Managing divergences in IT infrastructure standardization

Abstract

Standardization of an infrastructure is seldom a straightforward process, local adaptations of the global standard are often necessary in order to make the standard viable. These adaptations can be seen as divergences from the standard and need to be managed in order to avoid disorder within the IT infrastructure. The existing literature does not go in to detail regarding divergences; it focuses mainly on the phenomenon of drift. The purpose of this thesis is therefore to increase the understanding of divergences and how they can be managed. Our research questions are; “how can the divergences from a standard be described?” and “how can divergences from a standard be managed?”. The findings of this thesis are of interest for both academia and practice in order to better understand divergences. Two opposing views of infrastructures and standardization are presented; one emphasizes the need for control and the other focuses on the phenomena of drift. By using both these views together with the results from interviews held with managers at Volvo Group we define the dimensions of divergences. We describe five dimensions that are necessary to understand; type, underlying causes, time of creation, context and effects. We also present a process model, containing four stages, for managing divergences in a structured way. The first stage concerns the detection of either existing or anticipated divergences. In the second stage the aim is to understand the divergence based on the five dimensions presented above. In the third stage a decision to permit, adapt or remove the divergence is made. The last stage focuses on communicating the decision made in the third stage across the organization.