In the 1980s the IS function grew rapidly in many organizations, sometimes becoming the largest function in information-intensive firms. In the 1990s economic, technological, and organizational factors are combining to change the shape of the function and to often lead to 'downsizing'. IS activities, however, still continue and sometimes grow; specialists may no longer be the dominant group but users may have many more informationoriented responsibilities. So how to organize IS remains an important issue for IS executives, general managers, and line managers alike. Some valuable research of practical benefit on these matters is reported in this section.
The organization of IS can be seen in two dimensions: the centralization versus decentralization question and the issue of users versus specialists. The first is addressed by Earl et al., where the evidence suggests that effectiveness in IS is dependent upon the IS function being designed to fit the characteristics, especially structure, of the host organization. This may seem like common sense, but at least four significant further points are made. First, efficiency in IS is not so dependent on organizational fit. The operational, service, or utility activities of IS can be run efficiently anywhere, within or without the organization. Then combining the efficiency and effectiveness criteria, Earl et al. suggest five ideal types of IS configuration and how they should be selected. The drivers of change in organizing IS, however, are not only seen to be the characteristics of the host organization but the track record or heritage of IS performance. Finally, in most multidivisional organizations the federal form is seen to be the most viable. Not only does it provide a means of balancing centralization and decentralization, but it allows less costly structural changes to be made over time. And as Earl et al. show, change seems to be continual.
Feeny et al., using data from the same project which underpins the first chapter in this section, address the user--specialist dimensions. They find that integration of users and specialists is important whatever the structural configuration of IS. This is because there are differences between the two groups, each of which may have different roles, bringing different capabilities to the different activities of IS. Feeny et al. propose a contingency model of user--specialist integration, where the determining factor is relative maturity of the information technology in question. By 'relative' they mean the degree of experience of using and managing a particular technology in a particular organization. This hypothesis is closely related to the underpinning of both 'stages theories' and assimilation frameworks of managing IS. However, the Feeny et al. model seems to provide insights on a number of issues related to all information technologies throughout their life-cycle.