Hm. I realise this is an advertisement but that just makes its deficient treatment of failure all the more worrying.

I have yet to write the article on systems-theoretic accident analysis that is long overdue, but I don't subscribe to the idea that there is "a" root cause, much less "the" root cause. What people call a root cause is almost always just an arbitrary stopping point for analysis. This stopping point is selected either based on convenience or for political reasons.

Any real-world accident comes after running the system in a hazardous state, often for quite some time. The decision to do this is rarely explicit or the result of one thing, but a multitude of interacting factors and competing interests. A safe system needs feedback loops and control actions to keep it out of hazardous states. For any accident, we will usually find many missing or deficient control systems and feedback paths.

This is usually easier to phrase in reverse: we would never ask someone to identify the root cause of success -- and it's just as silly to do it in the case of failure.