Pacific Connection(英語)

The Google Empire

When you log onto the Google search engine, the first thing you notice is its simplicity. There's the Google logo in its trademark colors of blue, red, yellow and green. There's a single field for your search, and just two buttons-one that canvasses over four billion Web pages, and the other, called "I'm Feeling Lucky," that takes you directly to the first hit. There are a few non-descript links to other Google services. But compared to any other major website you can think of-Yahoo!, MSN, eBay-the Google homepage is a Zen-like oasis of tranquility. There is no advertising, animated pictures, popup windows or sounds.

That simplicity belies one of the most cunning strategies for making money on the Web while offering one of the Web's single most useful tool. When Google came on the scene, it competed with at least a half dozen other browsers. But because Google searches honed in on the desired results more quickly, it became the research tool of choice on the web. (The name Google comes from Milton Sirotta, the nephew of American mathematician Edward Kasner. Sirotta, then a child, defined a google as a one followed by 100 zeroes. A googleplex is even larger-10 to the 100th power.)

Google appeared at a time when all the buzz was going toward "Web portals" the archetype of which remains Yahoo!. A portal gives you access to all kinds of things-weather, stock reports, movies. As a central hub, a portal makes good sense as a home page. And yet Google has managed not only to survive, but thrive, with a very different model and a very different look. Maybe the two go together. Taken on its own, the Web is a sprawling morass. It is chaos. It is a sea of self-serving URLs, personal web pages of people you never heard of and companies you don't care about. It is filled with bad designs and dead links-scattered thinly among them, a few places you'd really want get to. Google's simple design seems reassuring. It seems to say: you've come to the right starting place. Enter something and we'll take you

Hidden behind Google's simple search interface are a host of tools. Enter a calculation and it will act like a calculator. Put in an address and you get a map. Put the suffix "define:" before an entry and it will give you a dictionary definition. Google not only finds HTML pages, but 11 other formats, including PDF, Microsoft Office, Postscript, Corel, and Lotus 1-2-3. You can give it a specific web page in the advanced search and it will confine its search to a specific website. Give it a tracking number for a carrier like UPS or FedEx and it will tell you where your package is at. Give it a stock symbol and it will give you a quote. Give it a U.S. flight number and it will return the status of the flight; or put in an airport code with the prefix "airport" and you get weather conditions and delays. Use the suffix "link:" followed by the URL and you will see pages that point to that URL.

One of the hallmarks of a Google search is that it is fast, because it is highly distributed. Google is famous in the Linux community for its use of the operating system, which runs on thousands of servers. Whereas most search engines employ just a few large servers, Google uses many linked PCs to plow through the web.

Google is also famous for how it looks up information. Google's approach looks, in part, at what pages are linked to a given page. The company claims its "PageRank technology" employs more than 500 million variables and two billion terms. "PageRank interprets a link from Page A to Page B as a vote for Page B by Page A," says Google in its technology overview. "PageRank then assesses a page's importance by the number of votes it receives. But not all "votes" are given equal weight. A link from what Google ranks as an important page is given greater value. The search engine also analyzes page content. But because publishers can manipulate the content of a page to try to elevate its importance, Google also considers such things as the size of the font and the location of each word and the content of neighboring web pages.

These days, of course, Google is much more than a search engine. It is becoming something of an empire. The company keeps adding on features, some of which are for public consumption, others are strictly in test phase, with access sometimes limited. But even with these added services, the differences between Google and a classic portal are still pronounced. Google has avoided the all-in-one dashboard look. Its features are hidden away, each available only on a discrete screen. The company has made no effort to integrate its blog tool, shopping engine, Usenet search engine, and other features into a single location.

Google has not done much advertising. The company claims that when the boom exploded and competitors were spending millions on market brand campaigns, it continued to work on the quality of its search engine-relying on word of mouth to bring in more users. Google makes money in two different ways. Search for the term television, for example, and a right hand column gives a list of "sponsored links" including some television retailers and some shopping engine sites, including Yahoo! and BizRate. Google charges advertisers a small fee to list. But the real source of income is the cost-per-click model in which companies essentially bid for a user mouse click. Bids range from 5 cents to $50, and that fee determines the placement of the link or whether it actually appears there at all. For Google, the result has been a steady stream of income, and yet for Google users, the place links are unobtrusive and are never confused with the unsponsored search results.

From BackRub to Google

Google was founded in September 1998 by Larry Page and Sergey Brin. The two met in 1995 as graduate students at Stanford University. Page, a 24-year old alumnus, was visiting from the University of Michigan. Brin was one of a group of students showing him the campus. Page, the son of a computer science professor, once built a working printer out of Lego blocks, and created a server environment out of cheap PCs. Brin is the son of a NASA scientist and a math professor, and was born in Russia. The company now employs more than 1900 people and is headed by Eric Schmidt, who was the lead researcher at Sun and later headed Novell. Brin stays on as the head of technology while Page heads products.

The two collaborated on a search engine called BackRub in early 1996, named for its ability to analyze the links that pointed to a given website. On the cheap, they bought a terabyte worth of disks and built the company's first data center in Page's dorm room. For a while they tried to find buyers who might license the technology, including a friend of theirs, David Filo, founder of Yahoo! But there were no takers. As one CEO of a portal told them, "As long as we are 80 percent as good as our competitors, that's good enough. Our users don't really care about search." An early investor was Sun co-founder Andy Bechtolsheim, who invested $100,000. Eventually the initial investment grew to a million dollars. In September 1998, Google relocated to Menlo Park near Stanford, operating out of a garage complete with clothes washer, dryer, and hot tub.

In February 1999, the company moved back to Palo Alto with eight employees and a service that replied to more than 500,000 daily queries. Google signed on Red Hat as its first commercial search customer and secured $25 million from two Silicon Valley venture capital firms, Sequoia Capital, and Kleiner Perkins Caufield & Byers. On September 21, 1999 the company finally took the beta label off its website. (That tradition of long Beta trials remains a Google hallmark.) The $25 million let the company move to its current location back in Mountain View, which it calls the Googleplex. The headquarters is known for it's unorthodox design, including rubber exercise balls that are used as mobile office chairs, lava lamps, and large pet dogs. Larry Page has been known to traverse the building on roller skates. The company hired a chef, Charles Ayers, who once cooked for the Grateful Dead. It roped off its parking lot for roller hockey games, and in May 2000, passed a billion conducted searches of more than a billion pages.

Later in 2000, the company announced partnerships with NEC's BIGLOBE portal and China's NetEase. In 2001, the company expanded it's search to Usenet by purchasing the assets of Click on the term groups on the Google search engine and you can search through years of Usenet postings. The company signed agreements with Lycos Korea, with Universo Online in Latin America. It opened new sales offices in Tokyo and Hamburg, supporting searches in multiple languages.

The Google empire has kept expanding. The company introduced Google News in September 2002, providing news stories from 4500 different sources around the world. While the page looks as though it were edited by human being, a computer program does all the selecting and arranging. Three months later, the company launched Froogle, which searches websites for products and prices. In 2003 the company acquired Pyra Labs, the home of Blogger. Google even has a useful task bar, downloadable for free, that is installed on Internet Explorer, making Google searches easier while preventing pop-up windows.

Google has generally done a good job of staying non-controversial. It avoids pop-up Windows that drive users crazy. Its user interface is so simple that you can access it from any operating system, using any browser. Its ads are subtle-not in your face. But with its new email service, Gmail, Google ran into some rougher waters.

On the face of it, Gmail looks like a good deal. It offers a full gigabyte of storage, much more than its biggest competitor: Microsoft's HotMail. The interface is as clean as the search engine-simple, easy to understand, highly functional. The comparison to Hotmail's crowded screen is striking. You can use Google's search engine to run through your own correspondence. The company makes a point of this, saying you don't need to categorize your letters-just go ahead and surf them. Gmail doesn't tag ads on to the bottom of each email. It's a nice deal-so much so that for a while, black-market Gmail accounts were selling on eBay for around $80 each.

The problem, for some, is that text from incoming mail is scanned and while no human reads it, the content is used to display advertising-or rather, a set of "sponsored links" similar to those found on the search engine. That practice has alarmed privacy advocates like the Electronic Privacy Information Center, and the Privacy Rights Clearing House. The company has said that contrary to rumor, it does not index the messages it receives. No information is sent to advertisers. Nothing is free in this life. Gmail, like the rest of Google, looks like a great deal.

Sidebar: The Google IPO

As befitting its corporate culture, Google's public offering raised eyebrows among investors. Its official registration statement filed with the U.S. Securities and Exchange Commission, starts with a letter from the founders entitles: "An owner's manual for Google's shareholders." The first two sentences set the tone: "Google is not a conventional company. We do not intend to become one."

The mechanics of its offering were also unconventional. The company is doing what eBay users might recognize as a "Dutch auction" for its 150 million shares. Bidders obtain a bidder ID, specify the quantity of shares you want and the amount you are willing to pay. After the auction, the underwriters will figure out the lowest bid in which all shares would sell-and that will be the initial share price for everyone who bid that amount or higher.

What the shares are actually worth is another matter. Even some observers who like and use the service fear that unconstrained enthusiasm, undiluted by wisdom, will push the shares' value well beyond their worth.

Sidebar: An interview with Danny Sullivan, creator of Search Engine Watch

Danny Sullivan has been covering search engines since 1995, when he undertook a study of web pages indexing. The study, along with tips and information, was published online as "A Webmaster's Guide To Search Engines," which lead to the Search Engine Watch website. An American living in England, he spoke to me by phone from his office near Salisbury.

Why Is Google so successful?

Because their search results have very good relevancy-which is because they had better technology when they first debuted. Today, the technologies aren't so different, and others have caught up.

Google claims to have done two things-figuring out what links are pointed to a given page, and figuring out whether the search term plays a prominent role on that page.

Yes, that's the mixture. The thing that really brought success to Google was that they leveraged links across the Web and mixed that into analyzing page text itself. Earlier search engines relied primarily on what are called "on page" factors-the words on the page. Google came along with a very robust, powerful, relevant way of looking at off-the-page factors-in particular looking at links, understanding which links seem to be more important than other links, and use the words that are in and near links to determine what the linked pages are about.

Has that mixture become a closely held secret because of people trying to make their own web pages more visible?

The exact nature on how Google works isn't revealed, and that's true of other search engines as well. The basic principles are pretty well known, though, and that's why if you type the search term "miserable failure" into Google, you get George Bush's biography. That shows that that the formula is not as much a secret as one might think. But when you deal with topics where there are many links, it's hard to have that kind of an influence on Google or any other search engine.

The top three search engines are Google, Yahoo! and MSN. Is the distinction then not so much in what they pull up, but in things like positioning and market perception?

That's still part of it. Google now has a brand reputation for good search, so it doesn't matter if anybody else is better in search. I'm that way with Coke-I'm convinced Coke is better than Pepsi because I like the Coke brand, but actually Pepsi may taste better. But people should be trying other search engines. And while Google does have that great advantage-people should realize there are other things as well.

What other search engines are out there that are worthy and overlooked?

Ask Jeeves is the main one that's overlooked. Yahoo! is still overlooked by many people.

Yahoo has changed it's search engine more than once.

Yes, but it is still a good search engine. Yahoo! Has been overlooked in part because it was basically being powered by Google, so there wasn't a lot of reasons for trying Yahoo! rather than Google-you might just as well just use Google. Now, Yahoo! definitely is something people should be looking at. Yahoo! is my number one choice after Google, with Ask Jeeves running a very close third.

Are there specialized search engines that have a niche of their own?

There are all sorts of specialized search engines. FindLaw is an example. They run a crawler for finding legal information. is a science search engine. focuses on business-to-business activity.

Has Google succeeded in part because of its simple opening screen?

Yes. That started out because they only had search, and because their home page was designed by Sergey Brin, who is not a Web page designer. And it's worked.