Schmidt on scaling Google

The first time I heard Eric Schmidt speak was in June 1995. I’d flown to Honolulu to cover the annual INET conference for the newspaper I then worked for. The Internet Society’s conclave was a sort of victory lap for the wizards and graybeards who’d designed the open network decades before and were finally witnessing its come-from-behind triumph over the proprietary online services. It was plain, at that point in time, that the Internet was going to be the foundation of future digital communications.

But it wasn’t necessarily clear how big it was going to get. In fact, at that event Schmidt predicted that the Internet would grow to 187 million hosts within 5 years. If I understand this chart at Netcraft properly, we actually reached that number only recently. (Netcraft tracks web hosts, so maybe I’m comparing apples and oranges).

I thought of this today at the Web 2.0 Expo, where Eric Schmidt, now Google’s CEO, talked on stage with John Battelle. (Dan Farber has a good summary.) He discussed Google’s new lightweight Web-based presentation app (the PowerPoint entry in Google’s app suite), the recent deal to acquire DoubleClick, and of Microsoft’s hilarious antitrust gripe about it, and of Google’s commitment to letting its users pack up their data and take it elsewhere (a commitment that remains theoretical — not a simple thing to deliver, but if anyone has the brainpower resources to make it happen, Google does).

But what struck me was a more philosophical point near the end. Battelle asked Schmidt what he thinks about when he first wakes up in the morning (I suppose this is a variant of the old “what keeps you up at night”). After joshing about doing his e-mail, Schmidt launched into a discourse on what he worries about these days: “scaling.”

It surprised me to hear this, since Google has been so successful at keeping up with the demands on its infrastructure — successful at building it smartly, and at funding it, too. Schmidt was also, of course, talking about “scaling” the company itself.

“When the Internet really took off in the mid 90s, a lot of people talked about the scale, of how big it would be,” Schmidt said. It was obvious at the time there’d be a handful of defining Net companies, and each would need a “scaling strategy.”

Mostly, though, he was remarking on “how early we are in the scaling of the Internet” itself: “We’re just at the beginning of getting all the information that has been kept in small networks and groups onto these platforms.”

Tim O’Reilly made a similar point at the conference kick-off: In the era of Web-based computing, he said, we’re still at the VisiCalc stage.

Google famously defines its mission as “to organize the world’s information and make it universally accessible and useful.” But the work of getting the universe of individual and small-group knowledge onto the Net is something Google can only aid. Ultimately, this work belongs to the millions of bloggers and photographers and YouTubers and users of services yet to be imagined who provide the grist for Google’s algorithmic mills.

I find it bracing and helpful to recall all this at a show like the Web 2.0 Expo — which, while rewarding in many ways, gives off a lot of mid-to-late dotcom-bubble fumes. Froth will come and go. The vast project of building, and scaling, a global information network to absorb everything we can throw into it — that remains essential. And for all the impressive dimensions of Google, and the oodles of Wikipedia pages, and the zillions of blogs, we’ve only just begun to post.

[tags]google, eric schmidt, internet growth, web 2.0, web 2.0 expo[/tags]

Post Revisions:

There are no revisions for this post.

Get Scott’s weekly Wordyard email


  1. The big difference between now and the late 1990s is that in the late 1990s people had visions for the potential of the Internet to utterly change the way we live, but the necessary infrastructure was not yet in place (for example, high-speed Internet access was not available in most of the U.S.). So, it was a case of unrealistic expectations being chased by money that correctly understood the long-term potential of the Internet, but misunderstood the impossibility of it happening at that time.

    Now the infrastructure is in place. This is not a bubble, but the start of the actual long-term wave of developments that will indeed utterly transform the way humans live.

    It’s very similar to the history of automobiles. “Horseless carriages” in a sense failed, they were laughed at legitimately, because the necessary infrastructure (good roads) did not exist. A great many companies went out of business. But once sufficient effort was put into building the infrastructure, a long-term wave of “adoption” of automobiles (including trucks) began, lasting for decades, and indeed the way people live was changed forever.

    Web 2.0 is the start of that long-term wave for the Internet.

Post a comment