11 ms·
We should stop using the Five Whys for root cause analysis, for this and especially for post mortems. The reasons are obvious. A “root cause” will almost never
by stevebmark 5y ago
We should stop using the Five Whys for root cause analysis, for this and especially for post mortems. The reasons are obvious. A “root cause” will almost never be exactly 5 whys down, there will almost never be only one root cause, and the reasons are a cyclic graph, not a line. The five whys are also random. They change based on on who does them and which why you randomly start on. If it feels like an awkward and difficult root cause analysis tool, that’s because it is.
- hinkley 5y agoThat’s why it comes down to who is steering the conversation. And someone is even if it looks like nobody is. A lot of circular reasons come down to human factors, like juggling three things at once is distracting. Adding on an extra thing doesn’t fix that. Making one or two of the three less distracting doesn’t fix it, but it makes it better, it makes progress, and it establishes precedent for policy changes. It’s an agenda you can hammer on every time something breaks until the problem only happens once in a great while. It’s the fact that 5 Why’s develop their own epicycles that sticks with management. Of the last five three of them had the same kind of conclusion, I guess we’re spending time on that now.