The role of data is, surely, to answer "what would have to be true?"

Ever since you wrote Playing To Win, I have hauled hapless strategists into a room and asked them to write a list of "what would have to be true?" questions, and my role (as consumer researcher) is to answer those questions.

If it would have to be true for the target consumer to own an iphone, and instead they own an Android, it is not true, and therefore invalidated.

None of that guarantees that you'll land on your winning idea, but it sure does weed out the crap ones.