I think the way its done in biotechnology is that the company pays and some 3rd party lab (e.g. university) then does the experiment.

There is also a positive bias from industry studies due to the fact that products tested internally that don't work will not be pursued and therefore that research is not published.