I've had times where I stopped the "5 Why's" process short of getting into the human-squishy organizational root causes because I knew the problems were too big to be solved in the context of that particular COE/postmortem. I've also had times where I knew my director/senior manager would've taken issue with the publication of those squishy organizational issues, /even if/ they were true, simply because he/she surmised it wasn't worth burning the political capital for it. (And I've been asked to revise my 5 Whys by leadership exactly because of that reason before.)
I've successfully pushed through organizational changes at least a few times from the output of the 5 Whys process too. If the COE/postmortem revolves around a critical issue or has gravitas, it's easier to put your full weight behind it.
To be honest, my personal hit rate on these is probably about right. If you insist on it every time, you're probably going to burn out on it, personally or organizationally. Sometimes it's OK to try to address a symptom, if the medicine is cheap, easy to administer, and has little to no side effects. And in tech, the symptoms are often inherently layered -- you can often address entire classes of symptoms in reasonable ways, /even if/ that solution is itself working around a deeper organizational shortcoming / root cause.
The level of "5 why's" depends on the level of the person doing it, going as far as its needed to reach that level where you should be driving change.
In the "car won't start" example of the original article, the motor pool mechanic needs to stop the whys at the broken alternator belt, because that is something they can and should fix; the motor pool manager needs to go on until reaching the practice/policy for how often preventive maintenance is done, and the top management (if the consequences are severe enough to warrant attention) need to go on until the "why do we have poor car maintenance" questions reach things like mid-management forcing people to ignore the maintenance schedules, or staffing policy, budget for whole departments, and/or the choice to outsource/insource that function.
But going on much further than your own level is not going to be productive - you can make an actionable suggestion to your boss if some root cause isn't fixable at your level but is fixable there, and that's about it, at least in my experience.
I've successfully pushed through organizational changes at least a few times from the output of the 5 Whys process too. If the COE/postmortem revolves around a critical issue or has gravitas, it's easier to put your full weight behind it.
To be honest, my personal hit rate on these is probably about right. If you insist on it every time, you're probably going to burn out on it, personally or organizationally. Sometimes it's OK to try to address a symptom, if the medicine is cheap, easy to administer, and has little to no side effects. And in tech, the symptoms are often inherently layered -- you can often address entire classes of symptoms in reasonable ways, /even if/ that solution is itself working around a deeper organizational shortcoming / root cause.
Spend the friction tokens where it matters most.