Predicting Incident Reports for IT Help Desk

Costs associated with IT help desk operations present challenges to profitability goals in an organization. Minimizing software failures in an operational environment is important to customer experience, but more practically to the cost model for offering IT services. The user-facing behavior of software systems such as unscheduled downtime, slow performance and anomalous behavior affect the overall perception of a software product. Critical business applications and systems with human life at stake demand reliability and continuous availability of services. Software failures are costly. In this case study we investigate software reliability models and their applicability to improvement processes at an IT help desk. We propose a model selection process and demonstrate its success using real help desk incident data from eighteen desktop software applications. Our results show direct applicability to meeting cost challenges in IT help desk operations.