Take Isabel, a social worker from Kenya. Ramani got her a job interview in May at a nonprofit that helps HIV/AIDS patients in San Francisco. When random Web searches left Isabel more confused than enlightened about her new field, "I sat her down in front of the computer and called up Google," recalls Ramani. "I typed in "HIV AIDS, San Francisco, organizations," and up popped all these Web sites on the topic. Isabel said: Oh my gosh! That's exactly what I need."
WORLDWIDE CLEARINGHOUSE. Ramani's experience is probably typical of that of the 81 million people who visit Google each month. The site provides answers to upwards of 200 million queries each day -- as many as 75% of all Web-search requests in North America by some estimates. It does this either through Google.com or via the Web interfaces of partner sites such as Yahoo! (YHOO
) and the America Online unit of AOL Time Warner (AOL
), which pay Google per query (but keep the revenue from advertising they sell against its searches).
Privately held Google doesn't reveal its financial results, but analysts say its revenues were about $300 million in 2002, when it delivered operating margins of 20% to 30%, and that its sales are expected to more than double over the next year. In short, Google's role as a worldwide clearinghouse for information has made it the social Web's linchpin.
Google has grown so fast that even those who use it religiously have started to worry about the tendency of its huge audience to increasingly regard the site as comprehensive when in reality its compilation of information is anything but. Here and there, moreover, Webmasters are starting to question the opaque rules that Google uses to determine what to index -- or not. Some businesses complain that because of these rules, they've been banished from the search engine without warning, explanation, or cause -- a development that can cost them dearly in terms of lost sales. All of which raises the question: Is Google getting too big for its britches?
REGULATION NEEDED? "It's the kind of question the Internet is putting to us again and again," says Jonathan Zittrain, co-director of the Berman Center for Internet & Society at Harvard Law School -- "the Net feels awfully public, but in fact almost every element of it isn't public or government-run or even subsidized. When these private parties start making individualistic decisions about who's favored and who's not, it does create serious legal questions." Zittrain argues that some regulation of Web searching might prove necessary should Google kill all its competition and become the only viable search provider.
That the issue even comes up is astonishing for a business that didn't exist until five years ago. That's when Larry Page and Sergey Brin, two Stanford University computer-science grad students, got a good idea for a search engine based on an innovation called PageRank. The system works by analyzing the linking structure of the Internet as a way to rank pages. Essentially, a Web site's influence and ranking are determined by counting how many other sites link to it.
To uncover this hierarchy, Google scours the Net using autonomous software called "bots." They scan the Web continuously, screening out blatant attempts to game Google by loading up a site with extra links -- while also giving extra weight to pages that are linked to by very prominent sites, such as CNN.com. This approach isn't really new. In academia, researchers have long rated the importance of any single paper by how many times it's cited elsewhere.
LET'S GOOGLE. What's unusual, however, is Google's ability to quickly deliver the information a Web visitor wants in a simple, uncluttered format. That's why it continues to enjoy enormous popularity, despite attempted incursions by competitors such as AllTheWeb (purchased by paid search provider Overture last March) and Teoma (which search company AskJeeves bought in September, 2001) -- both of which now rival Google in search technology.
Indeed, Google is the only site whose name has become a verb. Like Faith, most job applicants "google" their potential employers and vice versa. And googling a prospective date to learn more about the person has become de rigeur. If no more than six degrees had previously separated all the people of the world, Google has surely cut the number to four.
Take the case of Kate Ulbricht, a senior pharmacist at Massachusetts General Hospital in Boston. She's also one of the founders and chief editors of NaturalStandard.com, an online journal and research tool for physicans and pharmacists that seeks to catalog discoveries in the emerging field of nontraditional medicine. While Ulbricht regularly checks a host of well-known medical sites, she says she could never accomplish her task without Google. It can automatically translate Web pages in 12 languages into English, thus widening her web of contacts to researchers whose work she previously wouldn't have been able to read. Within the next year, much of that accumulated knowledge will go into the first-ever comprehensive U.S.-published textbook summarizing research into nontraditional cures.MISSING POLICIES. That's the shiny side of Google -- the friendly search engine that makes the world a smaller and better place. But as Google has moved toward holding a de facto monopoly on Web searches, critics have started to see another side -- one that worries them. Privacy advocates have begun to question what Google is doing with all the data it collects on the people who do the hundreds of millions of searches it processes. According to the Web site Google-Watch.org, a nonprofit group dedicated to informing the public about the search giant's inner workings, Google has no policies protecting the confidentiality of information it collects on its visitors via Web cookies. (Google execs declined to comment for this article.)
Recently, Google Watch nominated Google for the 2003 Big Brother Award, a tongue-in-cheek honor it gives each year to presumed privacy offenders. What data Google collects and keeps are of more than academic interest to privacy advocates: In an October, 2001, analysis of the Patriot Act, respected cyber-liberties advocacy group the Electronic Frontier Foundation warned: The government may now spy on innocent Americans' Web surfing, including terms entered into search engines, by merely telling a judge anywhere in the U.S. that the spying could lead to information that's "relevant" to an ongoing criminal investigation." The biggest repository of such information would be Google.
Even more disturbing to some Web-search observers are the contextual nuances that can influence Google's results. For example, type in the search term "Ruby Ridge" -- referring to a 1992 shootout in Idaho in which the FBI killed an alleged white supremacist -- and Google returns thousands of pages with tons of information. But hardly any of the links at the top of the list provide balanced information on the issue. Rather, most of the top results present extreme views of the incident.
WHAT ABOUT BOOKS? Were Google less influential that probably wouldn't matter. But as Americans, especially young ones, come to regard the leading search site as the source of all human knowledge, the effect could become pernicious. "People perceive it as the one and only place they need to go for information," says James Rettig, head librarian of the University of Richmond's Boatwright Memorial Library, in Richmond, Va. "That's unfortunate, because people who use only search engines will miss things" -- such as books, which represent most of humankind's body of knowledge to date.
Rettig notes that most of those don't show on Google and that many that do are buried in its rankings. What's more, he predicts, most books that have ever been published will never end up online. "There's going to be this vast retrospective body of information and knowledge that never becomes electronic," he says. "The costs of conversion are astronomical, especially compared with the potential payback."
Call it the Google Gap -- the difference between the growing perception that the site is omniscient and the fact that it isn't. This can be particularly damaging for businesses that run afoul of the rules Google uses to ensure that sites don't manipulate its rankings. While Google's concern over cheaters is justified, too often its rules punish innocent sites, critics claim.
"LIVID" WEBMASTERS. Witness the case of haute couture swimwear concern Graham Kandiah, whose site Google banished in May. John Hutzler, president of New York-based Now Interactive Services, which manages Graham Kandiah's online presence, believes Google shut the site out because of a Web-page popup that was used to launch a music player -- a popup that surfers would never see. But he doesn't know for certain because Google never answered the outfit's questions about what had happened. In the three days it took to diagnose the problem, Graham Kandiah says it lost 20% of its traffic -- which might have been avoided had Google simply sent it an automated e-mail explaining the reason for its action.
Accounts of similar scuffles with other sites permeate Google discussion groups on the Web. Many Webmasters "are livid about what's happening," complains Hutzler. He ended up purchasing advertising keywords on Google to get Graham Kandiah back on the site. But even though he's paying, he still can't use popups in Graham Kandiah pages.
Should Google further consolidate its control of the search world, argues Harvard's Zittrain, the Web could become the equivalent of a company town. That prospect prompts him to argue that "we need a concept of the public interest for the Internet that we haven't completely worked through yet. Somewhere between 'the market will magically take care of this' and 'let's regulate the heck out of it,' there has to be a solution."
Zittrain isn't sure exactly what it would be. He suggests that Google may be amenable to suggestions, since it prides itself on listening to the faithful and making their wishes come true. If Google wants its image to remain untarnished, it may have no other choice. By Alex Salkever, Technology editor for BusinessWeek Online