I haven't tried any of this myself since it raises a big issue for me: Doing this makes a promise that you don't intend to keep (at that moment). If trust is what gets users to buy, isn't it counter-productive to violate a user's act of trust?
I'd like to get over this because I can see how experiments like this can be useful and I hope you'll keep us all posted with your success rate. Some things I'm curious to know are:
* How much time passed between your AdWords test and your MVP launch? * Of the emails captured, how many participated in the beta and eventually became paying users?
Best of luck!