Other reasons to prefer the multi-app approach would be:
-Minimal programming: Even though programming can be a revenue stream, it also slows down turn-around time on jobs and may be a barrier to more profitable parts of a job if they get cut from the budget. It also makes those jobs harder to service in the future.
-Apps requires little to no intervention on our behalf when things change: For instance, Sonos will adopt new music services far before most "integration friendly" companies or solutions will, and it's a fully automatic process. Thinking of any arbitrary subsystem, it's nice to know the manufacturer is usually quietly working behind the scenes to roll out a new feature, such as when Lutron released the level editor for iPhone. Subsystem manufacturers have way deeper pockets than we do and can afford to develop the apps far beyond what we may be able to, all in the name of keeping the customer happy which is OK by me.
When you bring A/V control into the mix, one way we have done it that works well is to use something along the lines of URC CCP or ProControl which are a self-contained A/V control system complete with app.
All this being said, I still think a truly integrated solution is right for most installations of let's say $25-30k or better.