Software project management towards failure avoidance

Development tasks, in the thousands or more, are involved in a complex system/software project. These tasks aim at three objectives in the project, namely on budget, on time and on performance. Behind these tasks are decisions made on them to move the project forward. Project failure is an aggregated and cumulative failure of these tasks due to errors in tasks and/or decisions made. The project errors from faulty strategies to wrong builds, collectively called exceptions are much harder to detect than cost overrun (amounts spent) or time delay. The paper suggests a biologically-inspired approach to project failure avoidance which is different than most. It focuses on exposing exceptions as they occur and understanding the decisions made on them. Thus, there are two pieces needed for the proposed approach. The base piece is a management by exception (MBE) framework to monitor development exceptions occurred for management attention. The second piece is an adapted measurement method which will elicit, analyse and evaluate the decisions made on the reported exceptions. We argue that using the proposed approach, failure avoidance is possible and software project performance (in scope, intended features and desired quality) would be under control, and so are expected cost and time.