Who knows best?

Reading Mike Cohn's book on Agile Estimating and Planning all over, I can't help reflecting on how easy to simply do the wrong stuff. One of our customers in facility management said that of what an average technician did for the customers, 30% was something he shouldn't have done (mostly: while I'm already at it, I can as well do this too). The consequence: well, 30% of the stuff that needed to be done isn't. And the problem is that even if the extra is nice, the lack off the stuff not having been done is much more noticed.

I also believe this is a huge risk in software development. Mike Cohn describes it so well. We have a task which should take 4 hours. Three hours fun coding and one hour of boring documentation. The three hours of coding is up and the developer gets an idea of something "very very important". So he does that instead of the documentation for the final hour.

So, what is the problem: well he makes himself the product owner. He thinks that he knows best: that the extra coding is more important. Well, he doesn't think he thinks that, but that is really what the consequence is.