Business-driven IT Management

Business-driven IT management (BDIM) aims at ensuring successful alignment of business and IT through thorough understanding of the impact of IT on business results, and vice versa. In this dissertation, we review the state of the art of BDIM research and we position our intended contribution within the BDIM research space along the dimensions of decision support (as opposed of automation) and its application to IT service management processes. Within these research dimensions, we advance the state of the art by 1) contributing a decision theoretical framework for BDIM and 2) presenting two novel BDIM solutions in the IT service management space. First we present a simpler BDIM solution for prioritizing incidents, which can be used as a template for creating BDIM solutions in other IT service management processes. Then, we present a more comprehensive solution for optimizing the business-related performance of an IT support organization in dealing with incidents. Our decision theoretical framework and models for BDIM bring the concepts of business impact and risk to the fore, and are able to cope with both monetizable and intangible aspects of business impact. We start from a constructive and quantitative re-definition of some terms that are widely used in IT service management but for which was never given a rigorous decision: business impact, cost, benefit, risk and urgency. On top of that, we build a coherent methodology for linking IT-level metrics with business level metrics and make progress toward solving the business-IT alignment problem. Our methodology uses a constructive and quantitative definition of alignment with business objectives, taken as the likelihood – to the best of one’s knowledge – that such objectives will be met. That is used as the basis for building an engine for business impact calculation that is in fact an alignment computation engine. We show a sample BDIM solution for incident prioritization that is built using the decision theoretical framework, the methodology and the tools developed. We show how the sample BDIM solution could be used as a blueprint to build BDIM solutions for decision support in other IT service management processes, such as change management for example. However, the full power of BDIM can be best understood by studying the second fully fledged BDIM application that we present in this thesis. While incident management is used as a scenario for this second application as well, the main contribution that it brings about is really to provide a solution for business-driven organizational redesign to optimize the performance of an IT support organization. The solution is quite rich, and features components that orchestrate together advanced techniques in visualization, simulation, data mining and operations research. We show that the techniques we use - in particular the simulation of an IT organization enacting the incident management process – bring considerable benefits both when the performance is measured in terms of traditional IT metrics (mean time to resolution of incidents), and even more so when business impact metrics are brought into the picture, thereby providing a justification for investing time and effort in creating BDIM solutions. In terms of impact, the work presented in this thesis produced about twenty conference and journal publications, and resulted so far in three patent applications. Moreover this work has greatly influenced the design and implementation of Business Impact Optimization module of HP DecisionCenter™: a leading commercial software product for IT optimization, whose core has been re-designed to work as described here.