Reconstructing QA dwell time from GitHub history

Our team had one QA engineer and a persistent feeling that QA was slow. Feelings don’t make good arguments, so I wrote a script that walks the status-change history of every issue on the board and reconstructs how long each one actually sat in the QA queue.

What it showed

• Security fixes cleared QA first-time 89% of the time.
• Customer-reported bugs cleared first-time only 37%, and sat a median of 43 days.

The problem wasn’t a slow person. It was that sequential QA was surfacing bugs late, after developers had moved on. That reframing led to a workflow change: low-risk fixes now go to Customer Support on disposable preview sites and never reach the QA queue at all.