We really benefited from allowing ourselves ample time to test the app. Firstly, it helps you perfect the app but we found more importantly, it actually helped with the marketing.

We used TestFlight but HockeyApp is also great. (neither solution is a joy to use, mainly due to limitations Apple places on distribution of test builds). Getting 10-50 users to test is usually sufficient unless you're testing a multi-player game that requires more players to test the in-game dynamics. In those cases, it's often better to go straight to a soft-launch in a smaller English-speaking country (see http://blog.tapstream.com/post/71538606229/the-art-of-soft-launching )

I mentioned how your beta test can actually help your marketing efforts: use the beta to create buzz and aura of exclusivity around your app. Get people with large social followings to test it, enjoy it and talk about it before it's released. Have a LaunchRock or similar lander to capitalize on that buzz and to collect emails for a waiting list to use for launch. This can give you a serious ranking boost on the launch day!