“Software Quality” has been in my title for over 25 years. In that time, I've been tasked with many Root Cause Analyses (RCAs).
When software managers fail to handle these properly, future risk increases.
One simple improvement? Never focus an RCA on why “QA” missed a bug.
Whenever a troubling bug in mission-critical software makes it to production, emotions naturally run hot. People turn to the Testing group for an explanation. How could the group tasked to “assure quality” miss such a bug?
Such a limited focus robs the company of real opportunity. Detecting bugs is only one tool in an org’s arsenal.
A small startup produced software consisting of five components. A bug deemed unacceptable was discovered just before a release. The developers quickly modified the code, QA verified the fix, and the company released the new build.
Later, support phones were off the hook. Customers reported fundamental functionality had been dropped.
What happened?
The RCA uncovered that after the final code change, the Test Group tested only that fix, then approved for release.
It also uncovered:
- The functionality drop wasn’t introduced by the developer; rather, a build compilation's manual step was done in haste resulting in entire components getting dropped.
- The bug would have been unacceptable to only one customer. The salesperson for that customer (and their commission) unduly influenced the Engineering decision.
The test group implemented a Release Candidate process. The company implemented formal release decision-making, thus reducing the influence any single salesperson can have.
When accepting the RCA challenge, accept the broader scope. Accept the challenge, clarifying you will search for root-causes, wherever those may be.
Uncover any management failure, and tactfully handle. A salesperson influencing the release process isn’t a salesperson or Test group failure; it is a management failure. While the tactful handling of this is outside the scope here, uncovering changes management must make are key to a good RCA.
If you want to effect actual change, expand your “RCA” focus beyond the Test group.