At IMVU, when Ries and his team finally launched their first product, they were embarrassed by how basic it was. The avatars were poor quality. The interface was rough. Many features were missing. They expected customers to be put off. They were not. Customers gave feedback. They paid. They came back. The imperfections that the team had agonised over were largely irrelevant to the customers who found the product useful. This is the core insight behind the MVP: you almost always know less than you think about what customers actually care about. Building a polished, complete product before testing that assumption is enormously expensive. Launching something basic and learning from real users is cheap. Ries is clear: an MVP is not a bad product. It is a carefully designed learning tool. Every feature that is included should be there because it tests something you need to know. Every feature that is excluded should be excluded because building it would not add to your learning. The question is not 'Is this product good enough?' It is 'Does this product teach me what I most need to know right now?'