What if we stopped looking at each instance of a previously-engaged Hubski member inactivating off the face of the spokewheel-web as a stand-alone incident unrelated to any others - what if instead we started to look at departed users, the dataset, for commonalities, trends, and interesting conclusions - because right now someone leaves and we talk about it, if we notice, and go, "oh how sad, wonder why that happened," and gossip a little I guess, and move on. Then someone leaves again next month or the month after. Wash, rinse, and repeat. We really going with the assumption that all these previously-engaged users are disengaging, visibly and repeatedly, and none of them have any major factors in common? Or is that other people's problems because they're singular points which stand alone, and Hubski's already perfect? Hubski's already perfect never seemed like mk's approach to me.