Governance and resource interaction in networks. The role of venture capital in a biotech start-up

Journal of Business Research 65 (2012) 232–244

Contents lists available at ScienceDirect

Journal of Business Research

Governance and resource interaction in networks. The role of venture capital in a biotech start-up Torkel Strömsten a,⁎, Alexandra Waluszewski b, 1 a b

Stockholm School of Economics, Box 6501, 113 83 Stockholm, Sweden Uppsala STS, Uppsala University, Box 513, 751 20 Uppsala, Sweden

a r t i c l e

i n f o

Article history: Received 1 July 2009 Received in revised form 1 September 2010 Accepted 1 November 2010 Available online 24 August 2011 Keywords: Resources Innovation Networks Corporate governance Venture capital Speed

a b s t r a c t This paper examines venture capital (VC) governance in innovation processes. The VC literature often presents the relationship between a VC firm and a start-up as dyadic and analyzes it with agency theory. In contrast, this paper deploys the resource interaction framework presented in Håkansson and Waluszewski (2002) to governance and innovation in networks. The paper reports an in-depth case study of Pyrosequencing, a Swedish biotech firm financed with VC. The results from this study reveal how the relationship between a VC and a start-up company is embedded in a wider network and how the governance of the VC spreads in the surrounding network and influences a start-up's possibilities to develop organizational and technical resource interfaces to critical counterparts such as suppliers and customers. © 2011 Elsevier Inc. All rights reserved.

1. Introduction Venture capital (VC) firms invest in entrepreneurial firms perceived as having high potential but also high risk, only 2 of 10 projects that VC finances survive (Gompers & Lerner, 1999). VC firms are active owners, they often sit on the board of directors and participate actively in decision-making in the firms they finance (Jain, 2001; Sahlman, 1990). Research about the relationship between a VC firm and a start-up is extensive (e.g. Murray, 1996; Perry, 1988; Sapienza & De Clercq, 2000; Shepherd & Zacharakis, 2001). However, despite considerable literature on how VCs add value as active owners (Amit, Brander, & Zott, 1998; Sahlman, 1990; Sapienza, Manigart, & Vermeir, 1996; Timmons & Bygrave, 1986), there is little research on how VC governance influences innovation in networks. This stems from an important analytical assumption. Most existing research on VC assumes that the best way to analyze the relationship between a VC firm and a start-up is through agency theory (Sahlman, 1990; Sapienza & De Clercq, 2000). As a consequence, such research treats VC-start-up relationships as dyadic, whereby the principal (the VC firm) designs controls to govern its investments in an agent (the start-up and its management). The assumption is that the agent is driven by self-interest and guile. This perspective ignores the network ⁎ Corresponding author. Tel.: + 46 8 7369305. E-mail addresses: [email protected] (T. Strömsten), [email protected] (A. Waluszewski). 1 Tel.: + 46 18 471 5601. 0148-2963/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.jbusres.2010.11.030

that embeds the relationship (Granovetter, 1992; Håkansson & Snehota, 1995). Further, agency theory is not a theory concerned with innovation and ignores therefore also the using of physical artifacts, forming technical systems (Hughes, 1983), and their influence upon exchange relationships and development of networks (Anderson, Håkansson, & Johanson, 1994; Baraldi & Strömsten, 2009; Håkansson & Lind, 2004; Håkansson & Waluszewski, 2002; Halinen & Törnroos, 1998). Some researchers claim that VC is an impatient form of ownership (e.g. Van de Ven, Polley, Garud, & Venkataraman, 1999). One reason is because third parties, such as institutional investors, finance VC firms' investment in start-ups in a fund often with a limited lifetime, frequently 10 years. Hence a VC invests with the aim to exit within a certain period of time (Gompers & Lerner, 1999). This explains their impatience and desire to accelerate the innovation process in the firms they invest in. A dyadic agency-principal perspective that colors much of the VC literature ignores important consequences of accelerating the commercialization process and associated governance issues when relationships and networks matters (Ford et al., 1998). The interface between innovation and corporate governance is an under researched area (O'Sullivan, 2000). One reason is that dominant theories of corporate governance, such as agency theory, “do not systematically incorporate an analysis of the economics of innovation” (O'Sullivan, 2000, p. 413). This article explores how VC ownership and governance influence the innovation process when relationships and networks are critical contextual factors. It draws upon a framework (Håkansson & Waluszewski, 2002, 2007), which analyzes how resources (two technical—products and facilities; and two

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

organizational organizational units and relationships) interact in industrial networks. This article uses this framework to trace how VC ownership affects the development, purchase and sale of products; the production and utilization of facilities; organizational units' strategic decisions about scope and relevant knowledge areas; and a portfolio firm's ability to develop relationships with suppliers and customers. This article tries to remedy the knowledge gap and the call for more research on the interface between innovation and corporate governance. This article is organized as follows. First, it discusses the governance VC provides and how this governance can influence the innovation process, and how ensuing controls determine the priorities and behavior of a VC firm. It then introduces the resource interaction framework (Håkansson & Waluszewski, 2002) and discusses it with reference to innovation within networks, and possible effects of VC governance on organizational and technical resources and their interfaces. After describing the research methods, the article reports an in-depth case study of Pyrosequencing, a Swedish biotech firm backed with VC money. The case illustrates how a 4R approach raises fresh issues about governance and its role in innovation processes in networks. The paper ends by discussing the implications of VC for innovation, especially the tensions between the VC firm's need for speed and a start-up's time-consuming innovation processes where resources and interfaces constantly change in the surrounding network. 2. Venture capital governance in the innovation process Sahlman (1990) defined VC as a professionally managed pool of capital invested in private companies at various stages of their development. VC firms actively participate in the decision making processes of the ventures they invest in. Typically they also become members of the board of directors and assist management with advice and support. Hence, VC represents a specific type of corporate governance that takes an active part in start-up companies' innovation processes. In the discussions by researchers, policy-makers or journalists, the most common interpretation is that VC is a key resource in a start-up's journey toward becoming an established company. Well known companies (most often US companies like Apple, Cisco, Intel, Microsoft, Genentech, Yahoo! and Amazon.com), that have received VC financing often serve as examples of what is possible when innovative ideas are combined with “intelligent” financing. Without VC, argue Gompers and Lerner (2001, p. 1), “many entrepreneurs would never attract the resources they need to quickly turn their promising ideas into commercial success”. Or, as Powell, Koput, Bowie, and Smith-Doerrs (2002, p. 293), put it “Venture capital is one of the key elements of the infrastructure of innovation”. A VC firm provides three critical resources to a start-up enterprise. First, money expands the capacity to transform an idea or a new solution from individuals or a project into a company with established customer interfaces (i.e., Barney, Fiet, Busenitz, & Moesel, 1996). Money aids start-ups to grow, initiate product development projects, extend its specialized activities, invest in equipment, hire new staff, and use outside partners for product development work. Second, the VC firm not only provides capital but also experience, in the form of knowledge and foresight about the risks and opportunities that entrepreneurs face (Barney et al., 1996; Sahlman, 1990). Powell et al. (2002) stress the importance of combining money and knowledge when financing high-tech enterprises. As Gompers and Lerner (2001, p. 19) state, “Most hightechnology entrepreneurs are convinced that they have exciting and dynamic ideas… What most entrepreneurs do not see clearly, however, are the risks facing their business.” This uncertainty encompasses critical issues, such as who are the potential users,

233

suppliers and strategic partners, and potential market size (Gompers and Lerner, 2001, p. 20). Third, the VC firm provides a network of relationships including financial, commercial or technology-based contacts, for instance a skillful venture capitalist can often activate a wide network of industrial contracts to help a new venture find suppliers and customers in various ways (Fried, Bruton, & Hisrich, 1998). Being part of a VC network can help transfer experience and knowledge between firms and establish contacts with third parties such as new investors and investment banks. Being part of a “good” network helps young firms because they often face the “liability of newness” (Smith & Lohrke, 2008; Stinchcombe, 1965). The VC literature reflects the extensive influence of agency theory (e.g. Fama & Jensen, 1983). This theory analyzes how a principal, the VC firm, can monitor an agent, the start-up and its management, to align their interests to create shareholder value. A major theme is information asymmetry between the parties and how this creates “moral hazards”. Basically, the agent knows more about the firm's operations and could, therefore, cheat the principal, who develops mechanisms to protect himself. Alignment of interests is one way of avoiding opportunistic behavior; consequently VC often encourages managers to become part-owners. (Gompers & Lerner, 2001; Sahlman, 1990). Agency theory helps explicate the legal structure of a VC firm, which often involves a private equity partnership (e.g. Gompers & Lerner, 1999; Sahlman, 1990). The general partners (the VC management) manage the firm, monitor its investments funded mainly by limited partners (often institutional investors like mutual funds, pension funds or insurance companies), and find investments to increase the limited partner's money over a set time (often 10 years) after which equity in the fund must be returned to the limited partners. The limited time horizon exists to protect investors from VCs acting opportunistically (Freeman, 1999; Gompers & Lerner, 2001; Sahlman, 1990). The exit is always present, explicitly or implicitly, in relationships between a VC firm and a start-up. Further, there is also a tendency among young VC's to take their portfolio firms public through initial public offering (IPO) too early in order to “signal their ability to potential investors” (Gompers & Lerner, 1999, p. 239). This phenomenon is called “grandstanding”. One cost of grandstanding is a lower price for the portfolio firm when it goes public. However, there are also other costs, less well researched, associated with grandstanding that might be harmful for the portfolio firm in the long run (Gompers & Lerner, 1999). VC firms often use milestones or similar controls to secure rapid results and to monitor themselves and the firms they have invested in (e.g., Davila, 2005). Making the emerging company's management accountable for attaining milestones within a set timescale introduces stepwise financing through a stage/gate process. When a VC firm uses staged financing, the start-up must meet milestones before the VC invests more money (Gompers & Lerner, 1999). The aim of this procedure is to reduce the inherent risk of innovation processes and hinder opportunistic behavior from management in the portfolio firms. In summary, the governance mechanisms of a VC firm reinforce its interest in compressing the innovation process of the start-ups it invests in. However, emphasizing principal-agent attributes and consequently picturing the relationship between the VC firm and the start-up (Sahlman, 1990) as strictly dyadic ignores how relationships are embedded in a wider network of customers, suppliers and third parties (Håkansson & Snehota, 1995). These networks provide the topic of discussion below. 3. Resource interaction and innovation in networks Van de Ven et al. (1999: ix) characterize the transformation of innovations into commercial solutions as “highly unpredictable and

234

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

uncontrollable”. Tidd, Pavitt, and Bessant (1997) depict it as messy and describe ways of handling it as “trial-and-error” and “muddling through” (see also Basalla, 1988; Bijker, 1997; Hughes, 1983; Rosenberg, 1982). These researchers share the belief that an “innovation journey” (Van de Ven et al., 1999) is so complex that it resists a complete understanding as these journeys require the joint application of new and existing solutions, the combination and recombination of resource pools and the management of unexpected events and consequences. This corresponds with findings on technological development in the Industrial Network Approach research (Axelsson & Easton, 1992; Holmen, 2001; Håkansson, 1987; Håkansson & Waluszewski, 2002; Laage-Hellman, 1989; Lundgren, 1991; Waluszewski, 2004; Wedin, 2001). Starting with the assumption that resources are heterogeneous (Alchian & Demsetz, 1972; Penrose, 1959), Håkansson and Waluszewski's (2002) resource interaction framework (hereafter the 4R framework) demonstrates first how the features and value of technical resources (products and production facilities) and organizational resources (organizational units and business relationships) rest on endeavors to combine them, and then how these features relate to one another and require remodeling both within and beyond organizational borders (Baraldi, 2003; Baraldi & Strömsten, 2006; Baraldi & Waluszewski, 2005; Forbord, 2003; Gressetvold, 2004; Hjelmgren, 2005; Håkansson & Waluszewski, 2002; Wedin, 2001). Hence, innovation concerns combining old and new resources in an industrial network where heterogeneity and interdependencies between organizational and technical resources are the normal state—not homogeneity and independence as in agency theory. First, the 4R framework does not view products as given but as results of historical and future interaction patterns. Products are part of both a “selling/buying” and a “using” system (Håkansson & Waluszewski, 2002, p. 35) and the logics governing these systems are not always the same. Second, facilities form a producing/using system where actors systematically try to improve interfaces between the facilities to save financial resources or time. A facility is never isolated from other facilities, instead actors should use and exploit its interdependencies with other facilities. Third, organizational units comprise of skills, experiences and structure. When an organizational unit cooperates or interacts with other units, it gains imprints from this—“it develops specific social features” (Håkansson & Waluszewski, 2002, p. 36). Fourth, relationships are important resources in their own right: they and their features can help achieve actors' objectives. Historical processes, future plans and interconnections of actors' relationships to those in wider networks can create opportunities and restrictions. Some relationships “open doors”, and encourage interesting endeavors whilst others act to the contrary. It is important to note that all four resources are interdependent, that is, “to produce a product, we need a facility that is owned by a business unit and in order to sell the product we need a relationship” (Håkansson & Waluszewski, 2002, p. 38). Between the resources, interfaces, that is, “meeting points” emerge. These can be both organizational and technical and a firm must manage several interfaces in its network to secure effective resource combinations (Baraldi, 2003; Baraldi & Strömsten, 2006; Dubois & Araujo, 2006; Gressetvold, 2004; Håkansson & Waluszewski, 2002, 2007; Lind, 2006; Lind & Strömsten, 2006; Wedin, 2001). Existing intra- and inter-organizational routines affect the adoption and enactment of a new innovation in a network. A company launching a new product must create economic value for itself and others connected to it, be they suppliers of money or resources, or users of the product on whom long-run product survival rests. A product launch is challenging for any firm but more so for a new venture or start-up. To succeed, it must embed itself in an intricate pattern of organizational and technological resources, that is, it must find a daily use in customers' facilities and in combinations with other products (Baraldi & Strömsten, 2006). The combination of a new innovation with other resources

spanning various organizational borders, rather than its innate qualities, creates economic value. Consequently, managing the combination of resources and interfaces in both the close and distant network is critical for firms pursuing innovative endeavors. To sum up, the resource interaction framework presents a different view of innovation than the view that dominates the VC literature. In the resource interaction literature, learning about heterogeneous resources is an experiential process that requires time (e.g., Strömsten & Håkansson, 2007). In the next section we will discuss the connection between these two views and how VC governance might influence the design and use of organizational and technical resources. 4. Resource interaction and governance in industrial networks According to the discussion above, the governance of a VC firm makes time management significant in its relationships with portfolio firms. This time issue impacts how the portfolio firms will organize interfaces between organizational and technical resources directly and indirectly. The interface between VC governance and developing innovative products contains an in-built tension: like established companies, start-ups need time to learn how to best combine heterogeneous resources, develop a product, and embed it in dynamic interfaces between organizational and technical resources. However, the VC firm's time constraints will in some cases compress the innovation process for its portfolio firms. The interface between VC governance and organizational units influences how and what capabilities a company will develop. Implicitly, the VC literature presume that learning occurs “before doing” as opposed to “by doing” (e.g., Rosenberg, 1982), and that the VC firm can already identify appropriate producer-user interfaces and a development path, that is, who will use what kind of product and how. Furthermore, the VC industry uses the Internal Rate of Return (IRR) as an important performance metric. As IRR incorporates the time value of money, the sooner a divestment becomes possible, the higher the IRR. This emphasizes the incentive to minimize the time from investment to exit and accentuates the incentive to focus on the already known in order to create a growing company and a potential exit instead of exploring new possible resource combinations, where the outcome is uncertain and risky. This tallies with agency theory as it helps protect principals against opportunistic agents. In contrast, Dosi (1988: 222) states: “Almost by definition, what is searched for cannot be known with any precision before the activity of search and experimentation itself”. For Dosi, enterprise involves handling unexpected effects and repeatedly trying out new and existing solutions. The most common depiction of this process is trial-anderror learning: a firm must examine how new resource combinations affect established interfaces across firm boundaries and adapt accordingly. Hence, time is according to this view not something to minimize but a prerequisite for creating value, as it permits adaptation and embedding to occur. A VC firm often funds several start-ups with similar or complementary activities, which prompts opportunities to connect resources in one relationship to broader constellations. The VC firm wants to exploit all its relationships to maximize the value of its portfolio, although this is not necessarily what each start-up firm would prefer. How a VC firm controls a start-up in terms of time will affect the nature of the customer and supplier relationships it develops. A VC can help the start-up in finding customers and users. Minimizing the time from an initial invention to establishing a growing company is critical for a VC firm but impositions of time constraints upon the start-up may diminish its critical resources and possibilities to develop long term interfaces with users and providers. There appear to be two logics at play; a VC logic based on financial reasoning, where time is a critical issue that should be minimized, and an innovation logic, which emphasizes combining and

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

exploratory learning (e.g., March, 1991) of heterogeneous resources, which in turn requires time. These two logics (e.g., Lounsbury, 2007, 2008) seem to be incompatible; yet still they may also be complementary as financing and innovation are frequently codependent. Capital is necessary to finance innovation and innovation, appropriately managed, can (not necessary though) increase the capital invested. However, the logic or theory that underlies finance does not capture or appreciate the uncertainty of innovation and the time it might take to commercialize basic research, especially not when firms are embedded in a network of organizational and technical resources. The description of the relationship between the VC firm and the start-up in the VC literature clearly illustrates this issue. In summary, while the VC literature to a large extent focuses on the relationship between the VC and the entrepreneur, the industrial network approach and the resource interaction framework emphasize the embeddedness of this relationship in a wider network. Further, while time is a resource that should be minimized in the VC literature, in the literature on resource interaction time is rather a necessary condition for innovation. In the next sections we will present the methods we applied and the empirical case. 5. Method How governance affects innovation in networks is an under researched area, hence the need for an exploratory and single, indepth case study (e.g., Yin, 1989). There is a vast group of researchers that argue for different uses of case study research (e.g., Dyer & Wilkins, 1991; Eisenhardt, 1989) and the value of case studies in developing and extending knowledge about organizations and their behavior (Abbot, 1992; Dubois & Araujo, 2004; Easton, 1995; Eisenhardt & Graebner, 2007). Generalization from a single case study, in this article, proceeds by analytical or theoretical generalization. This case study is related to a larger study on the biotech sector in the Uppsala–Stockholm region in Sweden (Waluszewski, 2004). Other studies have covered issues such as institutional drivers in innovation (Shih, 2009; Waluszewski, Baraldi, Shih, & Linné, 2009) and users as key players in product development (Harrison & Waluszewski, 2008; Wagrell & Waluszewski, 2009). Ingemansson and Waluszewski (2009) and Ingemansson (2010) also use Pyrosequencing as a study object and investigate the developing, producing and using contexts of the firm. The 4R framework (Håkansson & Waluszewski, 2002, 2007) helps to address how the start-up develops and combines critical resources during its attempts to embed a new solution into user and supplier interfaces and how a VC-governed innovation process affects this and associated strategic decisions. Analysis started during the processing of interviews. The 4R framework acted as a framing mechanism: this helped match interviewees' responses with the theoretical categories. The analysis thus relates priorities and technological choices to the governance exercised by the VC firm. In the second analytical stage the 4R framework helped trace the impact of the VC governance upon features of critical resources, and how decisions about resource interfaces in one period affected later possibilities to maneuver. The empirical material falls into three phases from 1996 to 2003. The first started the year before the firm's foundation and includes the first year thereafter. The second phase covers Pyrosequencing's foundation in 1997 and the start of the product development project until its first sale in 1999. The third phase extends from preparing the Initial Public Offering (IPO) up to Pyrosequencing's merger with Personal Chemistry, another of the VC firm's portfolio firms, and Biotage, an American life science company in 2003. The empirical material came primarily from interviews with employees of Pyrosequencing and organizations close within its network. 35 people were interviewed—33 face-to-face and two by telephone. Interviewees included past and present employees of

235

Pyrosequencing; employees from its suppliers, customers, and the VC firm financing it; analysts and investors in the VC firm. The appendix details the interviews conducted. The aim was to identify influences on the innovation process in Pyrosequencing, trace its surrounding network, and identify changes over time. In addition, company reports, annual reports, product information and internal accounting reports helped identify the objectives, key performance indicators and strategies of the company. The empirical material was collected over a period of six years beginning in 2001: some key informants provided several interviews. 6. The Pyrosequencing case 6.1. Phase 1: 1996–1997 In 1953 Crick and Watson discovered the “secret of life” when they built a model of the human genome, the double Helix. In the 1970s Fredrick Sanger invented a method to sequence human DNA, which earned him the Nobel Prize in 1977. The technology underpinning the Pyrosequencing business plan relates to these breakthroughs. The case starts in a laboratory in Cambridge, UK. Pål Nyrén, a founder of Pyrosequencing, earned his PhD in chemistry at Stockholm University in 1985, whereupon a post-doctoral grant took him to the Laboratory for Molecular Biology at the Medical Research Council (MRC) in Cambridge to try and sequence the gene for bovine mitochondrial phosphate-carrier protein (Nyrén, 2006) and to learn how to sequence DNA according to the dominant Sanger method. During his work, he realized that the Sanger method was labor intensive and involved several tedious and expensive operations, like using gel slabs. Nyrén started thinking about different methods of sequencing DNA. His background in photosynthesis and bioenergetic research led him to consider how enzymes could help, especially using pyrophosphates to achieve light-driven pyrophosphate synthesis. Nyrén's idea was consistent with Watson and Crick's discovery that the nucleotides Adenine, Guanine, Thymine and Cytosine only bind together two DNA strings when they appear in certain combinations (adenine with thymine and guanine with cytosine), thereby creating a double helix. The novelty of Nyrén's idea in comparison to the Sanger method was its use of enzymes to determine the order of nucleotides when detecting DNA sequences. He endeavored to improve and simplify his concept by following “the activity of DNA polymerase during nucleotide incorporation into a DNA strand by analyzing the pyrophosphate during the process” (Nyrén, 2006; interview 26, interview 25). Back in Sweden Nyrén unsuccessfully sought funding to develop this method. Using resources from other projects he published on this topic (Nyrén, 1987) but no research funding came. In 1990 he moved to the Royal Institute of Technology where he met Professor Mattias Uhlén, who had researched solid phase DNA sequencing. The researchers had common research interests and Nyrén started working with Uhlén. This was important for identifying possibilities of commercializing the method: Uhlén was well connected in academia and in Swedish business, and had started two ventures based on his research. Nyrén had slowly developed the pyrosequencing method alone but his new colleagues and the availability of resources helped advance it. At this point the method remained manual but needed automation to become commercially viable (Interview 26, interview 1). A major impediment to the method's automation and commercial development was removing excessive nucleotides whilst detecting the DNA sequence. The breakthrough—using the enzyme apyrase to wash unnecessary nucleotides away—meant that the method now contained four enzymes incorporating the correct nucleotides to a DNA template (P. Nyrén, 1998/2001). This created a light-signal which could be detected and registered in real-time. (Ingemansson, 2010; Nyrén, 2006). Further, Nyrén and his colleagues published an

236

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

article in the prestigious journal Science (Ronaghi, Uhlén, & Nyrén, 1998) which also meant that the method gained acceptance in the scientific community. These developments occurred simultaneously to increased media attention for the Human Genome Project (HUGO). Nyrén and Uhlén discussed how to link their invention to the HUGO project. When the latter started to deliver positive results, they realized that more applied methods to analyze the bases or nucleotides that constitute a DNA sequence would be necessary. Developing a DNA sequencing method to compete with established technologies like the Sanger method became the long-term research aim of the Royal Institute of Technology team. They found their method could sequence short and medium-short DNA strands and analyze single bases and their variations, called Single Nucleotide Polymorphism (SNP)—genetic variations in the human genome occurring when a nucleobase (Adenine, Guanine, Thymine or Cytosine) differs between populations of individuals. When the HUGO project concluded in the late 1990s, SNPs were seen as the next step in developing new drugs. But the inventors' vision remained unchanged, namely to develop a generic or general DNA sequencing method (or technology) capable of reading longer DNA sequences than single nucleotide mutations. To achieve this, the “read length” of the Pyrosequencing technology (circa 15–20 base pairs in 1997) needed to extend to around 200–300 base pairs (Interview 26, interview 24, interview 33, Interview 5). The Royal Institute of Technology researchers initiated a collaborative project with Pharmacia Biotech, with the aim of eventually licensing the technology to the firm. However, Pharmacia Biotech proved to be uninterested (because of an intense merger process with UK-based Amersham) so the inventors decided to start their own company. In this process they realized they needed capital. Uhlén was a scientific advisory board member of HealthCap, a VC life science firm founded the previous year. After hearing a presentation of Pyrosequencing invention, HealthCap decided to invest 2 million USD in the new company. This was HealthCap's first fund and Pyrosequencing was their fourth investment. From the outset, the founding team and HealthCap had a specific business model in mind, namely to offer pharmaceutical firms, small drug discovery firms and academic researchers a system consisting of an instrument, a software package and a kit of reagents, which incorporated the system of enzymes. The reagent kit would be the invention from the Royal Institute of Technology, which would become the company's core intellectual property (Interview 26, interview 1, interview 2, interview 9). Once the researchers realized they needed VC finance, they modeled their actions on how they believed a venture capitalist would view their team and their innovation. For example, they recruited, Björn Ekström, the manager at Pharmacia Biotech who had evaluated the technology, to help create a company around the Pyrosequencing method. As Ekström would be a key actor as Pyrosequencing's Chief Technology Officer (CTO) in developing the technology, he became part of the founding team and not merely an employee. In addition, in the due diligence process that followed from the contact with the VC, the founders acquired a patent on the sequencing-by-synthesis method (Melamade, 1985). According to one of its inventors this was not a strong patent but Pyrosequencing's founders and the VCs believed they should control and secure ownership of it to stem any anxieties about potential competitors copying the method and to facilitate future discussions with potential investors (Interview 1, interview 2, interview 25, interview 26). Fig. 1 below illustrates the critical organizational and technical resources involved in the first phase where thick lines indicate close relationships and interfaces, while thin lines indicate weaker ones. 6.2. Phase 2: 1997–2000 Pyrosequencing, named after its technology, was formally founded in March 1997. The board of directors, consisting of the

founders and representatives from HealthCap, agreed that it needed a CEO with experience in start-up companies to develop effective administrative routines. This job went to a venture partner from HealthCap who had worked as a CEO and general manager at several start-up firms. He brought a “start kit” to get Pyrosequencing up and running quickly. For example, he appointed an auditing firm and a software company, and bought a business reporting system he had used previously. The CEO and the VC firm HealthCap influenced the development of administrative routines and, to some extent, relationships with critical suppliers (Interview 1, interview 9, interview 6, interview 7). The board, management and owners identified a time-frame that consisted of milestones, for instance, for when prototypes and products would be finished. To save time, they outsourced some product development work to local partners. Timing was a critical means of exercising control over product development. A product developer stated: “Everything was designed to shorten the time needed to achieve different milestones. Time objectives were of great importance. The quality was certainly important as well but cost was never an issue”. This had consequences not only for how the firm developed a product around the innovation from the Royal Institute of Technology, but even for which product it developed. For example, to save time, product development was directed at a particular application and knowledge area. The inventors' original intention had been to create a “general” or generic DNA sequencing system, which requires a system of reagents that is able to read long DNA sequences, but the board and management soon prioritized other areas because of the inadequate read length of the Pyrosequencing method. Creating a product quickly and getting it to market as soon as possible became the highest priority. The strategy that the board formulated was to focus on the market for SNP, where the read length not was an issue as SNPs consists of single nucleotides, not long sequences. Their belief was that the Pyrosequencing technology would help customers identify SNPs in an automated, reliable and simple fashion. The assessment of the SNP market was further very positive: it would grow 35% annually and in 2004 would be worth 750 million USD (Dagens Industri, 1999-1110). The firm planned to break even by 2002 and attain a turnover of 200 million USD in 2004, representing 25% of the SNP market (Pyrosequencing prospectus, 1999, interview 21, interview 2, interview 3, interview 1, interview 29). The board formalized the informal discussions between the founders and HealthCap about the business model. They decided that the product, or rather the system, would have three parts: an instrument (to perform SNP analysis), a reagent kit (the invention, namely advanced chemistry and combinations of enzymes to aid analysis), and a software program (to visualize the analysis and its results from databases). The aim was to sell the hardware that consumes reagents: subsequent sale of reagents would make most of the profits. Ekström orchestrated product development by involving external partners. A local company, ESSDE, well-known in the Uppsala biotech cluster, was for example instrumental in designing and developing the instrument. Time was so critical for Pyrosequencing that they eventually paid the cost of a full-time ESSDE employee to ensure that requisite development capacity was available. Further, to quickly build up its software development unit, they outsourced software development to an Uppsala company, Prevas (interview 1, interview 11, interview 13, interview 7). However, once the firm focused on SNPs, the inventors' priorities associated with increasing the read length became subservient to those of the board and the owners. Thus, despite researchers at the Royal Institute of Technology continuing to increase the read length (it eventually reached over 100 base pairs), the firm was reluctant to incorporate this into product developments due to the tight time schedule and controls established (Interview 15, interview 29, interview 25, interview 26).

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

237

Fig. 1. Network of resources in the first phase of Pyrosequencing's innovation process.

In 1998, the board appointed a new CEO, Erik Walldén, experienced in the biotech sector. A prototype was finished in 1998 and test versions were sent to potential customers. Their feedback was positive: researchers seeking drug discoveries found the technology appealing. The instrument, PSQ96 (from the company name—Pyrosequencing—and the instrument's capacity to read 96 samples simultaneously) could perform about 5000 tests daily, which surpassed all alternative technologies, and the software program analyzed results and created a database for SNP sequences. Pyrosequencing's most important feature was however not speed or throughput, but the high quality of data that the system created. Most participants believed the product development phase was a success. Nevertheless, there was one noteworthy problem: costs had heavily overrun the budget: originally the development cost of PSQ96was set at 7.5 million USD but it exceeded 25 million USD. Working for Pyrosequencing was different than in large multinationals, where most employees had come from. For example, an R&D manager realized that development costs exceeded calculations and he feared that management would terminate the project. The venture partner commented: “In just a few hours I phoned the board members and secured the money. This was unheard of, and of course encouraging for the R&D people to see; it increased loyalty in the company.” (Interview 9, interview 33, interview 28, interview 6). There were three rounds of financing. HealthCap in VC industry jargon was the “bell cow”, that is, lead investor. The 30 million USD raised enabled Pyrosequencing to initiate development work, invest in facilities, and increase the scale and scope of its activities. Additionally, during this phase Pyrosequencing needed to manage relations with the VC firm's other portfolio firms. For example, when one of HealthCap's investments faced financial problems in 1999, HealthCap encouraged Pyrosequencing to buy some of the firm's patents. In addition, the way its institutional investors evaluated

HealthCap in effect required Pyrosequencing to follow a path conducive to flotation on the stock exchange within a short time period. However, before doing so, the company needed a product and customers (Interview 1, interview 8, interview 18, interview 34). During spring 1999, the board decided to seek an initial public offering (IPO) for Pyrosequencing. The original ambition was for listings on the Stockholm and New York (Nasdaq) Stock Exchanges but the board of directors soon determined that a Stockholm Stock Exchange listing was sufficient. Consequently, the focus on time in product development intensified. One of Pyrosequencing's employees put it like this: “The product development project was very effective and it went very quickly. The price we paid for accelerating the product development process was lack of flexibility and a lack of possibility to adapt to alternative solutions…and to use “add-ons”. All these things require testing and testing takes time… (Interview 34, interview 4, interview 7). In late 1999, when AstraZeneca—the Swedish–British pharmaceutical company, purchased a system, Pyrosequencing could commence its path to a listing by attending “road shows” and entering negotiations with investment banks (Interview 12, Interview 8, interview 34). With the exception of the cost, most product development targets were met. The CEO's 1999 report stated: “We met every key milestone set for 1999, culminating in the commercial introduction of PSQ96 and our first sales”. Key milestones listed were: alpha and beta site testing completed; serial production started; commercial availability; sales and support office opened in Boston, USA; a sales force established; initial orders from the USA and Europe received; patent portfolio strengthened; and private placement of 15 million USD raised. Having developed a robust, simple and easy-to-use system for DNA analysis, technological and commercial success appeared to be within reach (Interview 6, interview 8, interview 22).

238

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

The production plan for reagents grew out of a sales prognosis by the owners and management in the late 1990s. This emphasized the importance of being ready early to serve the booming SNP market. Hence, Pyrosequencing made substantial investments in manufacturing capacity including: a clean room, a filling line, a refrigerating dryer, and chromatography instruments. Around 12 million USD went into building manufacturing lines for producing reagents. The strategy determined that Pyrosequencing would not manufacture the hardware or the software: these were outsourced. The firm PartnerTech, who since earlier served the booming biotech industry in Sweden, produced the hardware. The important organizational and technical resources identified during phase two appear below in Fig. 2, where thick lines indicate close and strong relationships and interfaces while thin lines indicate weaker or more distant relationships (Interview 28, interview 29, interview 31). 6.3. Phase 3: 2000–2003 During spring 2000, Pyrosequencing launched its PSQ96 system and continued presentations to potential investors. Results from initial installations revealed that users found the product strong and robust. In retrospect, an R&D manager at AstraZeneca stated: “It was an important system for us when it first came. It was indeed reliable as it had a built-in control, which made the instrument very fascinating. I think everybody engaged in drug discovery found it fascinating.” Indeed, AstraZeneca soon purchased four more systems and became a prime reference customer for Pyrosequencing. However, whilst users considered the system a success on its launch, they later experienced problems with the software, so a review of the outsourced development eventually brought it in-house (Interview 14, interview 12, interview 19).

By June 2000, three years after its foundation, Pyrosequencing went public: now HealthCap could provide its investors with an exit. The timing was extraordinary: one week before the IPO, results of the HUGO project and the competing Celera project were copresented at a White house ceremony, hosted by President Bill Clinton. Media attention was impressive. It was important for HealthCap to show potential investors that it could pick good investments, add value to them, and generate high returns, especially as it wanted to embark on a second fund soon. Pyrosequencing had given HealthCap's investors high returns and because investors in HealthCap's fund used the standard metric IRR to evaluate HealthCap, Pyrosequencing's very fast listing on the stock exchange was significant. Nevertheless HealthCap remained an owner after the IPO to signal their belief in Pyrosequencing's future (Interview 9, interview 30, interview 20, interview 23). Pyrosequencing received 125 million USD from the IPO. It launched at 12 USD per share in 2000 and later that year the share value reached 25 USD. The company's market value rose and six months after the IPO it reached 625 million USD. Then Pyrosequencing was indisputably a success story. The Swedish Civil Engineering Association named it “spin-off of the year” and the business magazine Forbes gave it the prestigious accolade of “one of the world's most promising 300 companies” (Interview 23, interview 8). As a public company, Pyrosequencing now had to submit quarterly financial reports to the Stockholm Stock Exchange. “Getting instruments out to the customers” became an important metric for quarter ends, as market valuation depended upon sales growth and profit margins. Thus, sales and technical support prioritized sales of new instruments. However, the need to report sales figures to the stock exchange influenced Pyrosequencing's relationships with its

Fig. 2. Network of resources in the second phase of Pyrosequencing's innovation process.

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

users by, for example, pressure to rapidly conclude a sale. Sometimes, this resulted in the personnel training that formed part of the sales package being conducted after the purchase, which contradicted the aim of getting customers to use the instrument frequently, and then purchase the economically important reagent kits (Interview 23, interview 34, interview 7). Initially, customers appeared content with Pyrosequencing's SNP analysis and its instrument, PSQ96. They considered the technology fascinating: users repeatedly described it as “elegant”. However, large scale DNA analysis entails a system of interrelated activities: SNP analysis is only the last of several closely related steps in a chain of activities. For example, before conducting an SNP analysis (using PSQ96), customers must choose which SNP to analyze. This depends upon the area of interest and the number of SNPs under consideration, which can range from one to hundreds. Second, they must design and develop an assay for each SNP, which is time-consuming and, depending on the analysis system used, can be labor intensive. The third step involves Polymerase Chain Reaction (PCR) amplification, and preparation of tests that can run into hundreds for specific applications. The final step is the DNA and SNP analysis. From a user perspective, a thorough and effective SNP analysis should incorporate all these steps. Pyrosequencing was only being able to provide the last step. This was a drawback. It became evident that customers wanted more than an “elegant” product (Interview 12, interview 7, interview 34). From 2001 increasing numbers of customers requested integrated solutions and methods to make the entire activity chain more efficient. Competitors offered ways to integrate these activities. Pyrosequencing's instruments were faster and more reliable than their competitors' but their overall offering was less effective. For example, according to an AstraZeneca researcher, “Preparation of the test was really time-consuming and that was a problem for us.” A Pyrosequencing product manager commented: “I think we have focused too much on our instrument… and perhaps that is not so strange … we had fantastic benefits early on as our instrument is so easy to use. But other suppliers of instruments have addressed the other steps in the chain of activities… so their instrument is faster, when one considers the whole chain … we have not been able to integrate [PSQ96] as successfully in the other steps”. Hence, after the initial success, sales did not develop according to plan. Moreover, the world economy dipped in 2001, but despite of this in December that year Pyrosequencing was still growing and had 130 employees. In June 2002 it had 156 employees. However, Pyrosequencing had misinterpreted customer preferences and the SNP market failed to develop as expected. In 2001 it was worth only 150 million USD, of which Pyrosequencing had about 5–10% (Interview 1, interview 12, interview 19). As mentioned, Pyrosequencing started in the wake of the HUGO project and sought to target academia, big pharmaceutical companies and the booming research and drug discovery market, where other small, expanding start-ups also financed by VC firms were active. In 2001 when the economy slowed their market and finances became troublesome—the large anticipated market did not materialize and the number of small drug discovery firms diminished. This hurt Pyrosequencing as its intended market lay in this sector. In 2002 over 70% of Pyrosequencing's customers were within academia (Interview 24, interview 21, interview 32, interview 35). In response to users' demands, Pyrosequencing started to develop a second-generation instrument—an automated high through-put system, PTP348, (Preferred Technology Program) that could analyze 348 samples simultaneously. Customers would with this system be able to integrate and speed up their processes. An engineer commented: “We added more features to the system, made it faster, more automated etc. We aimed for the big pharmaceuticals companies and, in order to secure them as customers we believed that a high through-put system was essential.” However, the second-

239

generation instrument's development was more complex, and needed more time and cost than the original innovation. There were technical problems that developers never fully solved, for example, problems with interfaces between components adding nucleotides and the sample volume. The time allocated for product development project proved unrealistic, and when users tested the new product, its multiple drawbacks meant it never reached market. Consequently, Pyrosequencing's value on the stock exchange fell in 2001 and by 2002 it was 35 million USD. Eventually, Pyrosequencing's cash holdings exceeded its market value (Interview 1, interview 6, interview 23, interview 34). In response, management reduced costs, cutting staff by almost 20%. A press release from Pyrosequencing in October 2002 claimed the focus now was “near term profitability”—no longer expansion and growth. However, sales continued to drop and, in the first quarter of 2003, the firm was far from profitable (see Fig. 3 below). In spring 2003, a new member joined the board with a mandate to restore the company to profitability and as a consequence to cut R&D spending significantly, and by June employees totaled 130. An engineer commented: “The board saw that their company was on their way to burning all its money. They had to do something” (Interview 28, interview 29, interview 8, interview 33). Pyrosequencing not only failed to develop an integrated system but despite its reagent kits gaining a reputation for good quality they proved too expensive. Following low sales, Pyrosequencing formed an alliance with another company, Corbett, to secure a more integrated system with added features to satisfy customers. The results were poor in relation to expectations and Pyrosequencing's shares plummeted further. In the search for near-term profits, but also because of the problematic experiences from developing a high throughput system, Pyrosequencing licensed its application to sequence the whole human genome to the US firm, 454 Life Sciences. Pyrosequencing would receive a royalty, depending on how 454's business developed. Thus another company was realizing Pyrosequencing's initial vision of providing an alternative to the Sanger sequencing method (Interview 26, interview 25, interview 7). The last months of 2003 proved turbulent. In August, Pyrosequencing merged (technically an acquisition by Pyrosequencing) with Personal Chemistry, another life science start-up in Uppsala owned by HealthCap and Investor Growth Capital. The CEO and most of the management team resigned. A commentator, experienced in the Swedish biotech industry, reflected that HealthCap and Investor Growth Capital wanted Personal Chemistry to acquire two critical resources: “That Pyrosequencing is listed on the stock exchange is important. By merging the two companies, this gives Personal Chemistry a “back door in” to the stock exchange with everything that means from a financial perspective, such as access to capital and liquidity in the share. The second resource that Pyrosequencing controls is, of course, its large amount of cash, which through the merger stays within the same group”. Later, following a second acquisition, Pyrosequencing changed its name to Biotage and created two divisions, Biosystems (Pyrosequencing) and Chemistry (the two other firms). Development activities within Pyrosequencing virtually ceased and its financial resources were instead used by the other division (Interview 24, interview 21, interview 2). A strategic review during 2003 and 2004 identified areas where users would appreciate the features of the Pyrosequencing method. One such area was Cpg methylation used in diagnostics to detect cancer. One inventor reflected that “For this, totally different features are needed. Only a few samples are needed, not thousands and cost per sample is not such an important parameter”. With hindsight, HealthCap realized that positioning the Pyrosequencing technology in the SNP market had been wrong. An investment

240

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

Year

Sales

Gross margin

Operating loss

1997

_

1998

23,125 USD

100%

1999

162,500 USD

80,7%

8,8 million USD

2000

5,8 million USD

77,0%

12,8 million USD

2001

13,5 million USD

71,8%

21,8 million USD

2002

12,2 million USD

67,2%

21,6 million USD

_ 4,4 million USD

Fig. 3. Sales, gross margins and operating results for Pyrosequencing for the period 1997–2002.

manager at HealthCap reflected: “We are back with the applications we discussed when we first became interested in the technology in 1996” (Interview 12, interview 1, interview 5; interview 17). Fig. 4 below illustrates the critical organizational and technical resources during the third phase of Pyrosequencing's innovation process. Thick lines indicate a close or strong relationship, while thin lines weaker ones.

Thus, six years from its foundation, Pyrosequencing had ceased to exist as an independent company. Even if the technology was a success within the scientific community, users had not sufficiently embraced it, expectations derived from forecasted sales had not materialized, and the firm had failed to survive alone. The technological review revealed that users had activated features of the technology differently than anticipated. Nevertheless, this remains an interesting route for some, since Qiagen, a German life

Fig. 4. Network of resources in the third phase of Pyrosequencing's innovation process.

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

science company, acquired Pyrosequencing (the Biosystem Division) in 2008. 7. Governing innovation in networks: analysis and discussion In this section we will analyze the case by addressing the issue of how the effects of VC governance unfolded during Pyrosequencing's innovation process. We focus first on the organizational resources (units and relationships) and then on the technical resources (products and facilities). Looking at the organizational resources and how VC governance might shape these, the VC firm's influence on Pyrosequencing as an organizational unit was not very explicit in the first phase. Nevertheless, prior to any VC investment, the Royal Institute of Technology researchers rebuilt their founding team. Further, they acquired a patent based upon recommendations from the VC firm. Hence, even before its foundation, VC influenced Pyrosequencing. In the second phase, the VC firm's experience of building start-up companies dominated the design of Pyrosequencing as an organizational unit. VC governance influenced the start-up firm's strategies, scope and controls, how and with whom it would interact, and therefore also its interfaces with external counterparts like suppliers and customers. The VC firm recommended a venture partner as the firm's first CEO, and it established controls and milestones that emphasized rapid product development following recommendations from the VC literature (e.g., Gompers & Lerner, 2001). VC money further enabled Pyrosequencing to increase its product development activities but only to develop SNP and related knowledge and skills. For example, Pyrosequencing's relationship with the Royal Institute of Technology was tense; as the researchers continued to develop the read length which they believed was a critical feature of the technology. From this perspective, Pyrosequencing substituted exploring new features of the technology for exploitation (March, 1991), that is, focusing on what the firm had already mastered—the SNP dimension of the method. Furthermore, the VC firm's focus on securing an IPO during the second phase influenced Pyrosequence's relationships with other units as the firm prioritized to develop an SNP application. The third phase started well for Pyrosequencing but ended in a merger with a portfolio firm owned by HealthCap and Investor Growth Capital. Ultimately, innovation and development work ceased. This period was dominated by the IPO in 2000. Initially VC demands made the start-up company reduce development time and freeze product features but the financial crisis from 2001 made it seek alternative applications for its technology. However, when sales slipped, stock exchange pressures induced cost cutting and eventually customers and the financial market lost confidence in Pyrosequencing. Pyrosequencing had become what in VC jargon is called “a living dead” (Bourgeois & Eisenhardt, 1987). The owners sought to protect what remained, that is, the large cash balances, rather than saving Pyrosequencing as an independent unit. To sum up, several studies using the resource interaction approach have illustrated that an organizational unit gain imprints from the close counterparts it interacts with over time. Most studies within this area (e.g., Baraldi, 2003; Håkansson & Waluszewski, 2002) concern interactions with customers and/or suppliers. This study shows that a financial logic not only influence an organizational unit that it hierarchically controls, but also how this logic can spread through the interactions with other actors that take place from the organizational unit's perspective. This takes us to the next resource in focus, relationships. The VC firm's involvement influenced relationships, not only Pyrosequencing's ability to develop relationships but, also relationships of other actors in Pyrosequencing's network. The VC governance spread in the network through relationships as carriers of this governance or

241

institutional logic (e.g., Lounsbury, 2008). For example, an effect of the way investors evaluate the VC firm is the tight time schedule put on Pyrosequencing during the three phases that we presented above. During the first phase the VC involvement influenced the formation of relationships as is evident in how the research groups at the Royal Institute of Technology combined their respective skills and resources to aid potential commercialization of their technology. Nyrén's group had fewer contacts within the business community than Uhlén's. His relationship with HealthCap was crucial for it secured VC finance which enabled the firm's foundation. VC influence on relationships became more explicit during the second phase from 1997 to 2000. The VC firm drew on existing relationships to help Pyrosequencing save development time. The VC firm's relationships with potential suppliers and customers were vital for Pyrosequencing's establishment of its relational interfaces. However, these relationships rested upon interactions of products, facilities and organizational units with other firms. For example, the venture partner used a set of relationships with organizations that he had used before in other start-ups. Another example of this is when the VCs encouraged Pyrosequencing to buy patents from another portfolio firm that was suffering financially. In the third phase, VC governance influenced Pyrosequencing's relationships even more. HealthCap's relationships with their investors made an IPO for Pyrosequencing a high priority. The IPO and the logic of the stock exchange, indirectly influenced how Pyrosequencing's relationships with customers evolved as the case illustrated. The Pyrosequencing and Personal Chemistry merger is of course the most obvious illustration of how the relationship between Pyrosequencing and HealthCap is embedded in a rather intricate network of financial ties on the one hand, and industrial relationships that Pyrosequencing tries to develop on the other. Here one can see how the VC firm systematically used its network position to add value to its portfolio as a whole. Hence, a VC firm sometimes uses its relationships to maximize the economic advantage of all its investments, which sometimes may conflict with a start-up's interests. Being financed by a VC can lead to a network of relationships that initially benefits the firm but subsequently creates conflicts of interests. This is an observation that mainstream VC literature (e.g., Gompers & Lerner, 1999; Shepherd & Zacharakis, 2001) would hardly make, due to its dyadic principal-agency view of the relationship between the VC and the entrepreneur. VC governance also affected Pyrosequencing's technical or physical resources and their relation to those of other actors. Any start-up must effectively integrate new facilities into its operations either by building its own facilities or utilizing external ones. To attract venture capitalists (or other investors) during the first phase the Royal Institute of Technology researchers related their invention to the HUGO project's facilities that read the human genome using the Sanger method. Challenging and relating to this method and associated facilities formed part of the inventors' sales pitch to VC firms. Venture capitalists influenced important decisions on the firm's scope during the second phase, for instance, the core competence, chemistry behind the reagent kits, and the decision not to develop and manufacture instruments and software. Due to the financial logic of the VC, Pyrosequencing out-sourced these operations and facilities. Further, following the forecast that the SNP market would grow exponentially, Pyrosequencing invested in a facility to manufacture reagent kits to reap the anticipated profits. This investment was long-term and involved retaining intellectual property resources to avoid others exploiting the innovation. When PSQ96 met users during the third and final phase, it became evident that outsourcing software development had been a mistake and it became an activity performed in-house, although this was also done to save money in troubled financial times. Furthermore, the board and management misinterpreted user patterns and demand for reagent kits: the latter was much lower than expected, particularly

242

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

from industry, which meant that the facility's capacity was barely used. As was discussed in the theoretical section, the VC literature, based on agency theory is not a theory of innovation (O'Sullivan, 2000). The use and development of facilities is a critical component in a firm's innovation journey (Van de Ven et al., 1999) as the case illustrates. Therefore it is rather remarkable that this literature does not pay more attention to how a start-up's products and facilities must fit in an existing or emerging technical system (Hughes, 1983). Lastly, although the researchers had developed a simple prototype at the Royal Institute of Technology, the product did not exist when the VC firm invested in Pyrosequencing during the first phase— it only existed as a potential product. The product was part of an idea structure with the ambition to get it embedded into an activated and physical structure (Baraldi & Strömsten, 2006; Håkansson & Waluszewski, 2002). The emergent firm had to sell this potential product, or idea, and its potential interaction with potential customers to the VC firm and other potential investors. The inventors' initial vision was to develop a general sequencing technology and they believed further development would resolve the unimpressive read length. But they needed something more tangible, with prospects of a growing market to sell the concept to venture capitalists, so management agreed to quickly “freeze” the new product and its features to speed up development, where for example possibilities to integrate customers' processes were not of high priority. Consequently, there was little opportunity to adapt it through the selling and buying interface. Pyrosequencing agreed to use the technology for SNPs, seeing this as a temporary step towards the greater aim—increasing the read length of the human genome, but the controls used within the firm and that focused on SNPs did not facilitate this in the second phase. The VC firm supplied Pyrosequencing with capital in stages depending on targets being met, which allowed product development activities to increase—but with a focus on SNP applications and analysis, not related activities. Nevertheless, early freezing of product features in the first and second phases did create economic advantages: without something tangible to sell to customers the IPO would not have happened. In the third phase, following its speedy development, PSQ96 soon connected with the technical interfaces of users because its interfaces were “open” in relation to the other technical resources in the customers' production processes: several instruments were sold in the first years of the product launch. Later in this phase, freezing features and interfaces between the PSQ96 and other resource interfaces in the network surrounding Pyrosequencing became problematic. PSQ96's principle features (easy to use, elegant and reliable) that were advantageous in early buying–selling interactions proved a drawback in the third phase and in producing–using interactions (Håkansson & Waluszewski, 2002) when the firm tried to embed PSQ96 into customers' production systems. Frozen features and interfaces between PSQ96 and customers' equipment precluded customers gaining integrated solutions across their entire technical system (Hughes, 1983), and incrementally upgrading and easily adding features. Speeding up the process from designing prototypes to developing an industrialized process in a large-scale facility can have positive effects, but only if the input and output of the new facility meshes with interfaces of other resources (Eisenhardt & Tabrizi, 1995). Pyrosequencing tried to develop a second generation system but this was more complex, dispersed, and increased interdependencies along technical and organizational interfaces (see Baraldi & Strömsten, 2006). The stock exchange listing and developing a complex system proved incompatible for a young company struggling to understand financial analysts' logic and how valuations affected its strategic discretion. Consequently, the new generation failed to add value or achieve positive economic results in resource combinations where it was usable.

To sum up, the industrial network and resource interaction frameworks helped reveal how VC governance can affect a startup's development, especially its organizational and technical interfaces with suppliers and customers (Håkansson & Waluszewski, 2002). The VC firm's governance mechanisms emphasized a speedy development cycle that had ramifications for resources (as in the 4R framework) and their interfaces. It influenced all resources (organizational units, products, production facilities and business relationships) but this only became apparent over time and during subsequent innovation phases. The management plans of the venture capitalists emphasizing speed seemed reasonable when adopted. However their effect upon resources and interfaces meant that these resources failed to mesh with the uncertain network environment and changing resource interfaces. Hence Pyrosequencing failed to embed and create a daily use for its products and services (Baraldi & Strömsten, 2006). The conflicting demands of an uncertain network and an accelerated innovation strategy proved problematic (see also Eisenhardt & Tabrizi, 1995). 8. Conclusions This study contributes to the literature on innovation by investigating how the governance by VC investors affected a start-up's innovation process. Further, the industrial network approach and the 4R framework detected effects beyond the single firm and its dyadic relationship with the VC firm as is often the case with studies based upon an agency theory perspective. The study highlighted the governance dimension of the 4R framework and demonstrates that relationships between VC firms and start ups, and how they are embedded, need greater recognition. This requires movement from a dyadic agency theory perspective that has long colored this literature. Mainstream VC research focuses on how the VC firm creates governance mechanisms to align the startup's interests with its own. The network perspective elaborates how this can influence a start-up's commercialization of new technology, especially when third parties in the network are involved. Our study makes it possible to draw the following conclusions regarding governance and innovation processes when relationships and networks matter. A first conclusion relates to the way a VC governs and controls a portfolio firm. This influences the innovation process in a profound way, directly and indirectly. The involvement of VC is akin to being in a boat that moves in a set direction at a certain speed. Its governance mechanisms, adherence to a tight time schedule, and eagerness to eliminate uncertainties is understandable (Gompers & Lerner, 2001): it follows advice from agency theory scholars to “meet uncertainty by identifying it in advance”. A VC firm with such skills can “get a better sense of the risks… set clear goals and timelines… communicate clearly… and think critically about financial and product market cycles”, thereby gaining a shorter and safer way to commercial success (Gompers & Lerner, 2001, p. 40). However, we can see from the case study that this perspective, fails to take into account the changing resource interfaces in an industrial network. For example interfaces between a product and other technical resources (forming a technical system) used by customers received low priority in order to hasten product development and proceed towards an exit for the VC firm and its investors. Hence, what seemed perfectly rational from the VC's perspective when it implicitly applied a principal-agency perspective with its static and dyadic view, did not make sense when the wider network and its many resource interfaces eventually entered the picture. A second conclusion relates to how portfolio firms act when they are under VC governance. We saw from the empirical case that the VC governance and the control it exercised led to an early freezing of

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

resource features as a way to save time in the development process. The academic world can cope with divergent approaches to empirical phenomena but in industry, as the Pyrosequencing story reveals, this is more difficult: a VC financed company must cope with linear and non-linear features of business life simultaneously. To “go with the flow”, or to embed new solutions in resource interfaces, firms engaged in technological and commercial development must create space for redirection, but to get VC finance for development, the start-up must produce detailed plans on what products, facilities, internal organization, and type of relationships it will develop. Thus, its only viable path must include both “going with the flow” and “setting clear goals and timelines”. Hence, being an innovative enterprise financed with VC is the same as living with two competing logics (e.g., Lounsbury, 2007) and there is a need to master both of them in order to survive. Furthermore, for the single company, redirection becomes more difficult as an embryonic company with innovative ideas acquires more tangible resources. The greater the variety in its resource base, the more it can try to embed new combinations in the user's technological and economic logic. This requires the start-up to develop organizational and technical interfaces in its surrounding network. Thus VC firms should govern the innovation process to ensure options evolve, and control the new venture so it has a possibility to redirect its innovation route accordingly. A third conclusion relates to the focus on speed and time-related controls. The case illustrated that achieving milestones within a predefined period was an important control mechanism in VC governance of a portfolio firm. Do time-related controls that emphasize identifying solutions at early stages of development have positive effects on creating economic value? The answer depends on the perspective chosen. From the VC firm's perspective, a quick exit is preferable when the contextual conditions permit so. This seems to be the case when the VC is young and needs to attract new investors. This seems also to have been the case in the present study. Hence a VC prioritizes a rapid establishment of a company and getting a product to a market. For the start-up firm, this can be a successful innovation journey if it develops products that immediately fit users' activity systems and resource interfaces. However, if the development proves inadequate and requires redirection, timeconstrained innovation can be detrimental. As the case illustrates, embedding a new technological solution into a structure of suppliers and users is seldom a quick fix. Even if the financial and industrial perspectives or logics can work in tandem, they are not always a perfect match. This study focused on one organization and its network. More research is necessary on how VC influences the innovation process when relationships and networks matters. Future research should continue to examine the interface between innovation and financial logics, represented by VC, and how these logics affect innovation on both micro as well as a broader level. For example, when do these logics match and when do they clash? Furthermore, the VC literature often claims that VC support is no guarantee of safety for a new project. According to Gompers and Lerner (1999), fewer than 2 in 10 VC financed projects survive the development journey. What role does VC governance play in this taken-for-granted pattern and survival rate of young ventures? Do companies that are funded by VC and fail provide products and solutions that lack long-run positive economic value or do they fail because their embedding processes do not match the logic of VC? Acknowledgements The authors thank Trevor Hopper and the three anonymous reviewers for their valuable and constructive comments. Financial aid was provided by Svenska Handelsbanken (Jan Wallander and Tom Hedelius research foundations).

243

Appendix A List of interviewees ID

Position

Company/Organization

Type of interview

Number of interviews

1 2

Manager Investment manager Investment manager Investment manager Sales support manager Product manager Product manager CEO Venture partner R&D manager CEO R&D manager R&D manager R&D manager Product manager Production manager Manager Purchasing manager Researcher Investment manager Financial analyst R&D manager CFO CEO Founder

Pyrosequencing HealthCap

Face-to-face Face-to-face

3 1

HealthCap

Face-to-face

1

HealthCap

Face-to-face

1

Pyrosequencing

Face-to-face

2

Pyrosequencing Pyrosequencing Pyrosequencing HealthCap/Pyrosequencing Supplier organization Supplier organization Customer organization Supplier organization Customer organization Pyrosequencing Supplier organization

Face-to-face Face-to-face Face-to-face Face-to-face Face-to-face Face-to-face Face-to-face Face-to-face Telephone Telephone Face-to-face

1 3 1 2 1 1 1 1 1 1 1

Pyrosequencing Pyrosequencing

Face-to-face Face-to-face

1 1

Customer organization Investor

Face-to-face Face-to-face

2 1

Brokerage firm Pyrosequencing Pyrosequencing Supplier organization Royal Institute of Technology Royal Institute of Technology Supplier organization Pyrosequencing

Face-to-face Face-to-face Face-to-face Face-to-face Face-to-face

2 1 1 2 2

Face-to-face

2

Face-to-face Face-to-face

1 2

Pyrosequencing

Face-to-face

1

Investor

Face-to-face

1

Supplier organization Uppsala university Investor Growth Capital

Face-to-face Face-to-face Face-to-face

1 1 1

Pyrosequencing Brokerage firm

Face-to-face Face-to-face

1 2

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

26 Founder 27 R&D manager 28 Key account manager 29 Head of manufacturing 30 Investment manager 31 Account manager 32 Researcher 33 Investment manager 34 CEO 35 Financial analyst

References Abbot A. What do cases do? Some notes on activity in sociological analysis. In: Ragin C, Becker HS, editors. What is a case? Exploring the foundations of social inquiry. Cambridge: University Press; 1992. p. 53–82. Alchian A, Demsetz H. Production, information costs, and economic organization. Am Econ Rev 1972;62:777–95. Amit R, Brander J, Zott C. Why do venture capital firms exist? Theory and Canadian evidence. J Bus Venturing 1998;13:441–66. Anderson J, Håkansson H, Johanson J. Dyadic business relationships within a business network context. J Mark 1994;58:1–15. Axelsson B, Easton G, editors. Industrial networks. A new view of reality. London: Routledge; 1992. Baraldi E. When information technology faces resource interaction. Using IT tools to handle products at IKEA and Edsbyn PhD Thesis. Sweden: Department of Business Studies, Uppsala University; 2003. Baraldi E, Strömsten T. Embedding, producing and using low weight: value creation and the role of the configuration of resource interfaces in the networks around IKEA: s Lac and Holmen's newsprint. IMP J 2006;1:52–97. Baraldi E, Strömsten T. Controlling and combining resources in networks. From Uppsala to Stanford, and back again. The case of a biotech innovation. Ind Mark Manag 2009;38:541–52.

244

T. Strömsten, A. Waluszewski / Journal of Business Research 65 (2012) 232–244

Baraldi E, Waluszewski A. Information technology at IKEA: an “open sesame” solution or just another type of facility? J Bus Res 2005;58:1251–60. Barney JB, Fiet JO, Busenitz LW, Moesel DD. The substitution of bonding for monitoring in venture capitalist relations with high technology enterprises. J High Technol Manage Res 1996;7:91–105. Basalla G. The evolution of technology. Cambridge: Cambridge History of Science Series; 1988. Bijker WE. Of bicycles, bakelites and bulbs. Toward a theory of sociotechnical change. Cambridge: MIT Press; 1997. Bourgeois LJ, Eisenhardt KM. Strategic decision processes in Silicon Valley: the anatomy of a living dead. Calif Manage Rev 1987;30:143–59. Davila T. An exploratory study on the emergence of management control systems: formalizing human resources in small growing firms. Account Organ Soc 2005;30: 223–48. Dosi G. The nature of the innovative process. In: Dosi G, Freeman C, Nelson R, Silverberg R, Soete L, editors. Technical change and economic theory. London: Pinter; 1988. Dubois A, Araujo L. Research methods in industrial marketing studies. In: Håkansson H, Harrisson D, Waluszewski A, editors. Rethinking marketing: developing a new understanding of markets. Chichester: John Wiley & Sons; 2004. Dubois A, Araujo L. The relationship between technical and organisational interfaces. IMP J 2006;1:28–51. Dyer WG, Wilkins AL. Better stories, not better constructs, to generate better theory: a rejoinder to Eisenhardt. Acad Manage Rev 1991;16:613–9. Easton G. Methodology in industrial networks. In: Möller K, Wilson D, editors. Business marketing: an interaction and network perspective. Norwell MA: Klüwer; 1995. p. 411–92. Eisenhardt KM. Building theories from case study research. Acad Manage Rev 1989;14: 532–50. Eisenhardt KM, Graebner ME. Theory building from cases: opportunities and challenges. Acad Manage J 2007;50:25–32. Eisenhardt KM, Tabrizi BN. Accelerating adaptive processes: product innovation in the global computer industry. Adm Sci Q 1995;40:84–110. Fama EF, Jensen MC. Separation of ownership and control. J Law Econ 1983;26:301–25. Forbord M. New uses of an agricultural product? A case study of development in an industrial network. R-5/03. Trondheim: Centre for Rural Research; 2003. Ford D, Gadde L-E, Håkansson H, Lundgren A, Snehota I, Turnbull P, et al. Managing business relationships. hichester: John Wiley & Sons; 1998. Freeman J. Venture capital as an economy of time. In: Lenders R, Galbay S, editors. Corporate social capital. Norwell, MA: Kluwer; 1999. p. 460–79. Fried VH, Bruton GD, Hisrich RD. Strategy and the board of directors in venture capital backed firms. J Bus Venturing 1998;13:493–503. Gompers P, Lerner J. The venture capital cycle. Cambridge: The MIT Press; 1999. Gompers P, Lerner J. The money of invention. Cambridge: Harvard Business School Press; 2001. Granovetter M. Problems of explanation in economic sociology. In: Nohria N, Eccles R, editors. Networks and organizations. Structure, form and actionBoston: Harvard Business School Press; 1992. Gressetvold, E. (2004). Developing relationships within industrial networks—effects of product development. Phd Thesis. Trondheim: Department of Industrial Economics and Technology Management, Norwegian University of Science and Technology. Håkansson H. Industrial technological development. A network approach. New Hampshire: Croom Helm; 1987. Håkansson H, Lind J. Accounting and network coordination. Account Organ Soc 2004;29:51–72. Håkansson H, Snehota I. Developing relationships in business networks. London: Routledge; 1995. Håkansson H, Waluszewski A. Managing technological development. London: Routledge: IKEA, the Environment and Technology; 2002. Håkansson H, Waluszewski A, editors. Knowledge and innovation in business and industry. The importance of using others. London, New York: Routledge; 2007. Halinen A, Törnroos J-Å. The role of embeddedness in the evolution of business networks. Scand J Manag 1998;14:187–205. Harrison D, Waluszewski A. The development of a user network as a way to re-launch an unwanted product. Res Policy 2008;37:115–30. Hjelmgren, D. (2005). Exploring the interplay between standard products and customer specific solutions. PhD Thesis, Gothenburg: Chalmers tekniska högskola. Holmen, E. (2001). Notes on conceptualisation of resource related embeddedness of interorganisational product development. PhD Thesis, Institute of Marketing, University of Southern Denmark. Hughes TP. Networks of power: electrification in western society (1880–1930). Baltimore: John Hopkins University Press; 1983. Ingemansson, M. (2010). Success as science but burden for business? PhD Thesis. Uppsala: Department of Business Studies, Uppsala University. Ingemansson M, Waluszewski A. Success in science and burden in business. On the difficult relationship between science as a developing setting and business as a producer–user setting. IMP J 2009;3:20–56.

Jain B. Predictors of performance of venture capital-backed organizations. J Bus Res 2001;52:223–33. Laage-Hellman J. Technological development in industrial networks. Doctoral thesis. Sweden: Uppsala University; 1989. Lind, F. (2006). Resource combining across inter-organisational project boundaries. PhD Thesis. Gothenburg: Chalmers Tekniska Högskola. Lind J, Strömsten T. When do firms use different types of customer accounting? J Bus Res 2006;59:1257–66. Lounsbury MA. A tale of two cities. Competing logics and practice variation in the professionalizing of mutual funds. Acad Manage J 2007;50:287–307. Lounsbury MA. Institutional rationality and practice variation: new directions in the institutional analysis of practice. Account Organ Soc 2008;33:349–61. Lundgren A., (1991), Technological innovation and industrial evolution: the emergence of industrial networks. PhD Thesis. Stockholm: Stockholm School of Economics, EFI. March JG. Exploration and exploitation in organizational learning. Organ Sci 1991;2: 71–87. Melamade, R. J. (1985). US Patent 4863849. Murray GC. A synthesis of six exploratory, European case studies of successfully exited, venture capital-financed, new technology based firms. Entrepreneurship Theory Pract 1996;20:41–60. Nyrén P. (1987) Enzymatic method for continuous monitoring of DNA polymerase activity. Anal Biochem 1987;167:235–8. Nyrén, P. (1998/2001). Method of sequencing DNA patent WO 9828440/US 6258568. Nyrén P. The history of Pyrosequencing. In: Marsh S, editor. Methods in molecular biology, vol. 373. Totowa, NJ: Humana Press Inc.; 2006. O'Sullivan M. The innovative enterprise and corporate governance. Camb J Econ 2000;24:393–416. Penrose E. The theory of the growth of the firm. New York: John Wiley and Sons; 1959. Perry LT. The capital connection: how relationships between founders and venture capitalists affect innovation in new ventures. Acad Manage Exec 1988;2:205–12. Powell WW, Koput KW, Bowie JI, Smith-Doerrs L. The spatial clustering of science and capital: accounting for biotech firm–venture capital relationships. Reg Stud 2002;36:291–305. Ronaghi M, Uhlén M, Nyrén P. DNA sequencing. A sequencing method based on realtime pyrophosphate. Science 1998;281:263–365. Rosenberg N. Inside the black box. Technology and economics. Cambridge: Cambridge University Press; 1982. Sahlman W. The structure and governance of venture capital organizations. J Financ Econ 1990;27:473–521. Sapienza HJ, De Clercq D. Venture capitalist–entrepreneur relationships in technology based ventures. Enterprise Innov Manage Stud 2000;1:57–71. Sapienza HJ, Manigart S, Vermeir W. Venture capitalist governance and value added in four countries. J Bus Venturing 1996;11:439–69. Shepherd DA, Zacharakis AL. The venture capitalist–entrepreneur relationship: control, trust and confidence in co-operative behaviour. Venture Cap 2001;3:129–49. Shih, T. (2009). Scrutinizing a policy ambition to make business out of science. Lessons from Taiwan. PhD Thesis. Uppsala: Department of Business Studies. Uppsala University. Smith DA, Lohrke FT. Entrepreneurial network development: trusting in the process. J Bus Res 2008;61:315–22. Stinchcombe A. Social structures of organization. In: March J, editor. Handbook of organizations. Chicago: Rand McNally; 1965. p. 153–93. Strömsten T, Håkansson H. Resources in use. Embedded electricity. In: Håkansson H, Waluszewski A, editors. Knowledge and innovation in business and industry. The importance of using others. London, New York: Routledge; 2007. p. 27–53. Tidd J, Pavitt K, Bessant J. Managing innovation: integrating technological markets and organisational change. Chichester: Wiley; 1997. Timmons JA, Bygrave WD. Venture capital's role in financing innovation for economic growth. J Bus Venturing 1986;1:161–76. Van de Ven AH, Polley DE, Garud R, Venkataraman S. The innovation journey. New York: Oxford University Press; 1999. Wagrell S, Waluszewski A. The problem of using (medical) innovations. IMP J 2009;3: 57–85. Waluszewski A. A competing or co-operating cluster or seven decades of combinatory resources? What's behind a prospering biotech valley? Scand J Manag 2004;20: 125–50. Waluszewski A, Baraldi E, Shih T, Linné Å. Resource interfaces telling other stories of the commercial use of new technology: the embedding of biotech solutions in US, Taiwan and China. IMP J 2009;3:86–123. Wedin, T. (2001). Networks and demand. The use of electricity in an industrial process. PhD Thesis. Uppsala: Department of Business Studies, Uppsala University. Yin RK. Case study research: design and methods. Newbury Park, CA: Sage; 1989.