It's a nice theory, and I've seen places that did it well (1Password seems like a good example). Moving to a subscription does _not_ guarantee that the development process will improve. Scroll back through the comments here and you'll see that people had been complaining about weak upgrades back to TextExpander 4's release. I can't think of a single meaningful feature that they have added in the two years since switching to a subscription service. Two years of TextExpander-as-service disproves this theory. Aside: quoting the comment below 'Having read a lot of negative user feedback on Twitter about the subscription model, I am convinced, this step towards recurring payment makes perfect sense, because there is one big advantage for customers: constant improvement of the software.
