Can’t say this more emphatically…

Start With an Important, Identifiable Problem

NSF CIO’s Recipe for Innovation Success