not necessary. if you launch early a crappy feature with a ton of problems it is actually very harmful. time to market is just one dimension of a trade offs you need to balance if you want to succeed
I often tell my management, "If it doesn't need to work, I can get it to you any time you want." Everything other than that is a cost/benefit analysis. To be honest, it's hard enough writing a reasonable sized program and not ending up with a horrid mess when I'm doing my best. I'll always recommend giving up functionality over quality if early release is desired. It's easy to add functionality later. It's hard to add quality.
