In this post I will share yet another tool from my PM Toolbox: A simple method any Software Project Manager can use for predicting the likelihood of project completion, based on a model for defect prediction. This tool only scratches the surface of real software engineering defect management, but has nevertheless proved useful in my experience of managing a large software stack. For instance, this tool was borne from my direct experiences as a Development Owner for a large Software Stack for STB Middleware that comprised of more than eighty (80) software components, where each component was independently owned and subject to specific defect & software code quality requirements. I needed a tool to help me predict when the software was likely to be ready for launch based on the health of the open defects; as well as use it as evidence to motivate to senior management to implement recovery scenarios. It was quite useful (and relatively accurate within acceptable tolerances) as it offered perspective of a reality that more often than not, people underestimate the work required to get defects under control, the need for predicting project completion dates based of defect resolution rates, depending on the maturity of the team is often also misunderstood.
I created the first instance of this tool back in 2008/2009, the first public version was shared in this post on Effective Defect & Quality Management. Since then, I've upgraded the tool to be more generic, and also significantly cut down on the number of software components. I still use the tool in my ongoing projects, and have recently convinced senior management in my current project to pay heed to defect predictions.