I think this is one of those things you can't really understand until you're in the middle of it. One of the classic points of advice is to be embarrassed and release early, and release often.
The problem is, when it comes time to finally show what you've built to solve problem x to the world (best solution ever, in your mind), it's so tempting to delay. It's easy to find a laundry list of things that need to be fixed before your product is pushed to face reality.
As long as you don't release and keep building, you don't have to face reality. You can continue to believe you will have 100% conversion and 100% love. The longer you can stave off reality, the more confident you can make yourself feel. It's not until you push the release/invite button that reality strikes and you see your 6% conversion, with 2% providing feedback and 98% not saying a word or using the service.
We recenlty found plenty of reasons not to push HighScore House out to more parents, who were waiting to try it. We eventually realized the reasons sucked, and we did push it. We then saw how few actually converted to users. From the release though, we did find key early adopters who will help drive the product forward and we now know the work required in getting to our first 100 users.
Reality sucks, but it isn't until you face it that you can measure, learn and build. If it's a truly useful idea, your product will survive the "reality test," If the idea sucks, well, better to face reality sooner rather than later.