(First, thanks/apologies to Andy Baio for listening to my musings on a mullet-style app strategy and coining "Mullet-platform". It's horrible and wonderful.)
So, here's the theory: The web is good at driving big audiences of engaged users, and enables awesome stuff like viral adoption and social sharing and all the other benefits we get from permanent, easily-shared links. Apps are great at driving paid user revenues, and creating experiences that stay with people wherever they go, and connecting to smart sensors and networks.
Thus, drive the business up front with the paid revenues from mobile apps, and the party in the back with the unfettered social sharing of the web. Mullet! With apologies to Jonah Peretti's mullet strategy for content, this seems to be as appropriate an analogy for revenue models as it is for the HuffPo/BuzzFeed method of publishing a site.
Let each platform do what it's good at! Now of course, there's some overlap between what each platform can do, but this is about making a rational justification of how the platforms relate to each other. And if you follow the premise behind the idea that you should stop publishing web pages you very quickly arrive at the straightforward conclusion that you'll have to have a platform-neutral cloud service that powers these two different experiences for apps and web in a way that lets them be fully native on each. It should also make it clear how to assign resources to the two different types of platforms, since each has a distinct role to play in helping your app or game or content succeed.