This research adopts the commonly accepted definition of a project as ‘a temporary undertaking to create a unique product’ (PMI, 2000). The undertaking is temporary because it has defined start and end dates, and it is unique because its purpose is to fulfil a specific requirement. Its performance is typically measured on four dimensions: time, cost, quality and functionality (Kerzner, 1998; Schwalbe, 2002; Turner, 1993).

Based on these definitions, IS project performance to date has been poor. Table 1 reports the findings from a series of longitudinal surveys conducted by the Standish Group since 1994. So-called ‘challenged’ projects are defined as being over budget, over schedule, or under specification. It should be noted, however, that these measures are against ex ante estimates of project time, cost, quality and functionality, which are affected by other dimensions including the socio—technical complexities involved with major projects and the human ability to produce accurate predictions (Kahneman et al., 1982). The research program on which we draw in this paper is broad, including studies into IT planning and IT investment processes and the way in which project managers effectively decompose, structure and sequence project and business outcomes.

The data in Table 1 show that there is a large disparity between achieved and projected performance, and that learning has been slow. This is consistent with other research on IS project success. For example, Field (1997) finds that about 40% of projects are cancelled before completion, and Ambler (1999) reports that some practitioners claim that, for large scale, mission-critical software projects, the failure rate has been as high as 85%.

Much of the accumulation of practical knowledge in IS project management has been driven by practitioners, who have amassed their collective knowledge of ‘successful’ practices into Bodies of Knowledge (BOK) such as the US-based Project Management Institute’s (PMI) Project Management Body of Knowledge (PMBOK) (PMI, 2000). However, these bodies of knowledge lack a strong explicit theoretical base. In addition, there is often little formal evidence of the success of the espoused practices.

In an analysis of the implicit theory underpinning these bodies of professional knowledge, Williams (2005) identifies three meta-theoretical assumptions that characterise the dominant discourse in current project management. Table 2 presents these assumptions.

Koskela and Howell (2002) review the theories that underpin project management as espoused in the PMBOK and that are frequently applied in practice. They show that the espoused practice rests on three theories of management: management as planning, the dispatch model of execution and the thermostat model of control. They conclude that these implicit and narrow theories are of limited value and explanatory power. They also note that they have already been superseded in the original management field from which they were imported.

In summary, practice dominates IS project management, with weak underpinning theory that could be developed, extended and enriched to improve project performance.

Kilduff (2006) argues that ‘the route to good theory leads not through gaps in the literature but through an engagement with problems in the real-world that you find personally interesting’. He reiterates the observation of Hambrick (2005) that influential theories derive from the observation of real-life phenomena, not from ‘scholars struggling to find holes in the literature’.

When motivated by a hole in the literature, researchers generally start with a problem within an existing theory, extend or refine it in some way, and apply it to a specific context (Kuhn, 1996). The nature of this learning and knowledge creation is ‘exploitation’ of the existing theory (March, 1991), including processes captured by terms such as refinement, choice, production, efficiency, selection, implementation and execution. Alternatively, researchers can address a gap in theory by starting with a new theory and testing it in a specific context. The nature of this learning and knowledge creation is ‘exploration’, including processes such as search, variation, risk taking, experimentation, play, flexibility, discovery, and innovation.

When motivated by the observation of practice, the problem is practice-driven, framed by the phenomena rather than by a well defined research model (Zmud, 1998). By not adopting a well-defined research model ex ante, this approach acknowledges that the research team does not know, a priori, the solution or the theory to be developed. Practice-led research that uses existing theories to codify best practice is exploitative. This includes research that seeks to improve project performance by developing more methodologies, better execution and stronger governance. Alternatively, and the focus of this paper, the research can be exploratory, looking to solve problems by drawing on new theories, frequently borrowed from other research domains. Integrating the two categories of theory development motivation (Kilduff, 2006) and learning and knowledge creation (March, 1991), Figure 1 presents a framework with exploratory practice-driven research located in the bottom right-hand quadrant.

Progress typically involves a mix of research approaches from all of the quadrants in Figure 1. Kuhn (1996) outlines the importance of the existing paradigm for conducting ‘normal science’, with punctuated changes to the status quo involving ‘paradigm changes’. March (1991) contends that maintaining an appropriate balance between exploration and exploitation efforts is necessary for system survival and prosperity. This paper argues that the research in the bottom right hand quadrant holds great potential for unlocking the intransigent problems of IS project management performance, and that the current literature under-utilises it.

Exploratory practice-driven research provides an environment for researchers and practitioners to collaborate, with the objective of solving a specific problem in practice and developing new theory, thus producing research that is both rigorous and relevant. In contrast to the usual debate around binary choices of rigor/relevance or theory/practice, this approach is simultaneously pursuing good science, which leads to new understanding, and practical solutions to critical problems. This is also known as Pasteur’s quadrant (Mason, 2001).

In our view, exploratory practice-driven research is a high-potential approach to developing theory in the IS project management context where:

Practice-led research has an established acceptance and use within the IS discipline and various research designs are available to conduct this style of research, including case study and action research. Techniques to tackle the theory building process are rooted in the classic grounded theory paradigm from Glaser and Strauss (1967) and subsequent developments and debates — see Glaser (1992) and Strauss and Corbin (1997) . A notable application of grounded theory is Eisenhardt (1989), which provides an accessible framework for building theory based on case study research.

However, exploratory practice-driven research has not been extensively utilised in IS project management research. Nor are there clearly articulated steps to follow for theory development to derive theory from gaps in its practice. Instead, IS project management research has generally focused on holes in the existing literature, with researchers exploiting the limited existing theory to develop factor and process models (Markus and Robey, 1988; Robey and Newman, 1996; Sauer, 1999). As contexts appear where these do not hold, researchers introduce contingencies (Shenhar, 1998, 2001) and the various Bodies of Knowledge expand in detail and coverage. For example, as early as 1997, there were over 1,000 methodologies in use by the IS community (Fitzgerald, 1998).

Exploratory research intended to displace the existing dated theory has been limited, with a few notable exceptions such as the application of Adaptive Control Theory (Alleman, 2002) and a growing body of literature on the application, often using simulation techniques, of Complex Systems (Benbya and McKelvey, 2006; Morris, 2002; Williams, 2005). Exploratory practice-led research is all but absent in the IS project management literature.