Own the question, not the solution
This morning I found this talk given by Dan Milstein at the Lean Startup Conference in 2014. I was surprised to see that it had less than 1,000 views because it is exceptionally good. Take 20 minutes while you're warming up your brain today and give it a listen.
Perhaps one reason why Product Management came out of Product Marketing is because being successful at it requires a keen understanding of people. I've come to realize that self-awareness and understanding of human psychology is one of the most important skills needed to create something new that is valued by a wide variety of people.
Dan does a fantastic job of explaining why it is so hard to create something new. You're going to be wrong, many times, and if you can't get really comfortable with that fact you should probably not be trying to make something new. The "vision guy" loves to talk about his vision, which is great, but if he's not also able to hear why parts of his vision won't work (now) then it's going to be extremely difficult to ship something that matters. And by "matters" I mean that customers will use your new product and maybe even part with their hard-earned cash for it.
It's human to want to be correct. We think we look bad when we change our minds—especially if we're leading a team. But, the opposite is true for great teams: They'll support a leader that is willing to own the fact that his plan was incorrect. Great teams will support a leader who explains why the plan was incorrect and what they're going to do about it. Great teams will not support a leader who always tells them everything is great, even though revenue is flat and customers are vocally unhappy.
How do we fight this human nature to always feel that we need to be correct? Great product leaders own the question, not the solution. Whether you are a product manager, founder, or "visionary" spend the time necessary to deeply understand the customer problem that you're solving. The solution is going to change, many times, but the question that you're trying to answer should not.