Many organisations evolving in hyper-competitive environments need to design effective innovation strategies to deliver the fittest products and services to their clients (Ilinitch et al., 1996; Atamer, Durand and Raynaud, 2005). In this context, the capabilities required for inventing and commercialising new products and services are critical (Maupertuis, 1999). This is particularly so when the development of innovations involves a variety of knowledge available inside and outside the firm’s boundaries (Chesbrough, 2003; Baudry, 2004). Although the literature on innovation management has long acknowledged the role played by capabilities in providing the firm with competitive advantage (Wernerfelt, 1984; Prahalad and Hamel, 1990; Barney, 1991; Teece, Pisano and Schuen, 1997; Dosi, Faillo and Marengo, 2008), this literature does not rely on an unambiguous understanding of how firms manage to develop innovation capabilities, in particular within knowledge-intensive sectors. For example, Freiling, Gersch and Goeke (2008) identified dozens of conceptualisations belonging to the competence-based theory of the firm (e.g. resource-advantage, resource-based view, competences, capabilities, dynamic capabilities, competence-based strategy, routines, organisational learning...), each referring to distinctive paradigms (e.g. evolutionary theory, behavioural theory, strategic management...). This lack of coherence is often reinforced by the combination of competence-based researches with established neo-classical and transaction-costs theories of the firm (Barney, Wright and Ketchen, 2001; Foss and Foss, 2005) and entrepreneurship (Alvarez and Busenitz, 2001; Augier and Teece, 2008). As Eisenhardt and Martin noticed, the resource-based view has been called “conceptually vague and tautological, with inattention to the mechanisms by which resources actually contribute to competitive advantage [...]; it has also been criticized for lack of empirical grounding” (Eisenhardt and Martin, 2000, p. 1106).

The literature on innovation has reported new trends in the management of R&D and innovation. The most representative approaches of these new trends include user-driven innovation (von Hippel, 1988), community-based innovation (Brown and Duguid, 1991), innovation in complex product systems (Hobday, 1998), design-oriented views on innovation (Grandori and Furnari, 2008), and open innovation (Chesbrough, 2003). Despite heterogeneous theoretical and empirical backgrounds, these conceptualisations of innovation share a common assumption with regard to the virtues attached to the concepts of knowledge and capability. The user-driven model (von Hippel, 1988) for example, claims that integrating final users with R&D activities provides the firm with critical knowledge on user needs, constraints and preferences. Furthermore, user-driven modes of management are likely to support interactions among product designers, suppliers and customers (Scheid, 2007), and seemingly improve organisational responsiveness and flexibility. In this context, scholars have posited that community forms provide the firm with congenial models to manage knowledge-intensive activities (Brown and Duguid, 1991; Cohendet, Creplet and Dupouet, 2003; O’Mahony and Ferraro, 2007). Community-based governance, for example, is well established in the software industry where free/libre and open source initiatives (F/LOSS) have flourished in the last two decades (Foray and Zimmermann, 2001; Grand, von Krogh, Leonard, and Swap, 2004), in biotechnology (Shan et el., 1994) and in health care sectors (Lettl, Herstatt and Gemuenden, 2006). The importance of knowledge and capabilities has also been acknowledged by researchers investigating innovation in complex product systems (CoPS, Prencipe and Hobday, 2003). This literature suggests that the principles of dis-aggregation enable the firm to design modular product architecture which reduces communication and coordination costs (Hobday, 1998). Therein, the dis-aggregation of a complex system into discrete modules often triggers the creation of a distributed community of suppliers (e.g. modular networks, industrial districts; Langlois and Robertson, 1992) which collaborate to develop new products and services and bring them to the market (Frigant, 2007). This approach to innovation leads the firm to profit from critical resources available outside its organisational boundaries (Chesbrough and Teece, 1996) and thus requires being able to coordinate with a range of players involved in the production and diffusion of innovative outcomes (Cohendet, Diani and Lerch, 2005).

To study the research questions addressed, we conducted a single illustrative case study (Yin, 2003). We used a bundle of secondary data generated by contingent researchers, observers and participants involved in a project. Similar to Weick’s (1993) study of the Mann Gulf disaster, we built on secondary data to reconstruct the history of the OOS project and re-analyse the nature and logics of the capabilities involved in the development of the OOS. This research strategy allowed us to go deeper into initial questions addressed by researchers implicated in the project (e.g. is the OOS project open source?) and to address new research questions (e.g. what types of organisational capabilities are attached to OOS as an example of an innovation project?). Since we were not initially involved in the project, we sought to address the research question by interpreting data that have been formalised by the players involved in the project (e.g. corporate representatives, project managers, contributors, users, and scientific advisors).

The OOS initiative was launched in 1996 by the U.S. Army in order to (i) streamline its modelling and simulation software acquisition and development cycles, (ii) reach a larger set of users and (iii) allow for greater flexibility in combining and integrating various software applications, products and scenarios. The project manager designated by the Deputy Commanding General (DCG) was the Army Program Executive Office for Simulation, Training, and Instrumentation (PEO STI). The PEO STI aimed to leverage internal as well as external resources and technical skills, in particular by connecting the OOS military community with external modelling and simulation software users and developers (Wittman and Surdu, 2004). Long before the Army had awarded contracts to industrial suppliers, a set of government representatives was teamed up to articulate final users’ technical needs and requirements. The Architecture-Integrated Product team (A-IPT) comprised the project manager (Army representative), a small set of lead users (Army agencies), a government think tank (MITRE) and a Science and Technology Company (Alion). In April 1999, the A-IPT began to codify the initial architecture concepts. In December 1999, the A-IPT elected the prime contractor of the OOS: the Science Applications International Corporation (SAIC). This company was selected by the Army as it shares Lead Systems Integrators (LSI) status with The Boeing Company on the Army’s largest transformational program, called the Future Combat System (FCS). Therefore, the OOS project is tethered to the FCS program and forms a major building block of the U.S. Army’s transformation programme.

The next sections present our case study findings. First, we focus on heterogeneous users’ skills and profiles. Second, we emphasise the role played by flexible design and open architecture. Third, we insist on the use of a variety of digitised and non digitised technological artefacts to manage knowledge and disseminate innovation packages. Finally, we show how the accessibility, alterability and marketability of the OOS product relied on restrictive license terms.

The OOS fulfils three different user needs within the U.S. Army: training the forces (Training, Exercises and Military Operations, TEMO), assessing the forces (Research, Development and Acquisition, RDA), and testing the U.S. military Doctrine (Advanced Concepts and Requirements, ACR). The U.S. Army Training and Doctrine Command (TRADOC) Program Office OneSAF (TPO) has been elected to define specific cross-domain requirements with support from user communities. Together with the Army, the U.S. Marine Corps have adopted –and adapted- the OOS for use in their Combined Arms Staff Trainer. The U.S. Air Force and the U.S. Navy have also developed modelling and simulation systems that are fully inter-operable with the OOS. In addition to military use and sponsorship, the system remains accessible to non military organisations (e.g. research and education, university) providing that they agree on restrictive license terms articulated in the OneSAF Distribution Agreement (DA). Finally, the user domain has been broadened to include international collaborations with Allied nations. In particular, the OOS has been integrated to the ABCA (Australia, Britain, Canada and United-States of America) Standardization Program in order to reduce the likelihood that interoperability failures occur.

The central property of the OOS architecture is flexible composition. The flexible composition of a particular system configuration occurs through the integration of different products, each product involving the combination of several components selected from appropriate repositories thanks to dedicated support services. Specifically, the OOS architecture is made up with seven layers:

One of the outstanding features of the OOS is that components, applications and products available to final users are not exclusively based on proprietary (copyrighted) or open source software (copy-left). For example, the OneSAF website is operated by Apache web server and the e-mail interface which has been chosen to diffuse newsletters is Mailman. These software applications operate under open source licenses (i.e., GPL). At the same time, XMLspy (i.e., proprietary software) has been elected by the project manager to provide users with an effective application to design suitable system configurations (Parsons and Wittman, 2005). All these components, applications and products aim at maintaining the level of attraction, flexibility and usability, while stimulating interactions and feedbacks between developers and users. Furthermore, since 2004, the annual OneSAF User Conference has been established to provide the global OOS user community with the opportunity to meet and exchange ideas, experiences and best practices related to their own utilisation of the system. As a result, the level of technical skills required to use the OOS has been reduced to reach a larger set of users and the dialogue between military expert users and software developers have been strengthened.

The OneSAF Distribution Agreement (DA) imposes significant restrictions on the alteration and distribution of the source code. While the latter is made available for users who are free to modify it and develop their own scenarios and applications, any significant modifications of the source code have to fit the corporate sponsor’s expectations before being controlled and validated by the project manager. Furthermore, profit-seeking commercial exploitations of the OOS are strictly banned. According to the articles 3 and 4 of the Distribution Agreement Distribution Agreement (v.2.0) of the OneSAF Product Line Configuration, “software and data will not be modified, adapted, or otherwise altered except as required for the prior approved purpose […] OneSAF Software shall not be re-distributed, sold, or used for commercial purposes […], in whole or in part, without the express permission of the PEO STRI PM OneSAF […] The recipient is prohibited from using the OneSAF open source code for commercial practices and profitable gains” [1].

As a knowledge intensive activity, innovation requires specific capabilities to be developed and brought into play by the organisation. Specifically, when public and private organisations develop innovation in knowledge-intensive sectors they need to develop new organisational capabilities which enable them to manage freely available knowledge, access the community making contributions to disseminate useful knowledge, and assemble, distribute and monitor innovation outcomes for which restrictive property rights have been established. The next section elaborates on the nature and logics of these new organisational capabilities the firm must hold and develop in order to develop innovations.

Basically, capabilities “emphasizes the key role of strategic management in appropriately adapting, integrating, and reconfiguring internal and external organisational skills, resources, and functional competences to match the requirement of a changing environment” (Teece, Pisano and Schuen, 1997, p. 515). The foregoing suggests that the management of innovation processes entails specific capacities the firm must possess in order to (re)-combine and align a variety of resources available inside and outside its own organisational boundaries (Galunic and Rodan, 1998). When deployed effectively, these capabilities are likely to enhance the firm’s ability to invent and commercialise new products and services. In her study of a series of development projects, Leonard-Barton (1992) distinguished four types of dimensions that must be aligned by an organisation in order to become core and deliver strategic value. The first dimension is related to the personal knowledge and technical skills that are hardly transmissible and remain tacit. The second dimension is concerned with the technical systems which enable the accumulation, codification, classification, and communication of knowledge. The third dimension involves the managerial systems which represent formal and informal ways of creating and controlling knowledge and coordinating individual tasks. The fourth –and last- dimension is connected to the values and norms which are infused through the three dimensions above and correspond to the value assigned to the content of knowledge, artefacts, and governance mechanisms that are deployed and used within the firm. Drawing on our case study findings, we use Leonard-Barton’s (1992) conceptualisation of the firm’s core capabilities to develop a typology of the dimensions the firm must hold and align in order to meet innovation requirements within knowledge intensive environments (Barbaroux and Le Texier, 2007). The typology is presented in Table 2. Each dimension exhibits specific traits that are shaped by the attributes of the innovation project.

The first dimension is related to participants’ skills. From our case study findings, at least three categories of individual participant can be differentiated according to their technical level of expertise (von Krogh, Spaeth and Lakhani, 2003). First, low skilled users participate in the project as final users of the product. They do not directly participate in development activities, but provide critical feedbacks (e.g. signalling of technical bugs and requests for add-ons) to achieve projects’ sustainability. Examples of low-skilled users involved in the OOS project include model composers and scenario developers. Second, intermediary skilled users are active contributors whose technical knowledge skills are higher than that of the previous category of users. System administrators and data managers exemplify how intermediary skilled users participate in the OOS project. Such category of individual users provides technical-based assistance and offers specialised applications that fit with their own needs or fulfil specific users’ demands. As such, they are likely to promote product customisation through “problem-solving”-oriented attitudes. Third, high skilled users represent a class of individual participants having high-level technical capabilities. Their architectural know-how combined with local technical domains of expertise, gives them legitimate responsibility and authority within the global user community. In the OOS project, software developers and technical managers represent this category of highly skilled users.

The second dimension refers to the ways innovation outcomes are produced. Participants in the OOS project used a bundle of the technical systems (e.g. PLAF, TRD, OCD, RDG, Mailman, XMLspy) which enhanced the coordination of dispersed activities and fostered the dissemination of product packages to adopters. The OneSAF User Conference, which has been organised every year until 2004, can also be construed as an example of technical systems that foster interactions across distinctive user communities. Basically, technical systems rely on the combination of artefacts and contexts to support the codification, diffusion and sharing of knowledge. Using documentation, user manuals and peer-based discussions is vital for the community. Indeed, both product architecture and technical functionalities have to be unambiguously written, stored and accessed in order to be re-used. Our case study shows that the codification of knowledge is essential for any newcoming individual participant to join the user community. Although newcomers present sufficiently high technical skills to take part in development processes, contributions tend to be extremely complicated without technical systems assistance (cf. Varian and Shapiro, 2003). Documentation is also critical for low skilled users to implement software solutions in a simple and effective way. Finally, the dissemination and sharing of critical knowledge is likely to be enhanced through mailing lists and other electronic artefacts. In the OOS community, users obtained important information related to problem-solving procedures, for example by using the XMLspy software and technology.

The third dimension refers to managerial systems, incentive systems and reporting structures implemented by the corporate authority to manage innovation. Basically, the extent to which authority is distributed is essential to achieve success (West, 2003). The OOS project hinged on particular managerial mechanisms that tended to combine authority-based hierarchy with peer-based governance mechanisms. Depending on the form and legitimacy of the authority prevailing in the community, sponsorship and leadership strategies are likely to achieve technical excellence and/or fulfil users’ expectations (O’Mahony and Ferraro, 2007). Governance mechanisms are also a determinant of the capacity of the firm to capture value from innovation. The Distribution Agreement (DA) that ruled the conditions for accessing, altering and commercialising the OOS, is illustrative of the pervasiveness of hybrid strategy in the software industry (Shah, 2006). Together with the use of appropriate technical systems, well-designed governance mechanisms are essential to promote user engagement and harness a variety of feedbacks from heterogeneous user communities (MacCormack, Rusnak and Baldwin, 2006).

The fourth dimension is concerned with participants’ attitudinal profiles. The OOS innovation depicts a knowledge-intensive activity which is based on particular values and norms. Among these values, technical excellence seemingly influenced individual participants in the OOS project. Indeed, customisation resulting from the contributions of technically-oriented users improved software quality to the benefit of contributors themselves and the OOS community as a whole. Although OOS is a defense-oriented program, it is remarkable that the socio-cognitive values endorsed by individual participants are commonly shared within open source software communities. Technical excellence, reputation enhancement, learning, and even gift-oriented behaviours are widespread in the OOS user community. Amazingly, the willingness to share time and knowledge for “ideological” reasons, which is considered as a central value within the open source movement (Bagozzi and Dholakia, 2006), is also critical in the OOS user community. However, the “ideological” motives shared by the participants in the OOS project certainly differ from those shared by those individuals involved in open source initiatives…

The main contention of this article was that innovation requires specific capabilities. In particular, the firm must be able to manage heterogeneous resources which are often distributed and incorporated in a variety of units and organisational forms (e.g. individuals, communities, corporate firms, networks). The foregoing requires that the firm is capable of designing and aligning products and organisational forms, and of capturing value from the commercialisation of innovation outcomes (cf. Table 1). Investigating the case of a modelling and simulation software used by the U.S. Army for education and training purposes, we built on Leonard-Barton (1992) to propose a typology of the dimensions involved in the deployment of those capabilities required to innovate within knowledge-intensive environments. This conceptualisation includes four dimensions: heterogeneous skills, diverse communication and codification artefacts, hybrid governance mechanisms and a variety of attitudinal profiles, values and norms (cf. Table 2). These results are supported by Dodgson, Gann and Salter’s (2005) analysis of the relations between new technology and innovation. The authors stated that “the competitive advantage to be derived by firms from innovation lies with the creative leadership in design and development linked with effective integration of other productive functions, the capacity to manage complexity, and in the ability to fully engage the users of innovation in the process of its realization” (Dodgson, Gann and Salter, 2005, p. 24). But while Dodgson, Gann and Salter focused on the role played by information and communication technology (ICT) in supporting superior innovation capabilities, our conceptualisation considers the use of digitised and non-digitised technical artefacts as one among four dimensions the firm has to handle in order to deploy innovation capabilities. The key implication of this research for both scholars and practitioners is that it is the capacity of the firm to handle and align these dimensions that best represents what innovation capabilities truly are.