There's still a lot of mainframe stuff out there in cobol, after all. And a crazy amount of DOS. Business moves sloooooow.
The beauty of what Google's doing (and a lot of their playing is creating a pointillist map) is in this space: they're making it so I can get (right now mostly) text from one place to another through multiple streams to a final display regardless of ANYTHING. There are still some places file metaphors exist (docs), but that's fading in their newer products. Add robust lines of demarcation (privacy/security via groups) and the 'features' of Word don't beat the 'killer app' of portable, trackable, malleable information that people can collaborate on without effort.
So: watch out, Office. (Though I love what you're doing in Excel)
You know the main problem with Excel? Finding out what changes were made, keeping updates between multiple teams/people in sync, sending the correct version of the doc. once updated, etc.
If Google Docs Excel was as good as Excel today (it still lacks quite a bit, esp in UI), we'd never use Excel, simply to save a huge amount of time syncing everybody.
However, there is a huge ecosystem around Excel - lots of enterprise apps have tight Excel integration and an amazing amount of complex stuff gets done in VBA in Excel.
As to your other point, I've actually persuaded quite a few (very bright) finance people how to use TortoiseSVN and they picked it up pretty easily and seems to work very well for them. Wire it up so that you use a sensible diff tool to show differences between revisions and it actually is pretty civilised. Almost like real code! ;-)
But it's by no means the only option if they fall behind.
I disagree with this basic premise. Let's put our investor hats on. There are these businesses, Windows and Office. They throw off a certain cash flow. It doesn't last forever.
So what to do? First, we re-invest some of that cash flow into extending the life of the cash flow from the profits themselves. Advertising, new versions, and so forth.
Second, we can pay a dividend to investors. If the taxes are onerous, we can do magical things with trusts or some such to reduce taxes.
Third, we can invest the cash flow in new businesses, basically treating ourselves as a VC or startup lab. We hope to turn the new businesses into streams of cash flow.
WHy do we assume that options one and three are always on the table? If investors can get a better rate of return investing the cash flow from Office and Windows than Ballmer is delivering, why are we asking Ballmer to 1. Generate cash from Office and Windows and 2. Produce better returns than we can?
I think the time has come to stop expecting so much from SteveB. Get him to generate tons of cash from Office and Windows and give the money to investors. Let them decide whether to invest that into Google stock or Apple stock or the next YCombinator round or whatever they like.