A further factor influencing the recommendation in this paper is the view that there is a weakness in existing portfolio-level IS theory that has contributed to management’s loss of faith in IS. That the management of IT issues is important from a bottom line perspective (Luehrman, 1997) does not equate to an interest in the technology or its applications for their own sake. The ongoing commitment to IT outsourcing by organisations in both the business and government sectors, now extending to encompass the off-shore transfer of some functions, seems indicative of a general loss of belief that IT is strategically significant (Hirschheim and Klein, 2003; Stewart, 2003). As has been pointed out by various theorists (for instance Hendry, 1995; Harvey and Lusch, 1997), organisations do not generally outsource functions they perceive to be part of their strategic core.

IS has had theories that address the relationships between IS and organisational structures and strategies. The strategic IS planning literature was a vibrant one at a time when theories of competitive advantage (Porter and Millar, 1985; Kettinger et al., 1995) and of the benefits of IS integration (Segars and Grover, 1996) were in fashion. But the theories on which the publications in these areas were based ultimately failed to convince, and the number of papers being produced on portfolio-level theory has fallen drastically. The problems these theories encountered are, however, useful to indicate in which directions the development of a general IS theory might go.

In broad terms, competitive advantage theory appears to have foundered on at least two related problems. These are the issues of imitation and structural change, which together refute the idea that IT applications can generally be considered to be reliable competitive instruments. What the available evidence shows is first that it is in most cases at least as good to be a fast IT imitator (i.e. to wait and copy a promising innovation, usually at a lower cost) than it is to be a first mover (Vitale, 1986; Clemons and Row, 1988), and second that IT innovations usually operate to effect structural industry change rather than entrench specific competitive edges (Copeland and McKenney, 1988; Kettinger et al., 1995; Clemons and Row, 1988). Both these findings have been available for some time, and have not been refuted.

The weaknesses of theories dealing with systems integration are less obvious in that they deal with ideal structures (Martin, 1990; Wyzalek, 2000), theorists have acknowledged the practical difficulties of achieving integration goals (Segars and Grover, 1996; Hamilton, 1999), and the integration of processing platforms is not only possible, but frequently very effective (Weill and Broadbent, 1998). The problem is not just that there is virtually no empirical support for the view that comprehensive IS integration is achievable (Segars and Grover, 1996; Goodhue et al., 1992; Allen and Boynton, 1991), but that consideration of the negative possibilities inherent in integration is not part of the theory. Yet evidence is available that integrated systems structures are relatively rigid and difficult to change in practice (Allen and Boynton, 1991) and that an organisation implementing such structures must lose some capacity for flexible response to change as a result. The issue of how to balance efficiency gains against losses of this type has been neither conceptualised nor researched.