One failure that I see very frequently is that the analysis reliably stops at the invisible line between the trenches and the management.
I.e. the people-problems are very well identified by everyone in the room, but nobody feels like calling out the guy who sits on the other end of the table in your next "performance review". It just doesn't seem like a good idea.
This is something that I'd like to see people like Eric talk about.
Because either I live in my own personal bubble here or that is a much more common problem than people not knowing how to trace back a technical issue to people and processes.
In fact, every engineer I know could sing you a song about it. But again, perhaps I really just happen to live in a particular bubble far away from Harvard...
Maybe the problem is that 5Ys worked great and seemed like a nice process improvement because we had healthy management. I think the reality is - the unhealthy managers who could really benefit from Eric telling them how to be better managers? They're never going to go anywhere near HN or the Harvard Business Review.
It's not a complicated process that requires certification. It's simple enough to be described completely in three minutes.
I'd also love to see an example of a problem you've encountered that is actually pure technology, since virtually all modern technology is created and maintained by humans. I've run into lots of crazy hardware and software problems, but pretty much all of them had a human involved here or there that could help make things better.