(Appealing, yet somewhat lonely, plain vanilla cupcakes from Euni's Flickr stream.)
What's the role of consensus in design and innovation?
Probably none. If everyone agrees, it might be time to go back to the drawing board.
Experience shows that one way you can tell if a design and innovation effort is working is that it's comfortably uncomfortable. The team should be exploring the edges and pushing towards the new.
If it isn't new, you aren't learning.
If it isn't new, it probably isn't a meaningfully differentiated
offering.
If it isn't new, you aren't going to get the attention of a
new user.
Since new offerings and new users are how you grow, then you're probably not growing.
When you're at the edges, people are bound to disagree on the right path forward. If everyone on your innovation or design team agrees, you probably aren't pushing hard enough. That's a tough reality and one of the hidden facets of what is usually a team sport.
So, what can you do to help inform your decision? Here are some options.
Observe and interpret what the ultimate user wants: Design research is meant, in part, to uncover explicit and implicit functional, emotional and reflective needs of a user. Getting out into the field to really look deeply and listen faithfully makes the difference.
Test, validate and repeat: Most large scale organizations know this well. Unfortunately, the thing about traditional validation and the use of benchmarks is that they are actually a form of consensus (albeit with folks or a standard that isn't even in the room).
Design for "yourself": There's a school of thought that says you can and should design for yourself. Steve Portigal has an article (the first in a series I believe) in Interactions magazine that discusses this point of view. I think this usually doesn't work for most large-scale companies (with notable exceptions of course) because most of their people are not the user.
Show the user: Build a prototype and show it to someone. Anyone. Projective methods, a type of design research, puts a question in front of someone to let them react. Let your users' reactions influence you (not necessarily guide you), refine what you're making and helping to craft how you'll tell the story of what you're making.
Build it, see what sticks and learn from it: The less the experiment costs, the better. Build a discovery-driven plan and you'll know what to learn.