One of the more important principles of nimble corporations is that their leaders do not expect to get things right the first time around. No matter how great something looks, in its first iteration, you are only guessing whether the target group will find it equally great. In fact, it will undoubtedly have weaknesses. Things that could be improved upon.
For this reason, Ries (author of The Lean Startup) is right. The goal is to get a prototype out there that can be improved upon as quickly as possible. Prototyping is done in a limited market test. Only after you get the bugs ironed out do you go to scale.
You might think of this as giving yourself the opportunity to make intelligent mistakes.