We see, know and experience information systems development failures in many domains and in many countries. This paper will explore some of the issues related to the study of these failures. Every year, billions of dollars are wasted on failed projects. The paper will emphasise the fact that the study of failures can only take place post-hoc, once a failure has been identified. Preparation is therefore different to normal scientific study where a situation is pre-selected in advance, the precise parameters are identified and decisions are made about the best methods for measuring them accurately and objectively. The literature reveals that researchers and practitioners have been experiencing projects failures for many years. Indeed, acknowledgements of failures go back at least thirty-six years. However, failures are still a prevalent problem. A significant obstacle related to the study of failures is the lack of acknowledged research methods for understanding such complex phenomena. The evidence collected during failure investigations emerges from a variety of sources, perspectives and contexts. Not surprisingly, it often appears to be ambiguous, incoherent and confused. The information collected tends to be rich, messy, contradictory and subjective. Such situations call for a new repertoire of methods to address the unique features of failures. This paper will introduce possible alternative ways of looking at and constructing failure stories. The techniques described below come under the umbrella term forensic analysis. The insights obtained from forensic analysis can be used for internal learning within organisations as well as externally within the discipline, thereby enabling practitioners worldwide to benefit from the mistakes of others