Causal analysis of system failures from IT project trouble reports

One of the leading causes of IT system failures is the failure of IT system development or operational projects. These developmental or operational failures may result from software faults or erroneous IT system setups or operation. To prevent such failures and their recurrence, it is essential to clarify the reasons for past failures. This paper proposes a method to clarify the causes of IT system development or operational project failures that lead to system failures. The proposed method extracts the causal relationship between a project failure item and a system failure using a statistical operation. Furthermore, structural information on the causes of failure is generated using interpretive structural modeling. Eleven failure items were structurized in an experiment that analyzed 40 trouble reports from IT projects. Consequently, a general pattern relating project failures to system failures was extracted.