I worked as a researcher at NIWA for a few years. Commercialisation of ideas was a mantra. All research was seen through the prism of commercialisation. The sort of people employed weren't particularly good scientists (as alluded to in the article) , but they had tended to have CV's with commercial science backgrounds. We ended up with a lot of chiefs and very few researchers capable of formulating experiments, or having innovative ideas.

Peer reviewed publication was viewed as quaint or an inconvenience.
Many experiments were done without the basic scientific requirements of replication or control because the data weren't intended for publication. The results were often inconclusive due to poor experimental design. There is a reason for peer review - it is a key element of the self-correcting process of science.

I am happy to hear people such as Bart Janssen and Peter Gluckman saying that the current CRI model isn't particularly productive, because that is my experience. I can't see more funding to CRI's increasing productivity, just more project managers.