I think the right balance is to at least have some kind of prototype to show off. The initial phase of "not having to write code" is pretty short. You are more likely to figure out who are the customers and who are not. You will also exhaust all options to get email / signup quickly... If you spend way more time into this phase and still don't want to write code, you will start losing people's interest. So when I said the "right balance" is to code in 2-weeks timeframe. That's what my buddy and I did with this TV social site with show casts/celebrities http://www.tivify.com/. It is something raw, brutal but works!! We at least could take it to Meetup and show off to people. That's easier to sell than slidedeck or launchpage.
