It can be argued that opinionated design is more successful that design by committee. I think the most effective approach is to begin with opinionated design prototypes then triage and refine using A/B testing.

MS can be accused of relying too heavily on user surveys, making their designs a bland blending of averages that please nobody (see: the Office Ribbon). Thus the Ford quote gets trotted out, "If I had asked people what they wanted, they would have said faster horses."

The point is that MS are using a greedy optimisation algorithm with which they arrive at a local maximum that is well below the global maximum for quality of end product.