You need a high level of traffic before you can really start drawing any statistically meaningful conclusions from A/B testing. People don't magically start showing up just because you publish a landing page, so some assumptions are necessary for the first version.

Sites like http://fivesecondtest.com/ or http://www.usertesting.com/ are a lot more useful than A/B testing when you're about to launch your first landing.