Often, blue sky research has the best ROI. You get smart people working really hard on it. There's a lot less crap to deal with, so they are totally focused on

Often, blue sky research has the best ROI. You get smart people working really hard on it. There's a lot less crap to deal with, so they are totally focused on solving the problem. And the problems are actually quite clear cut (is the plant back from the dead yet?), so a lot of management issues disappear. A well defined (yet useless) goal is easier to manage than a wishy-washy "improve it, but don't change anything".