And a common cause of getting stuck in local maxima (specifically, in committee/consensus-driven projects) is that each individual decomposed feature doesn't ob

And a common cause of getting stuck in local maxima (specifically, in committee/consensus-driven projects) is that each individual decomposed feature doesn't obviously deliver value on its own merits, but the sum of a group of such features can reach a higher local maxima. Aka, why argue about feature A, when the goal and long game is how A+B+C+D works?