Managing the risks of information systems implementation

Managing the risks imposed by the complexity of system design and the uncertainty of requirements are key issues in implementing an information system (IS). It is commonly argued that complexity can be reduced by using rational approaches such as abstraction, decomposition and structuring of the design problem. Similarly, uncertainty can be reduced by using experimental approaches allowing for user participation, learning and feedback to the developers. An important but only recently studied issue in managing the risks of implementation is, then, how to find a proper mix of the rational and experimental approaches in a given situation. To do so, we first extracted some guidelines from the literature and formulated a model describing the economics of the allocation of experimental and rational efforts. The model illustrates the trade-offs and optimal conditions for managing the risks of implementation. The predictions of the model were tested by studying major information system development projects in Finnish companies. The results confirmed the contingent guidelines in general. Some interesting deviations were detected, however, such as a tendency to rely on rather heavy formal design even in simple but uncertain projects, probably owing to the long tradition of standardised life cycle planning and documentation methods in Finnish companies. Use of a proper mix of the rational and experimental approaches turned out to be especially important in projects with both high complexity and high uncertainty. Finally, practical implications and applications of the model are discussed.