Most firms manage most of their new product projects using techniques we have been presenting: PICs, concept testing, screening models, protocols, and so on Other projects are wildcatting betting on a technology that hasn't yet been shown to work, betting on a new application where some end-user will partner with us to see what works, or just betting on a scientist with a good track record for coming up with sale able new products. None of these is appropriate for a protocol we just don't have the knowledge to write one, and its only effect would be to bother the developers, who, actually, will ignore it completely.