Yeah, the correct strategy is to hire a new parallel team (or, more practically, external developer again) to build a fresh platform from the ground up, optimized for multiplatform play, with a completely redesigned UI that supports common use cases like streaming. They'll have to pay a lot (I would guess on the order of $10 to $15 million) to do it right, but it'd pay for itself in a year.
They're stuck on this to some degree because of the paper game -- they can't lower pack prices and they can't move to a generous F2P model because they'd have to eliminate redemption.
The area they really have room to maneuver is phantom events. People -- especially bad players -- love to play some of these formats just because they're enjoyable and don't care about the prizes. Running way more phantom events, tweaking the pricing model (maybe let people buy into a season for a flat fee, idk) and changing the prizes up a bit would open up MTGO to a lot of more casual players, which would in aggregate have a very good effect on the game's popularity.
I played it at PAX East when it was revealed and this was pretty much my exact takeaway -- it's like Magic, but uglier and less strategic -- but it's cheap and fast to play which counts for a lot.