The importance of test metrics
Para acceder a post original click aquí
If you want to get a clear picture of software quality, or find confirmation that a new tool has widened your test coverage, then you really need to put metrics in place. Only by measuring the right things, can you can confirm that the software meets requirements, identify key areas for testing focus, and find the best processes and tools.
There are various different metrics that you might employ to gain some useful insight about your current project and your team.
Requirements and defects
One of the most obvious things to look at is requirements coverage. Does the software meet the requirements that were laid out at the start of the project? If not, why not? Consider how many test cases have been designed for each requirement, track what still has to be designed, what has been executed, what the results were, and how many defects were found.
You’re going to want to track defects closely. Don’t just look at overall defect numbers, look at where the most defects are concentrated and factor in the severity. If specific areas of the software consistently have a much higher severe defect rate, then perhaps there’s a deeper problem there. Encourage testers to log usability issues as well, and cross-reference these to identify areas of the code which need work.
By tracking all of this data you can work out the percentage of work completed and what has yet to be done. That enables you to understand the remaining time and resources required to complete the project and identify potential delays early. You can also send up early warning flags about potentially problematic parts of the design.
Tracking your team
You can drill down into the numbers and calculate defect discovery rates for individual testers, but be careful to assess them fairly. Defect rate in isolation is not a fair measure of effort, some bits of code have fewer defects, and encouraging bug count above all else can encourage testers to split bugs or enter duplicates to boost their numbers. Tracking these defect counts across multiple projects and comparing teams can reveal insights.
The rate at which defects are fixed is obviously important. Keep a close eye on how many are deemed as not defects by the developers or merged with existing defects as duplicates. You can also look at the efficiency of your testing by comparing the number of defects identified and dealt with during testing and the number found later by the end user.
Always compare your estimates at the beginning of each project with your actual performance. The two should get closer with every project that passes.
Assess tools and processes
You can apply the same analysis to the tools and processes that you employ to ensure they’re delivering value for money. Did that new management software really save you time? Are defects being fixed faster, now that you’re recording each defect? Does your new workflow get better results than the old one? Measure everything and you can filter the data in different ways to extract useful insights that will help you improve the way you test.
It’s good to keep questioning and looking for ways to improve, but only test metrics can give you the evidence you need to argue for changes.