Google’s Brian Bershad on the Search Giant’s “Second Act,” and Building More Trust
Google owns the world’s most popular Internet search engine, but to hear the leader of the company’s Seattle site talk, building that was “the easy part.” All the information on the Internet was free and available for indexing with its clever algorithms, but that’s just a tiny fraction of the world’s information, when you start thinking about books, business documents, photos, videos, e-mails, voicemails, instant messages, technical databases, and more.
“The Web was wide open,” Brian Bershad said this morning in Seattle during a presentation at the Science & Technology Discovery Series hosted by the Technology Alliance. “With non-Web data, we need to get people to give it to us, and we need to give them a reason to.”
This “second act” in Google’s quest to organize all the world’s information and make it accessible is what got Bershad, a former University of Washington computer science professor, fired up about becoming Google’s Seattle site director in 2007. The search engine is now used by 597 million unique visitors a month, and is responsible for 60 percent of all Web searches, he said. Google now has about 20,000 employees, and about 350 engineers between its sites in Kirkland, WA and Seattle, he says. At the very beginning of his talk, Bershad made sure to put up a disclaimer about how some of the things he said are his opinions, and not always necessarily the company’s. I doubt he said anything that’s going to stir trouble in Mountain View, CA, but there you have it.
So here were some of the highlights of his talk:
On the “user experience”: One of Google’s clever insights early on was that users preferred a clean, uncluttered page to do searches. Now the company has “long fights” in-house about things like whether search results should show up in blue letters or red, or which information should be placed on the left or right side of the page. It may sound trivial to some, but it matters a lot, Bershad says. When Google experimented with putting sponsored links in red to draw more attention to them, it found it could draw more clicks, and more revenue, but people didn’t stay with the ads as long, and it turned off some customers. So they switched back to baby blue text for sponsored links. “If we give our users a great experience, we’ll make money,” he said.
On keeping Google searches up to speed with breaking news: Bershad told the story of how in August of 2005, people who searched for “Katrina” were not likely to automatically get results related to hurricanes or New Orleans. Google’s engineers work to make sure search results stay relevant and timely in reaction to news events that change what certain terms mean. This is one feature that distinguishes Google from other search engines, he said. “News changes what people are interested in quickly,” he said.
On relevance of search results: The company’s goal is to give the user the search result they want, and get them off the site fast. If the user comes back to re-do a search, that’s judged a failure. This is tricky, because Google needs to pick up on what a user means when they type in some vague term like “fruit.” It usually wouldn’t be a definition, but probably something else. “We want to give the user what they want,” he said.
On speed: Google measures performance on search results down to the millisecond, and every one counts. “When we improve speed, we improve revenues,” he said. Getting all sorts of data centers humming in coordination, with all sorts of redundancy in case hardware breaks down is critical. A couple months ago, the Google search engine went down for a few minutes, and although the world didn’t stop spinning on its axis, it seemed like it for a while inside the company, he said. “There’s a saying at the company that speed is our favorite feature,” Bershad said.
On news: “Our model is to take what newspaper publishers are willing to give us,” Bershad says. The latest experiment, sure to cause heartburn in many publishers’ offices, is to place ads on Google News, which aggregates all that costly newsgathering that other people do. Google is experimenting with this model as long as it doesn’t turn off its users, Bershad said.
On how Google and Amazon have different relationships with customers: Aside from Google Checkout, the company’s online payment service, Google doesn’t store the kind of personal information on its customers that Amazon does, like credit card numbers and detailed history on buying habits. (Google did, however, introduce a behavioral targeting system for its keyword-based ads this week.) “With Amazon, there’s money at stake. You give them your credit card and trust them not to clean out your bank account,” Bershad said. Google would like to have more personal information to help make search results more relevant for individuals, but this is a tricky balancing act with concerns about privacy. Google has a strong bond of trust with users, but isn’t at the same level yet as Amazon, he said. “Google is more at a distance,” he said.
Trending on Xconomy
By posting a comment, you agree to our terms and conditions.