I've proposed new product ideas on both data and intuition. Sometimes it worked, sometimes I had to right it off to experience.

This is relevant to both new product development and just new features. Sometimes a new product is a good idea to a couple of your beta users but you may not get enough traction off the larger user base.