The main idea is really about learning. Whatever is the minimal you can do while still learning something valuable is what you should do. If releasing your MVP

The main idea is really about learning. Whatever is the minimal you can do while still learning something valuable is what you should do. If releasing your MVP won't help validate your core business assumptions then it probably isn't worth the cost of releasing it. However do you have to build the MVA to at least somewhat validate it? If people don't get it... Are you sure you get your customer base? It's all just about rapidly learning about your market.