The figure below offers a good starting point to determine what the most important metrics from a testing standpoint could be. The cost of detecting and fixing a defect escalates as the defect is detected later in the project. The obvious implication is that testing should aim to find defects earlier in the lifecycle. Another is that defects should be predictable- this can often be achieved based on historical data and a few heuristics.
The x axis in the graph below shows the phase of the project, and the y axis shows the relative cost of finding a defect in the software.
The next post will identify some of the metrics that can be derived from the above diagram. Meanwhile, can you take a shot at identifying what kind of testing metrics can be useful in reducing the cost of defects?