Description We can run this company virtually, why bother with staff? If discovery is discontinuous and we will depend on a small cadre of individuals for our creative thinking, then let's outsource all the other work and avoid the fixed expense of employees who will  be more or less busy depending on the flow of the creative process?

The output of the creative process will be discontinuous, but their will always be more than enough work. Indeed it is expected that key insights will come from our Big Mo'+ activities.

There are the core half-dozen players in any R&D pursuit, but an additional one to two dozen players need to be intimately involved in any R&D effort. Daily intimate involvement is needed to be able to productively contribute in more than a mechanical fashion+. The important information is implicit and not written down. You pick it up by being part of the activity.

So we can and should use full-time research contractors but only if we can engage them exclusively over extended periods (years) or we leverage them for quick hit insights to unstick internal thinking. And even in those cases, they won't sense the same behavioral fundamentals as they would as employees. External service providers are mostly volume operations: they demand and contract for standardized, explicit work tasks. For example, Contract Research Organizations (CROs) demand a firm, fixed clinical protocol. This is not what we seek for innovative product development.

See (listen to) Munger on the Nature of the Firm to see what happens when a firm outsources too many activities (jump to about 28:00 in the podcast or transcript for the specifics).

Appropriate Uses  The more mechanical and proscribed the R&D activity the more applicable the use of outside contract resources.
Further Reading