I am applying my tech purchase model, never a new product or release until its proven, to subscription software. If the developer shows a benefit to me, the user, to support the model over a reasonable stretch of time, the odds are good I will participate. But if the app development progresses at roughly the same pace as the point-purchase model, then there is little reason to subscribe. Implementing "X-as-a-Service" doesn't constitute a feature if it isn't optional.