Robin Li, center, signing the lectern during the close of Nasdaq trading on Aug. 5, 2005, Baidu’s first day as a listed stock on the market.

IN the summer of 1998 at a picnic in Silicon Valley, Eric Xu, a 34-year-old biochemist, introduced his shy, reserved friend Robin Li to John Wu, then the head of Yahoo’s search engine team.

Mr. Li, 30 at the time, was a frustrated staff engineer at Infoseek, an Internet search engine partly owned by Disney, a company whose fading commitment to Infoseek did not mesh with Mr. Li’s ongoing passion for search. Like Disney, Mr. Wu and Yahoo were also losing interest in the business prospects of search, and Yahoo — in a colossal corporate blunder — eventually outsourced all of its search functions to a little startup named Google.

Mr. Xu, who had called together some friends for a documentary he was making on Silicon Valley, thought the two search guys would hit it off. Mr. Wu says he exchanged greetings with Robin Li, but what most impressed him was that despite all of the pessimism surrounding search, Mr. Li remained undaunted.

“The people at Yahoo didn’t think search was all that important, and so neither did I,” says Mr. Wu, who is now the chief technology officer at the Chinese Internet company Alibaba.com. “But Robin, he seemed very determined to stick with it. And you have to admire what he accomplished.”

Graphic

Miranda Mimi Kuo for The New York Times

Robin Li’s passion for search led him to create Baidu, based in Beijing.

Indeed. A year after the picnic, in 1999, Mr. Li founded his own search company in China, naming it Baidu (pronounced “by-DOO”). Today, Baidu has a market value of $3 billion and operates the fourth-most trafficked Web site in the world. And Baidu is doing what no other Internet company has been able to do: clobbering Google and Yahoo in its home market.

While Baidu continues to gain market share in China — and does so with a Web site that the Chinese government heavily censors and that gives priority to advertising rather than relevant search results — some analysts question whether Baidu can withstand competition from Google and Yahoo, which possess superior technology and global work forces.

But Baidu’s evolution, and Mr. Li’s journey as an entrepreneur, offer textbook examples of the payoffs and perils of doing business in China and suggest that Baidu may prove to be far more resilient than some analysts believe. China has a population of 1.3 billion, about 130 million of whom are Internet users, an online market second in size only to the American market. Because China is the world’s fastest-growing major economy, analysts consider it the next great Internet battleground, with Baidu uniquely positioned to prosper from that competition.

In exchange for letting censors oversee its Web site, Baidu has sealed its dominance with support from the Chinese government, which regularly blocks Google here and imposes strict rules and censorship on other foreign Internet companies.

In addition, analysts say, entrepreneurs in China have a knack for pummeling American Internet giants. “The globally dominant U.S. Internet companies have failed to take the No. 1 market share position in any category,” says Jason D. Brueschke, a Citigroup analyst, of the Chinese market. “And they came with more money and major brand names. And so there’s something fundamentally different about this market.”

So fundamentally different, Mr. Brueschke believes, that Baidu will retain its hammerlock on the Chinese search industry. “The real battle in the competitive landscape is not about who’s No. 1, it’s about who’s going to be No. 2,” he says.

Google, of course, will have none of this, stressing the independence of its search results and the international reach it offers users. “People want information and they want global information,” says Kaifu Lee, the president of Google in China. “We can’t be bought.”

But Mr. Li says Baidu’s model is working supremely well and that the company has built a loyal base of users who value its search capabilities. “At the end of the day, if a user finds relevant information, they’ll come back,” he says.

ON its corporate Web site, Baidu says that it takes its name from a Song Dynasty poem written several centuries ago that “compares the search for a retreating beauty amid chaotic glamour with the search for one’s dream while confronted by life’s many obstacles.”

Mr. Li, born Li Yanhong in 1968 in what was then an impoverished city 200 miles southwest of Beijing, is familiar with life’s obstacles. The fourth of five children, he grew up during China’s brutal Cultural Revolution. Despite the oppression that surrounded him, he said he was always able to focus on stamp collecting, performing traditional opera and other interests — including, eventually, computers. He was bright enough to get into the country’s most prestigious school, Beijing University, where he majored in library science and dabbled in computer science.

The government infamously cracked down on pro-democracy demonstrations in Tiananmen Square in 1989 when Mr. Li was a sophomore, causing his college campus to be shut down. Mr. Li is mum on the events that followed, saying only that he was apolitical. But he does say that a year later he started thinking of studying abroad and that by the time he graduated in 1991 he was ready to leave his homeland.

“China was a depressing place,” he says. “I thought there was no hope.”

He applied to the top three graduate programs in computer science in America, but did not get into any of them (perhaps, he says, because China was considered an also-ran in technology). “I blindly sent out 20 applications,” he says. “SUNY Buffalo was the only program willing to give me a fellowship.”

He enrolled at Buffalo planning to earn a Ph.D. in computer science but grew disillusioned with academia. He completed his master’s degree in 1994 and then joined a New Jersey division of Dow Jones & Company, where he helped develop a software program for The Wall Street Journal’s online edition. During that time, he also became enamored of the technology boom taking shape in Silicon Valley. He spent much of his time trying to solve one of the Internet industry’s earliest problems: sorting information.

A breakthrough came in 1996, he says, when he developed a search mechanism he called “link analysis,” which involved ranking the popularity of a Web site based on how many other Web sites had linked to it.

“The moment I created this thing I was very excited,” he says. “I told my boss and pushed him. But he wasn’t very excited.” Soon after, he attended a computer conference in Silicon Valley and set up his own booth to demonstrate his search findings.

William I. Chang, then the chief technology officer at Infoseek, met Mr. Li at the conference and recruited him to oversee search engine development.

“Robin is possibly the single most brilliant and focused person I know,” Mr. Chang says. “And his inventions, now widely adopted, are still the gold standards in Web search relevance.”

After Disney acquired the small fraction of Infoseek stock it did not already own in 1999, it shifted the company’s focus away from search and toward content, leading Mr. Li to form his own Internet company with Eric Xu, who had a Ph.D. in biochemistry and good contacts in Silicon Valley.

The partners raised $1.2 million from two Silicon Valley venture capital firms, Integrity Partners and Peninsula Capital, and with their seed money in hand flew to China and founded Baidu in a hotel room overlooking Beijing University’s campus. Nine months later, in September 2000, two other venture capital firms, Draper Fisher Jurvetson and IDG Technology Venture, pumped another $10 million into the startup.

So it was that on the eve of the Internet bubble bursting in the United States, Baidu took off in China.

“When I came back I was prepared for a rough life,” Mr. Li says. “It turns out it wasn’t so bad.”

Baidu started out offering search services to other Chinese portals before developing its own stand-alone search engine. Some members of Baidu’s board of directors opposed the shift, saying it would turn customers into competitors. But Mr. Li said he sensed a shift in the market after watching the success of Overture, a company in Pasadena, Calif., that sold advertising space correlated with search results (which meant, for example, that ads for dental clinics might pop up next to search results for cavities).

“We were skeptical about search,” says Scott Walchek, a partner at Integrity Partners and a member of Baidu’s board. “But we weren’t as smart as Robin. Robin said he had a unique opportunity to build a brand around search. And he was right.”

In September 2001, Baidu began its own site — Baidu.com — which looked almost exactly like Google’s no-frills home page. And even before Google did it, Baidu allowed advertisers to bid for ad space and then pay Baidu every time a customer clicked on an ad. Small and medium-size companies loved it, the site became deluged with traffic and Baidu turned a profit in 2004. By then, Mr. Li was pushing for an initial public offering in the United States, insisting it would be a huge branding event for a company that had come to be called “China’s Google.”

BAIDU went public on Aug. 5, 2005, at $27 a share. When trading ended that day, shares of Baidu closed at $122, up 354 percent, the biggest opening on Nasdaq since the dot-com peak in 2000. Suddenly, Baidu was a $4 billion company and Mr. Li held stock worth more than $900 million. But not everyone cheered. Many analysts said that by almost every measure Baidu’s stock was ridiculously over-valued. It eventually tumbled to as low as $44 before rebounding. On Friday, its shares closed up $3.03 in regular trading, to $87.75, giving the company a market capitalization of about $2.94 billion.

At the time of the I.P.O., some critics attacked Baidu’s zealousness for ad revenues. They noted, for example, that a Baidu search for the word “cancer” turned up ads for hospitals that paid for top spots in results rather than returning information on cancer itself. In comparison, Google and Yahoo more clearly separate ads from relevant search results by placing them on the right side of the page.

The company’s revenue jumped 190 percent in the first half of this year, to $40.9 million; profit soared 550 percent, to $11.7 million. Baidu’s Web site is drawing millions of young people eager to download music files, create blogs or search for pictures of China’s “10 Most Beautiful Women.” While Baidu is growing fast, its revenue is still anemic compared with Google’s, which is expected to top $7 billion this year.

Analysts say Baidu is playing to a different audience than Western Internet companies because the Chinese are far more interested in entertainment than news, books or car rental rates. “The fact is 70 percent of China’s Internet users are under the age of 30,” says Richard Ji, an analyst with Morgan Stanley. “Most of them are single, only children. They’re looking for entertainment.”

That may explain why China’s dominant Internet companies are all entertainment focused, like Tencent (which hosts online communities and instant messaging) and Netease and Shanda (which are online gaming sites).

Yet no Internet company in China is growing as fast as Baidu, which had more than 50 percent of the pay-per-click market in the first half of year, up from a 37 percent share in the same period a year ago, according to Analysys, a research firm in Beijing. Google and Yahoo both lost ground, with each company holding 16 percent pay-per-click shares for the first six months of 2006.

Still, Baidu faces significant challenges. The company’s stock is in the stratosphere, putting pressure on management to deliver knockout growth every quarter. Google’s shares closed up $5.90 Friday in regular trading, to $409.88, meaning investors pay a hefty $60 for every $1 of profit in the stock, far more than other Internet companies. But Baidu investors pay a whopping $190 for every $1 of profit.

Baidu also faces legal challenges, including lawsuits claiming it violates copyright laws on music files. Baidu has been sued over the issue, but continues to provide links to sites that offer music files. The company says it does not believe it should be held responsible for simply offering linking to other sites. In a country rampant with claims of click fraud, a Beijing hospital recently claimed that Baidu orchestrated a scheme in which a Baidu affiliate kept clicking on the hospital’s ads to fatten the fees it had to pay Baidu. A Baidu spokeswoman says the company has not reviewed the case, but actively polices click fraud.

LOOMING on the horizon are Google and Yahoo. Google says it plans to spend hundreds of millions of dollars to compete in China, and Yahoo has merged its operations here with Chinese Internet behemoth Alibaba.com.

“Google is fierce,” Morgan Stanley’s Mr. Ji says. “And Alibaba has the best sales force. Baidu could get hurt on the technical side.”

But the Chinese market is littered with the wreckage of American Internet companies that have failed to dominate here. In 2003, eBay bought the largest Chinese auction company — and then lost market share. In 2004, Amazon bought the largest Chinese online merchandiser — and then lost market share.

Now, the real fight begins. Google, which invested $5 million in Baidu just before its public offering last year, sold that stake for a hefty $60 million in June. And now, Google is building up a huge research team in Beijing, not far from Baidu’s headquarters. But analysts say it won’t be easy for Google.

“The American Internet giants are dominant in the U.S. and dominant in Europe,” Mr. Brueschke at Citigroup says. “And then they come to China and fail. And so what I want to know is: What is Google going to do differently?”

For his part, Robin Li seems undaunted.

“Our traffic keeps increasing,” he says confidently. “We’re now the No. 1 Web site in China.”

This week’s Digital Trends explores the phenomenon known as the Long Tail, a phrase and now book by Wired editor-in-chief Chris Anderson, first published as an article for the magazine he heads in October 2004. Part 1 of Digital Trends explores the origins of the term and its mathematical origins in what the numbers guys’ call power law functions. It also mentions one of the more common rules of thumb, the 80/20 rule, which also behaves according to a power law function. Part 1 finishes by highlighting the main idea of the Long Tail, that it represents a new way of thinking about a number of markets, one rooted in opportunity by servicing the full range of options not just the few. Here in Part 2 we dig further into the mechanics of the Long Tail by elaborating on its components and trying to explain when tail opportunities exist.

Fig 1.1 – Graphical representation of the long tail, a power law function.

The Head: The items in this region constitute a) at worst, what Mr. Anderson describes as the lowest common denominator that we have had to accept (often due to real world constraints such as proximity and/or shelf space), and b) less unflatteringly, which is what we see in the search world, the limitation of our mind; it’s the keywords that we can think of given time constraints and a lack of tools. In closed systems such as a set advertiser base, the head would constitute 80% if not more of the volume, but if more open, liquid systems (stocks, keywords, etc.) the head constitutes often less than half.

The Tail: this section contains the full representation of supply; it would be all the movie titles made throughout the world, all the books, all the music, and with respect to paid search, all active keywords. As Mr. Anderson explains, these contain the non-hits that stores couldn’t carry but that given access, users prove they still desire. In some cases the tail can follow the Pareto principle making up a fifth, but in many cases it exceeds the sum of those items in the head.

Wired’s Chris Anderson provides three rules in his seminal Wired Magazine article extrapolated from the success of tail companies. Below is our more Internet advertising centric checklist for helping to identify the key components of a Tail economy.

• Large audience – while the tail speaks to the viability of niches, you still need volume. Without a large cumulative audience, the type of activity in the tail will not exist. DVD rental, music downloads, all qualify; whereas a niche like knitting would make up a group within a larger segment; it is not large enough to have its own Tail economics, at least not in scale. Netflix can leverage the tail because it can aggregate demand more efficiently than a retail location can. Its “stores” are its warehouses which no human sees, and instead of needing five, ten, twenty stores for a given region, it’s direct to consumer approach means it needs just one store for a major city, which means that it can fulfill component two, more choices.
• Lots of choices – the tail notion works because an actual tail exists. In an environment of only 100 choices, there is no tail. The more that number grows, the more likely the tail becomes. There isn’t a magic number, i.e. 1,000 or even 100,000. In the case of paid search, the inventory is words in any given language. This means, and it is the case, that search engines like Google actually stock millions of choices; it's why search makes for such a promising tail economy. It has a natural tail, an incredibly long list of choices, many of which receive little activity in a month. The more naturally occurring choices, the longer the tail, and the greater the opportunity, which means that breaking the 80/20 rule might be harder but more rewarding when it comes to search.   
• Search driven – requiring that user action be search driven suggests a slight departure from most discussions on the tail, and it could be argued that being search driven is optional, in describing a market/system/opportunity more examples than not seem to fit this. Search driven is the reason that Netflix can monetize the tail but broadcast television cannot. One matches user intent to the universe of options whereas the other tries to map a narrow band of options to user intent. Being search driven means allowing user intent to provide the starting point and a company trying to meet it rather than actively get customers for its selection. Companies can still do well with a narrow approach, but they won’t mine the tail, and their business could always come under pressure if it looks more like a tail economy, i.e. lots of buyers, lots of choices, and the next point, room for aggregation.
• Aggregation (Information asymmetry and facilitation) – one thing that makes tail economics, i.e. making more off the non-hits than the sum of the hits is a general lack of consumer awareness. In his article on the Long Tail, Mr. Anderson describes the case with Amazon and two books, one mainstream the other not. Its recommendation engine picked up that the more obscure book overlapped with the more popular one; users agreed and almost overnight, a non-seller became a source of revenue. The key takeaway is the existence of related intent, that items in the tail (in the exhaustive list of non-sellers / non-hits) map to those in the head group, in the big sellers. Amazon could make money without helping facilitate the mapping, but it makes much more because it does help map demand with supply. Doing that requires the final item, technology.
• Role of technology as facilitator – technology plays a crucial role in making Long Tail Economics possible. This takes place on two levels – the virtual store and the matching engine. Virtual stores can house more options than their physical “big box” counterparts. For example, Netflix can carry more titles than a local Blockbuster, and this is because a Netflix store has more customers than a local store; by having a virtual store, their coverage can be whole cities, and because their one to two business day approach means they need only a handful of “stores” to cover the whole US. Follow that logic and we see how each store can house thousands more titles. Secondly, companies that can service their customers without maintaining a retail store can start to play an active role in matching available supply with the demand. Amazon and Apple do this with their recommendation. Google, who has millions of titles, doesn’t have such an engine, which opens the door to specialists to figure out the relations, i.e. “mine the tail.” None of that would be possible, though, without technology.

In the end, talk of the long tail is about being able to make money from a wider range of products than was the case prior to doing business online. Long Tail Economics means leveraging technology to turn the classic 80/20 model on its head, about viewing opportunity as an iceberg. Embracing The Long Tail involves challenging us to rethink our notion of the fringe. The power is now in the 20 not the 80. With The Long Tail, 20 really equals 40, 50, or even 60. Long Tail Economics applies for many things, but it is not the magical solution to every business, only a way to view the world as having more choices and that having more choices can mean more money not less.

By Jay  Weintraub