambridge: Cambridge University Press). Bennett, R. C. & Cooper, R. G. (1982) The misuse of marketing: An american tragedy, Business Horizons, 25(2), pp. 51 –61. Berkhout, A. J. (2000) The Dynamic Role of Knowledge in Innovation. An Integrated Framework of Cyclic Networks for the Assessment of Technological Change and Sustainable Growth (Delft: Delft University Press). Berkhout, A. J. & Van der Duin, P. A. (2006) New ways of innovation: An application of the cyclic innovation model to the mobile telecom industry, International Journal of Technology Management, forthcoming. Berkhout, A. J., Van der Duin, P. A., Hartmann, L. & Ortt, J. R. (2006) Innovating the innovation process, International Journal of Technology Management, 34(3/4), pp. 390– 404. Bessant, J., Lamming, R., Noke, H. & Philips, W. (2005) Managing innovation beyond the steady state, Technovation, 25, pp. 1366–1376. Buderi, R. (2000) Engines of Tomorrow (New York: Simon & Schuster). Carlsson, B., Jacobsson, S., Holme´n, M. & Rickne, A. (2002) Innovation systems: Analytical and methodological issues, Research Policy, 31, pp. 233 –245. Channon, D. F. (1973) The Strategy and Structure of British Enterprises (London: MacMillan). Chesbrough, H. W. (2003) Open innovation. The New Imperative for Creating and Profiting from New Technology (Boston, MA: Harvard Business School Press). Chiesa, V. (2001) R&D Strategy and Organisation. Managing Technical Change in Dynamic Contexts (London: Imperial College Press). Corcoran, E. (1994) The changing role of US corporate research labs, Research Technology Management, July– August, pp. 14– 20. Day, G. S. (1981) Strategic market analysis and definition: An integrated approach, Strategic Management Journal, 2, pp. 281–299. Doloreux, D. & Parto, S. (2005) Regional innovation systems: Current discourse and unresolved issues, Technology in Society, 27, pp. 133–153. Fornell, C. & Menko, R. D. (1981) Problem analysis: A consumer-based methodology for the discovery of new product ideas, European Journal of Marketing, 15(5), pp. 61–72. Fromhold-Eisebith, M. & Eisebith, G. (2005) How to institutionalize innovative clusters? Comparing explicit topdown and implicit bottom-up approaches, Research Policy, 34, pp. 1250–1268. Gassmann, O. & von Zedtwitz, M. (1999) New concepts and trends in international R&D organization, Research Policy, 28, pp. 231–250. Griffin, A. (1997) PDMA research on new product development practices: Updating trends and benchmarking best practices, Journal of Product Innovation Management, 14(6), pp. 429–458. Gupta, A. K. & Wilemon, D. (1996) Changing patterns in industrial R&D management, Journal of Product Innovation Management, 13, pp. 497 –511. Hagedoorn, J. & Schakenraad, J. (1990) Leading Companies and the Structure of Strategic Alliances in Core Technologies (Maastricht: MERIT Working Paper). Hauser, J. R. & Clausing, D. (1988) The house of quality, Harvard Business Review, 66(3), pp. 63 –73. Howells, J. (1990) The location and organisation of research and development: New horizons, Research Policy, 19, pp. 133–146.


215

Levitt, T. (1960) Marketing myopia, Harvard Business Review, July–August, pp. 45–56. Liyanage, S., Greenfield, P. F. & Don, R. (1999) Towards a fourth generation R&D management model— research networks in knowledge management, International Journal of Technology Management, 18(3/ 4), pp. 372 –393. Miller, W. L. (2001) Innovation for business growth, Research Technology Management, September–October, pp. 26 –41. Mowery, D. & Rosenberg, N. (1979) The influence of market demand upon innovation: A critical review of some recent empirical studies, Research Policy, 8, pp. 102–153. Nessim, H., Ayers, D. J., Ridnour, R. E. & Gordon, G. L. (1995) New product development practices in consumer versus business products organizations, Journal of Product and Brand Management, 4(1), pp. 33– 55. Niosi, J. (1999) Fourth-generation R&D: From linear models to flexible innovation, Journal of Business Research, 45, pp. 111–117. Prahalad, C. K. & Hamel, G. (1994) Strategy: The search for new paradigms, Strategic Management Journal, Summer Special Issue, p. 11. Ronde. P. & Hussler, C. (2005) Innovation in regions: What does really matter?, Research Policy, 34, pp. 1150– 1172. Rothwell, R. (1994) Towards the fifth-generation innovation process, International Marketing Review, 11(1), pp. 7 –31. Roussel, P. A., Saad, K. M. & Erickson, T. J. (1991) Third Generation R&D. Managing the Link to Corporate Strategy (London: Arthur D. Little). Scott, G. (2001) Strategic planning for high-tech product development, Technology Analysis & Strategic Management, 13(3), pp. 343–364. Sissors, J. Z. (1960) What is a market?, Journal of Marketing, 30, pp. 17–21. Souder, W. E. (1988) Managing relations between R&D and marketing in new product development projects, Journal of Product Innovation Management, 5(1), pp. 6–19. Szakonyi, R. (1998) Leading R&D: How much progress in 10 years?, Research Technology Management, November–December, pp. 25–29. Tauber, E. M. (1974) How market research discourages major innovation, Business Horizons, 17(3), pp. 22–26. To¨dtling, F. & Trippl, M. (2005) One size fits all? Towards a differentiated regional innovation policy approach, Research Policy, 34, pp. 1203–1219. Trott, P. (2002) Innovation Management and New Product Development (Englewood Cliffs, NJ: Prentice Hall). Van der Duin, P. A., Ortt, J. R., Hartmann, L. & Berkhout, A. J. (2005) Innovation in context. From R&D management to innovation in networks, in: R. Verburg, J. R. Ortt & W. Dicken (Eds) Managing Technology and Innovation: An Introduction, pp. 227–247 (London/New York: Routledge). Van der Panne, G., Kleinknecht, C. P. & Van Beers, A. H. (2004) Success and failure of innovation: A literature review, International Journal of Innovation Management, 7(3), pp. 1–30. Van den Elst, J., Tol, R. & Smits, R. (2006) Innovation in practice—Philips applied technologies, International Journal of Technology Management, 34(3–4), pp. 217 –231. Van Gunsteren, L. A. (2003) Management of Industrial R&D. A Viewpoint from Practice (Delft: Eburon Publishers). Verloop, J. (2006) The Shell way to innovate, International Journal of Technology Management, 34(3–4), pp. 243 –259. Volberda, H. W. (1998) Building the Flexible Firm (Oxford: Oxford University Press). Von Hippel, E. (1986) Lead users: A source of novel product concepts, Management Science, 32 (July), pp. 57 –71.