Going after Windows and Office
In the business world I prefer being aggressive, and if someone is coming after me, I in turn am going to go after them -- but only if I think I can win the battle. So that becomes the question, can Google win this battle?
From a branding perspective, I think the only organization that can attack MS must (a) have a lot of money themselves, (b) have great technology, and (c) have a great brand. (Like Apple) Google fits all of those descriptions. So it seems possible they can win this battle.
Strategy 1: Attack Windows
Step 1 is creating a "GoogleOS" that is a binary-compatible replacement for Windows. That is, it must be something like a combination of Linux and WINE. This operating system must look like enough like Windows so minimal training is needed for users to make the transition. And, as the binary-compatibility implies, it must be able to run native Windows applications. Finally, with the Google brand, you have to convince IT managers around the globe that it's safe to make the switch. (Yes, I know that's a tall order, but that's what is needed.)
Beyond this, IT managers and consumers won't just switch to a Google operating system because it's made by Google, there must be something else. It can be things like a tagged-based file system, but I think it needs to be ... free, and open source.
I think this is the lure needed for organizations around the world to make the switch: you're promising to save them tons of money, and finally give them a viable desktop choice. I don't think there are that many IT managers that love Windows, in fact, I think most of them probably feel like they're stuck with it.
Strategy 2: Attack Office
Beyond Google's web applications approach, I think they can also attack MS Office much more effectively using one of two other options. One possibility is to start with the OpenOffice source code, and overhaul the user interface. IMHO, the OpenOffice designers/developers made a mistake in creating a product that requires a significant learning curve to replace MS Office. Put simply, the UI needs to be overhauled to eliminate this learning curve barrier. Ideally the replacement is nearly a click-for-click, keystroke-for-keystroke replacement for the existing MS Office applications.
Second, if OpenOffice is not the answer, the other obvious option is to buy your way into the market. I haven't looked at alternatives very hard, but my guess is that the best thing going is the Corel Office suite. Again, once you have the source code, overhaul the UI to make it as familiar to people as possible.
It's also a given that with either of these two choices you also need to enhance the products to integrate seamlessly with Google Apps. Personally, I don't like web applications as a replacement for Office, but if you really believe in the viability of web applications to replace fat client applications this is an obvious step.
Strategy summary
The end result of this strategy is that new PCs from Dell, HP, and others are available with either Windows or the new, free GoogleOS installed. And no matter which one you buy, Google Office is installed for free. (And in the GoogleOS, Google can integrate their search technology as deep as they want into the operating system. And hey, go ahead and throw in that tag based filesystem while you're at it.)
But they have to make money, right?
This all sounds great, but Google has to make money, right? They can't just spend money to develop a GoogleOS and Google Office to fight the battle on the MS turf, can they?
First, wouldn't you have said the same thing about Microsoft creating IE? Surely they wouldn't invest development money in a product and give it away for free just to kill a competitor, would they?
Second, I think this depends on cost. Whereas IE was much smaller (initially), creating a Windows binary-compatible operating system is probably a much more significant expense.
In short, if it could be done for a reasonable price, hell yeah, I'd invest some money to gut the cash cows of a competitor that's coming after my turf. In my mind, I keep coming back to the fact that MS will keep coming after you, as long as they have the Windows and Office revenue streams. If you cut those off, do you think MS will keep pursuing you? If you do it now, I think MS might back off, as they have no search and ad credibility. But if you do it years from now that same opportunity might be gone.
Third, these two products don't have to be total "loss leaders". Companies like Red Hat and even Sun have built service organizations around open source products, and that's also possible here. I don't know if Google has any type of service organization at this time, so that may be a big impediment.