I usually build an MVP all at once. I'll have a plan on Friday and start working first thing Saturday morning. Typically I work until the MVP is finished; 8 to 12 hours later. I believe an MVP should be as simple as possible and I'm no prefectionist. Some really poor quality software and websites are popular. If it feels a real need, I assume people will use it.

I haven't built anything terribly successful, however, because I often build the wrong idea or lack the marketing chops to get people interested. My next MVP will be a simple marketing page with a "coming soon" sign up box to see if people even care before I build it.