I'm of the opinion that you shouldn't make guesses when you can actually test the hypothesis. Us design types do that through primary research and observation, and one of the main tools is prototyping. It's important to actually try different things instead of just guessing at what is best, following someone else's research, or basing judgment on past experience. All of those things are valid parts of design and business, but they are all still forms of guessing.
The concept itself will constantly evolve and change in response to things like new technology, new market opportunities, or economic conditions. The difference is subtle, but traditional businesses do things like sticking to their "core competency." We don't live in that world anymore. I'm going to test a bunch of things, take the stuff that works and move forward and ditch the stuff that doesn't work...frequently.