The contemporary provision of software-centric information systems follows two main operating models: outsourcing and insourcing. Outsourcing involves the contracting of services to external vendors whereas insourcing involves contracting to wholly owned internal subsidiaries. The outsourcing lifecycle can be further divided into three main phases: decision, transition and operation. The decision phase mainly involves analysing internal requirements and looking for possible solutions, while the transition phase is an interim period in which implementation of a chosen solution begins or a handover to a new vendor is carried out. In the operation phase, previous arrangements cease to exist and the new arrangements take over. Disentangling these three phases and analysing them in isolation may not provide a complete picture of the outsourcing lifecycle in general and transition in specific. Therefore, in this thesis a holistic definition of transition is used which encapsulates the complete outsourcing lifecycle that covers all three phases of decision, transition and operation. It has been asserted that more than 80% of IT outsourcing contracts are renegotiated during their lifespan. As such the need for consideration of the client-vendor relationship in outsourcing is continuous – it does not end when the contract is signed or even in operation when the services are outsourced. What happens, for instance, when a contract expires or when circumstances in an organization change, resulting in an outsourcing agreement being reviewed? The available options are to continue with an existing vendor using the same contract, renegotiate the contract with the same vendor, re-tender the contract and switch to a new vendor, or backsource the outsourced activities in-house. Research about decisions to backsource or switch vendors is scarce; however, decisions regarding backsourcing or vendor switching are becoming increasingly common in practice as firms seek reductions in their IT costs and improvements in quality. There is therefore a gap between theory and practice concerning the transition phase of outsourcing. The research reported in this thesis contributes towards a deeper understanding of this complex phase.

In this research an interpretive research philosophy underpins the conduct of a case study with a longitudinal perspective. In order to develop knowledge about the conduct of transition a process-oriented approach is used to structure the case study analysis. Narratives and dilemmas are used in the data analysis and the presentation of the results, with the narratives then conceptualised as Antecedent-Consequence diagrams. The case study at the centre of this research follows the transition process in the Novopay project, a Public-Private Partnership in the New Zealand education sector. Methodologically this research demonstrates the use of publicly-available documents and sources to investigate a controversial topic, while leveraging multiple data analysis techniques provides rich insights. While the context for the work is the New Zealand public sector, readings from media reports suggest its wider applicability. Government agencies worldwide are frequently engaged in transitions for major software systems and e-government projects. Thus while this research benefits the New Zealand public sector by generating knowledge which is specific to the local context it can also be disseminated to other organizations involved in large-scale transitions – or that are planning to do so. Some of the high-level recommendations made based on this research are as follows: network-based relationships cannot be established instantaneously, rather, they are developed progressively; in a vendor lock-in project, sharing the bidding costs with interested vendors can increase their interest; data migration and responsibilities for its transformation should be clearly defined; in order to increase the acceptability of a new system, multiple training methods should be used based upon end-users’ capabilities; successful transition should be linked to Key Performance Indicators which indicate the number of end-users successfully using new business processes; when choosing ‘Custom off the Shelf’ based systems, complex business processes should be simplified; and extensions for ‘Business As Usual’ support should be negotiated in advance