Metrics are often effective in giving a project manager a good idea of the progress of the design as well as managing resources. Simply tracking the number of bugs found and the proportion of the code covered can provide a guide to the maturity of the code, as well as where test resources are best deployed as the project progresses. But it is worth remembering that a falling bug count does not necessarily mean there are fewer bugs left to find – just that tests are no longer finding them. It is possible that a showstopper bug still lurks. But metrics alone cannot target that bug.
This is why we need to think again about what is being measured. The number of bugs removed is not the reason why a product will work or fail in the field. The question is whether the product hits its design targets. You need a way to test the design against the targets. That is where the requirements driven verification and test (RDVT) methodology provides the missing link between metrics and the project’s real aims.