I don't see evidence in that article that the new pricing is better than the old. Did you test how well it performs before announcing that it is "fixed" now? Don't get me wrong, the reasoning for these changes makes some sense, but without validation that it really converts better now, it simply cannot be called "fixed".
The motivation for the change came from chatting with people looking at the site, and hearing strong reluctance from some about having to pay to evaluate the service.
The pattern we hear is that often people want to get it running in a sample app or side project so they can get a feel for it - even if their motivation is to eventually use it with their larger site.
The more precise quantitative results of the change will come now that the pricing change is public. We should do a followup in a week.
