As Google’s Atmosphere conference was winding down on Monday afternoon, I sat in the press box and ruminated on the vague discomfort that had been nagging me since lunch. Something was bothering me about the way the CIOs that Google had gathered to give testimonials were talking about “going Google,” but I couldn’t quite put my finger on it.Then suddenly, the image of Phil Hartman popped into my head, and I had my answer.

I imagined that I was a CIO who was out hiking on some glacier in 2001, and I fell in and was frozen there, only to thaw out ten years later (thanks, global warming!). My company was so glad to see me that they immediately gave me my old job back, and then they shipped me off to Google Atmosphere 2011 to get up-to-speed on this whole “cloud” thing that I had missed out on. What would I, the unfrozen caveman CIO, think about “the cloud” after a full day at Google’s conference? The answer is easy: I’d think that “cloud” was synonymous with “Gmail.”

Of course, Google was keen to present a multitude of cloud offerings at the conference, with Gmail as just one of ways that businesses can make the jump into the cloud with Google. But it all kept coming back to Gmail so often that, instead of Gmail being a very important but narrow example of the more general “cloud” in action, it felt more like Gmail was the whole cloudy enchilada, with a few other things (docs, a spreadsheet, video chat) thrown in as cheap side dishes.

Google was definitely aware of this Gmail-centric dynamic, as evidenced by the fact that Google VP of Product Management Dave Girouard kept insisting to attendees and the press that “email isn’t dead yet!” and stressing just how critical it still is to the enterprise. So there was some nervousness there around the fact that Gmail is to Google’s cloud offerings what AdWords is to Google’s bottom line—i.e., it’s the bread-and-butter, and everything else is sort of nascent and aspirational.

Now, to be clear, the assembled CIOs were asked in a panel if they were using the non-Gmail parts of Google Apps, and they answered in the affirmative to varying degrees. But it was obvious that none of them did a Google migration based on, say, the docs or the spreadsheet. Time and again their remarks came back to how many disparate email systems Gmail had replaced, and how much money that was saving. CIOs are converting their employees to Google Apps in order to get away from the headache of legacy email systems, and in the process they’re also trying to entice employees to switch from Microsoft Office to Google’s competing offerings.

Microsoft Office is a network, so Metcalfe’s law works against Google Apps

It’s worth thinking about why Gmail is the main way that Google is getting into the enterprise. Email is a communication network, and as such it obeys Metcalfe’s Law, which says that the value of a communication network is proportional to the square of the number of users. So every person who signs up for an email account anywhere on the Internet adds in a nonlinear fashion to the aggregate value of “email” as a network.

Given Metcalfe’s Law, there are two important things to recognize about email: 1) after the DNS system, it represents the largest, most valuable communications network at the application layer of the Internet, and 2) when you change email providers you don’t switch networks, so you retain the full value of the network for yourself and other users.

All of what I’ve said about email above is intuitively obvious and often recognized. But what’s not often recognized is that Microsoft Office is also a communications network that runs on top of the Internet at the app layer, and specifically it runs on top of the email layer. In the legal profession, Microsoft Word files, specifically with Word’s “track changes” feature enabled, is the standard way that lawyers communicate and collaborate with each other electronically. In finance, Excel is the standard way that every one in every part of the markets communicates and collaborates around the numbers and models that make markets run.

In general, all businesses everywhere communicate internally and externally via a critical Microsoft Office “network” that runs on top of the email network in the form of document attachments. Any of those businesses can move to Gmail without losing access to that incredibly large and valuable Office network, but they cannot move to Google Apps’ other offerings without losing access it. So Metcalfe’s law actually works against Google’s efforts to migrate users from Office to Google Apps.

Google Apps faces essentially the same problem with getting businesses to migrate to its non-Gmail products as ICANN faces in getting everyone to migrate to IPv6. It will be hard to do this gradually, and companies will have to keep up with their Microsoft licenses while the transition takes place. It would be easier if everyone held hands and jumped at the same time, which is what Vint Cerf told Atmosphere attendees that he hoped would happen on the upcoming World IPv6 day. But it’s not going to happen, so Google faces a long, grinding, uphill battle to get users to switch to its non-Gmail offerings, and that battle is largely independent of the quality of those offerings. As long as Microsoft Office remains a critical communication network runs atop email, Google will be confined to owning just the transport layer of the enterprise collaboration stack.