Root cause analysis

Tom Breur

2 August 2015

Why? If you’ve ever been haunted by kids living through their “why phase”, you know they can drive you crazy. “Why” seems such an innocent question. Yet it can generate so many different answers. And because you can answer on so many different levels, it isn’t always obvious what answer will satisfy, which answer will “work.”

Another profound aspect of why questions, is that the answer you give is only useful, or even relevant, for people in the right context. Why did the software build break? Maybe because we didn’t test it properly. Maybe because our production environment isn’t stable. Maybe because some method returned an out of bounds parameter, etc.

There are infinite levels you can explore, and unless you appreciate the context of the person asking this question, it may take repeated “whys” until you get to a satisfactory answer. It’s not until you hit a level that the listener can relate to, that your answer will “make sense” – be of any use.

In a similar vein, root cause analysis only arrives at the root –a useful, sensible root– after successful problem understanding (emphasis added by me)(Andersen & Fagerhaug, 2006). Unless you know who you’re solving a problem for, and why this is a “problem” to them (Gause & Weinberg, 1990) you might be spinning around in circles, as if “haunted” by a super annoying kid!

Isn’t it intriguing how it seems we always need to establish a “connection”, understand the world from the other person’s vantage point, before we can interact in a meaningful way at all?