Searchology: Google Employee #1 Speaks

A Google technology director named Craig Silverstein is talking about his first days as a Google employee--and it's an interesting topic, because he was the very first person hired by founders Sergey Brin and Larry Page, back when Google was a research project rather than a phenomenon.

"If Google had started a year or two earlier, it wouldn't have worked," he's saying, contending that in 1995 and 1996, the Web was small enough that folks didn't need a search engine to find stuff. By 1998, when Google launched, search was useful--but the initial Google search engine only spidered 25,000 pages which the algorithm believed to be important and of high-quality. Which meant that lots of stuff was left out.

Speed, Silverstein is saying, was a Google mantra from the get-go, but it was hard to balance that with comprehensiveness. In 2000, Google hit one billion indexed pages, an occasion that led to it calling itself "Giga Google," he says. But the company had trouble getting its west-coast servers to sync up with east-coast ones that would allow easterners to have a speedy search experience. It took a few days, but it fixed this problem.

Then there's relevance. Google knew from the start that people wouldn't burrow through multiple pages of results, but it became clear quickly that many wouldn't even scroll down the page. It was "supremely important" that the first couple of results be strong.

Google's impressive relevance led to favorable word of mouth, he's saying-which was good, since Google didn't really advertise. Silverstein knew the company would do well when a high-school buddy had heard about it.

Now Silverstein is talking about speed. Pre-Web, you might travel to a library to get information (slow) or just ask a friend (possibly faster). Then there were early Internet search tools like Archie, which were useful but not quick or simple.

"I can't imagine anybody using Archie to figure out how to open up a pineapple," he says, using an example that came up in a video of real people talking about Google. But if a search engine can get you information in about a second, why not use it as your principal source of all sorts of information?

"Google is now very fast...we'd like to do better, but there are some limitations we just haven't figured out." Like the surpassing the speed of light, and the fact that the planet is a large place. But the company has established data centers all over the earth to keep Google zippy everywhere.

In the early days, though, the European Internet infrastructure was fairly rudimentary--and it turned out that it was faster for European search queries to travel all the way to the U.S. and back.

Now Silverstein is showing the evolution of the Google home page over the years. He's pointing out that in the early days, there was a signup for a "Google Friends" newsletter right on the home page. Today, the home page has some stuff it didn't back in the day, such as tabs. But he's saying that it's still extremely important that it be simple and snappy.

In the early days, he's saying, Google wasn't very good at testing: At one point, a bug led the site to deliver the ten worst results. Users noticed. Today, the company is serious about testing new features before rolling them out.

Silverstein is saying that four things have always defined Google, and probably always will: comprehensiveness, relevance, speed, and user experience. And that's the end of his talk.

Comments

Post a comment Post a comment

0 / 1000 max characters