Imho, being a marketing guy, this whole soup comes kinda natural to me, as it seems obvious that the market (the early adopters in our case) has to be the one driving the product through a dynamical and incremental trial&error path.
But from my experience I can see that all of this typically sounds convoluted and pointless in the head of an engineer: engineers are used to analyze things based on available data and find an optimal solution, marketing guys are used to guess the wannabe-best solution based on the “feeling they have for the market” and the users’ feedback.

Currently we’re in this kinda grey 1/2 stealth mode: we didn’t go out to the press yet, but we promoted our software through GOTD and BDJ which gave us the little exposure we needed to build a little community of fans we could confront ourself with. This strategy so far is paying off pretty well: we’re still very poor but people starts finding our product more and more useful and user-friendly and this interaction led us to rethink many of the choices we did in the earliest days.

The way we handle this is mainly this thread on our forum where I write what is going on, and post links of it on twitter/blog/facebook:
http://www.tabbles.net/forum/viewtopic.php?f=6&t=69

Plus the “beta” section of the forum, where each beta version gets discussed.

I hope this can help someone else out there! :-)
(yeah, maybe I should post this stuff on my blog too)