I’m still sticking with my original narrow set of blames pointed out by @andreas7117 , viz.
- unclear/unknown requirements
- unknown code changes
- no understanding of the business
- no understanding of the technical surroundings/environment
I’m loathe to drop the whole dictionary on this one, because listing causes of a P1 being undetected, and I think this is what we are talking about here: P1=customer reported defect that has no trivial workaround. We are all relying on automation to catch most bugs, but for example MacOS “dark mode”, is not automated tested in your system, is it? Or…Much like 3rd party interactions like when a overseas cellular network operator takes ipv4-ipv6 shortcuts that break your app, it’s fatal, but something you could never have tested for without in depth knowledge!
Often, I want to stop messing about with long lists, mindmaps and threat diagrams. Check this clip by a fellow called Simon Wardley, about building real maps to map a problem space. SEACON:UK 2019 An Introduction to Wardley Maps - YouTube