If the revenue wasn't covering the development costs (plus profit) then they definitely needed to do something, I agree. I just utterly disagree that the best (or even a vaguely good) thing to do was rebuild the current product with less utility.
I can't really argue with any certainty one way or the other about this since I don't know all the inside info. From the outside though, I can see a HUGE amount of potential that I didn't see available (or as easily available) with the offline builder(s.)
There's a reason more and more software is moving to the cloud after all.
Suppose they had kept the offline CB and Adventure Tools, introduced one-time charges for having access to them (which got you a unique registration code) and tied updates and additional services (like upload access to the VTT, character and monster design sharing - both upload and download - and similar services) to being logged in with an account linked to such a unique code.
Might work, might not- we'd have to look at the costs involved with maintaining such a system.
Off hand it still seems susceptible to the unintended revenue leak I mentioned before.
Would it help one of their other goals which was to combat piracy (And by this I mean at least trying not to blatantly help piracy)
Also- what about Mac support?
What about future products? How easy/hard would it be to tie the old CB in with future plans?
(I get the feeling they kind of eventually want the online stuff to function as almost an xbox online continuous convention sort of thing... Log on and jump into a game at any time.)
The CB itself doesn't need the web stuff to work, but to take full advantage of the webstuff with the CB you need to have it registered to your account. I would still be subscribed, the 'occasional payer' would still be paying occasionally (plus a one-off for the program initially) and I think a lot of the flack that has hit over the whole shebang would never have happened.
I doubt it...

Anytime a change is made there will be flack.
I think the main amount of Flak seemed to come from loss of current functionality. I don't think that had anything to do with going to a web based system- I think it had to do with a release date that didn't match what was actually needed to complete the system (which could have happened with anything they did.)
And the development costs for the switch would almost certainly have been lower.
Sounds like a reasonable assumption, but without seeing the original code, it's hard to tell. If the original CB was a bloated mess coded with little to no ability to easily be manipulated and re-purposed, then no, it might end up being the opposite actually. (You might end up spending more development time just figuring out what the heck the original code DOES for instance.)
And that is just one example of an alternative schema that I think would have worked better.
Maybe- maybe not.
Anything seems good when you assume all the risks involved don't pan out.