Pattern-based solution risk model for strategic IT outsourcing

Abstract

We present a pattern-based solution risk model for assessing risk of incurring a cost-overrun in Strategic IT Outsourcing (SO) by the SO provider based on historical deals and their corresponding cost overruns. The approach is based on finding co-occurring patterns of solution elements and cost-overrun elements, i.e., elements that had to be implemented as not foreseen in the project planning phase. In order to find such co-occurring patterns we apply closed itemset-mining augmented with item cost information and build corresponding association rules with risk information. Such rules can be used by project managers of SO contracts to minimize the gap between the proposed and implemented solutions. In experiments, conducted on a sample of deals of a multi-national SO provider, we show the applicability of the framework for predicting significant cost-overruns. The introduced model is a general solution risk model for service delivery, whose task is to minimize the gap between proposed and implemented service elements by the provider based on historical deals. © 2013 Springer-Verlag.