Lean startup model, which has overtaken the startup world, doesn't make sense all the time. Launching early can work if you have no network or platform. ie. no marketing impact at launch. It works terribly if you already have a substantial audience, because that first impression is important. You get one chance to show those in your network who you are. Perhaps a middle ground: many alphas and betas w private groups of people who are invited to test the product; once they're excited and coming back, asking if they can invite their friends, then you launch.