That is hard as hell, to be honest. Paradoxically, I think it requires a top-down commitment. Companies usually want to try this sort of thing incrementally. So they put their best people into an R&D group that has a fair amount of autonomy. However, over time what seems to happen is that it's resented and attacked, and eventually has to spend more time fighting political battles than innovating. At that point, it's often just as underperforming as the rest of the organization, the experiment is judged a failure, and that autonomy is taken away. It's the "fish frying" problem (http://www.strikemag.org/bullshit-jobs/).

I'm also skeptical in relation to how well an open allocation culture would work out once a company grows beyond a certain inflection point. I'm not well read enough on this matter to put a stake in the ground and define that exact inflection point.

Gore-tex uses open allocation and is large (almost 10,000 employees, if I recall correctly). What I don't think OA can survive is too-rapid growth. I think it can scale to thousands of employees. I don't think it can tolerate personnel growth at the 100-150%/year levels that VCs want to see.