$15 Billion.

"A trusted referral is the holy grail of advertising."

Opening Facebook up to everybody had been a huge success. By the fall of 2007 more than half the site"s users were outside the United States. The explosive international growth was a powerful sign of Facebook"s growing universal appeal, since the company had done nothing to make it easy for non-Americans to join. All the text remained completely in English, for one thing.

But the growth also presented a serious business problem. Facebook had to start figuring out how to make money-providing service to people all over the world was expensive. All the ads were aimed solely at Americans, which meant that Facebook was not generating any appreciable revenue from more than half its users. The advertising deal Facebook had signed with Microsoft a year earlier only applied inside the United States. If Facebook was going to take advantage of its newfound global presence, it needed a partner to help it sell display advertising internationally. Microsoft had made it plain it would love to be that partner, extending its U.S. deal to a global one.

Zuckerberg had always been blase about advertising . But Facebook now had 50 million active users and the platform had transformed it into an industry darling. The company had to find a way to pay for it all. Hundreds of thousands of new users were joining every week. And Facebook continued to build its infrastructure based on the a.s.sumption it would be much, much larger in the future. That meant spending millions of dollars for new servers. If he had to have ads, Zuckerberg hoped to develop a new kind that would work uniquely well on Facebook, ads that wouldn"t interfere with a user"s experience. The last thing Zuckerberg wanted was for it to feel like watching network television, where the show is routinely interrupted by irrelevant and inane advertising.

The U.S. ad deal gave Microsoft the exclusive right to sell banner advertising on Facebook. That had to change. Relying primarily on Microsoft for the lion"s share of revenue was precarious. Facebook needed its own self-managed streams of revenue.

Separately, Zuckerberg and Facebook"s board decided it was time to raise more money. Peter Thiel in particular wanted to do it that fall. Thiel has a refined nose for the twists and turns of the financial markets. Stocks were at levels not seen since the dot-com bubble and investors were feeling buoyant. Facebook"s reputation and growth had been transformed by f8 and the platform launch, and it was time to take advantage of enthusiasm among investors. But Thiel also knew that if they went out asking investors for money, someone might try to buy the entire company. To Thiel and Jim Breyer that was an appealing idea, but it horrified Zuckerberg (whose two unfilled seats on the board of directors remained in his control as a form of horror insurance).

The CEO asked Van Natta and his newly hired chief financial officer, Gideon Yu (who had been CFO at YouTube), to see what kind of interest they could drum up for a small stake in the company. Yu now says he thought Facebook might be able to get investors to buy stock at a valuation of about $4 billion. That would have been a huge leap. A little more than a year earlier, its third round A little more than a year earlier, its third round of financing (called Series C) had raised $27.5 million, valuing Facebook at $525 million. of financing (called Series C) had raised $27.5 million, valuing Facebook at $525 million.

But this company, as usual, didn"t conform to the usual expectations. Several venture capital and private equity firms were willing to buy a chunk of Facebook at a $10 billion valuation. That was eye-opening to Yu. He"d clearly been thinking too small. But Zuckerberg wasn"t satisfied. He thought the company was worth $20 billion, says another confidant. He and Van Natta decided to try for $15 billion. Sure enough, they found interest from several parties, but not enthusiasm. n.o.body would invest at this level without doing some serious negotiating on the terms. "We found where the market was," says Yu. "We were going to be able to get a deal done at $15 billion."

This was right about the same time that talks with Microsoft about an international ad deal began in earnest. The software giant also wanted to hold on to its U.S. deal with Facebook, but Microsoft executives felt they couldn"t let it stay the same. It needed to renegotiate the deal as much as Facebook did. Microsoft was losing about $3 million a month on the U.S. ads. It was putting the lion"s share of its banner ads on Facebook pages that displayed photos, but people just didn"t much notice ads in that environment. The price Microsoft thus could charge advertisers was low. Yet it had agreed to pay Facebook a guaranteed minimum amount-around 30 cents for every thousand page views, regardless of what it was getting from advertisers.

Everything Microsoft was doing in online advertising was fundamentally a response to Google"s growing power. Google search ads were garnering over half of all online advertising dollars, even as the increasingly profitable search giant was starting to toy with other sorts of software that directly competed with Microsoft"s core PC products. To fight back and defend its turf, Microsoft was now going head to head with Google across the board in online advertising. As part of this effort, it was investing billions of dollars to improve its own online search software. Separately, it had in May made its biggest purchase ever-paying $6 billion for aQuantive, which distributed advertising across the Internet. Now that it owned this distribution engine, it badly needed additional inventory to sell through it.

Microsoft CEO Steve Ballmer was fed up with losing deals to Google. He had recently lost both of the industry"s two biggest partnership opportunities after coming exquisitely close to agreement. Each time, Google swooped in at the last minute and stole the deal away. Ballmer had flown to New York in December 2005 to negotiate a major ad partnership with Time Warner"s AOL. He left town thinking he had a deal. Google unleashed its ad team headed by Tim Armstrong and came in with a better offer in days and sealed a contract with a $1 billion investment in AOL that valued it at $20 billion. Then in August 2006, Microsoft had a deal with News Corp."s Mys.p.a.ce and was poised to guarantee $1.15 billion, according to one of the deal negotiators. Google pounced at the final hour and won with a three-year guarantee to News Corp. totaling about $900 million. News Corp. apparently wanted the status of partnering with Google so much it was willing to give up revenue to do it. Microsoft had been further galled when Google swept up Internet banner advertising network DoubleClick for $3.1 billion in early 2007. This time, Ballmer was resolved it would not happen again.

Van Natta"s forte is deal-making. He coolly played Microsoft off against its archrival. He knew that uttering the word "Google" was like a magic spell to tame Microsoft"s normally rapacious negotiating instincts. And in fact when Google heard that Facebook was looking for a partner for its international ads, it began aggressively pursuing a deal itself.

On October 10, 2007, Google hosted its signature annual event for its best advertising clients-called Google Zeitgeist. Not only did the biggest marketers and ad agencies come to its campus for the two-day conference, but Google"s board of directors converged too for one of their quarterly meetings. It was a good time to make deals. Tim Armstrong, Google"s ad chief, had talked to Van Natta and knew that Microsoft was well along in talks to win Facebook"s international ad contract. But it just so happened that Mark Zuckerberg was one of the signature speakers at Zeitgeist. Armstrong talked to Google"s board members and got an official go-ahead to use this opportunity to begin serious negotiations with Facebook to try to take the deal away from Microsoft. The board even approved talks about buying Facebook, if it made sense.

Google didn"t make any secret of its interest in an ad deal with Facebook. At a press conference during Zeitgeist, Google CEO Eric Schmidt called social networking "a very real phenomenon." He added, "People don"t appreciate how many page views on the Internet are in social networks." It was an early expression of what would become a long-standing concern-Google cannot search content that is behind proprietary walls on the Internet. A close relationship with Facebook might achieve even more than acquiring a bunch of new ad s.p.a.ce. It could help Google stay dominant as the Internet evolved.

That evening everybody was bused to a nearby park where Google had erected a gigantic white tent. After a lengthy c.o.c.ktail hour, all 250 or so Zeitgeist attendees sat down to a lavish, almost Baccha.n.a.lian feast. The first course was served on thick plates made of ice. Google was at the height of its powers-money was flowing in like manna. Here the company was thanking the people who were spending those billions on advertising while simultaneously proclaiming itself to be rich rich rich. At the center table, immersed in intense conversation, were Google co-founder Larry Page, Armstrong, deal expert Megan Smith, as well as Zuckerberg, Van Natta, and Facebook corporate development boss Dan Rose. The Facebook team said they were far along in negotiations with Microsoft. Armstrong impressed upon the Facebook executives that Google was serious about wanting to do the deal instead.

After the lavish dinner ended at around 10 P.M. P.M., Facebook"s trio and the Google executives retired to the company"s nearby headquarters building for some serious negotiating. They worked late into the night until they had the rough outlines of a deal. Google would take over both the U.S. and international ad deals. It also agreed to consider making a small investment in Facebook at the $15 billion level. For Google it made sense to buy the equity as a sweetener because Facebook would have to go through quite a bit of trouble to dislodge Microsoft. If Facebook pushed Microsoft out of the still-in-effect U.S. deal it was likely to result in legal unpleasantness. But Google had gone further. Executives told Zuckerberg that they were willing to consider buying Facebook outright, though at a price considerably less than $15 billion. This time, however, Zuckerberg was firm. Facebook was not for sale.

Even on the ad agreement, many on the Google side detected a lack of commitment on Zuckerberg"s part. They noticed he kept pushing for very specific concessions on things like the size and shape of display ads-the kind of thing usually left to underlings to iron out. It seemed to them he might be seeking specific promises from Google in order to strong-arm Microsoft to concede the same points. For all the talk, the Google team knew that Microsoft"s prior relationship with Facebook gave it a big advantage. The chances of pulling Facebook away remained small.

Microsoft had been carefully cultivating Zuckerberg. CEO Steve Ballmer had flown to Palo Alto to visit his young counterpart twice. Ray Ozzie, Microsoft"s Chief Software Architect, had also repeatedly visited Palo Alto. As Zuckerberg is wont to do, he took them on long walks. He told Ballmer that Facebook was raising money at a $15 billion valuation.

But Ballmer had come with something very specific in mind. "Why don"t we just buy you for $15 billion?" he replied, according to a very knowledgeable source. Zuckerberg, as usual, was unimpressed, even by this fabulously extravagant offer. It was so high that, had it been accepted, Microsoft"s shareholders might have raised serious obstacles to its completion.

"I don"t want to sell the company unless I can keep control," said Zuckerberg, as he always did in such situations. He knew that keeping control once Facebook sold would be almost impossible, so for him this was effectively a way to end the conversation.

Ballmer took this reply as a sort of challenge. He was emphatic that Microsoft wanted to buy Facebook. He went back to Microsoft"s Redmond headquarters and concocted a complicated plan intended to begin a process of acquisition without compromising Zuckerberg"s ability to call the shots. According to people close to the situation, Ballmer proposed that Microsoft acquire a minority stake in Facebook at a $15 billion valuation. Then, in a provision loosely modeled on a deal arrived at almost two decades earlier between giant Swiss pharmaceutical firm Hoffman-LaRoche and Silicon Valley biotech star Genentech, Microsoft would have the option, every six months, to buy another 5 percent of Facebook. A complete takeover of the company would take 5 to 7 years, depending on how much of the company Microsoft bought at the outset. The price Microsoft was obligated to pay would rise steadily over time, making Facebook"s ultimate price considerably higher than $15 billion. But from Ballmer"s point of view, the deal addressed Zuckerberg"s key concern-it allowed him to retain control, at least for another few years.

Ballmer flew down to San Francisco again and brought along Kevin Johnson, who oversaw all Microsoft"s ad-related business. Van Natta suggested that they all meet at his Palo Alto home in order to avoid attracting attention. Ballmer didn"t make that any easier. He arrived in a big black Cadillac Escalade with a contingent of security personnel wearing earpieces and microphones. As the security men cased the yard, Van Natta delivered bad news to Ballmer and Johnson, while Zuckerberg sat quietly and Microsoft software chief Ray Ozzie listened in via speakerphone.

Facebook wanted to alter the U.S. deal, Van Natta declared. In fact, it was going to start selling its own ads soon, whether Microsoft liked it or not. Ballmer was nonplussed. If Microsoft wanted the international deal, Van Natta continued, it had to agree to concessions on the U.S. one. Facebook needed to try out some new ad formats of its own. If Microsoft wouldn"t agree, well, Google was waiting in the wings.

The encounter of Van Natta and Ballmer must have been a sight to see. Van Natta may be fearless, a belligerent and uncompromising negotiator, but Ballmer is big, loud, and consummately forceful himself. You don"t screw around with him. Not to mention he"s CEO of what is in financial terms still the most powerful technology company in the world. Van Natta must have had an exquisite sense of just how far to push, because Ballmer didn"t lose his cool. He reiterated that Microsoft had no interest in reopening the U.S. ad deal. What he was really interested in, he said, was buying Facebook.

Zuckerberg was cautious. The young CEO had learned his lesson the year before with Yahoo-once you open the door to a possible sale it"s hard to close it. Zuckerberg made it clear he was inclined not to sell, but suggested that Microsoft would have to agree to a raft of conditions, including even more autonomy for Facebook than the tiered proposal antic.i.p.ated, with Zuckerberg remaining at the helm indefinitely. Says someone from Microsoft who heard about what happened, referring to Zuckerberg: "It wasn"t "If you pay $X billion we"ll do it." The guy"s not a seller. His expectations were too high." Despite all the work Ballmer had put into his acquisition proposal, Zuckerberg wouldn"t bite.

While Microsoft was almost desperately seeking the international deal, Facebook took advantage of the software giant"s pliability to resolve another dispute, a problem with Hotmail, Microsoft"s free Internet email service. The biggest tool for Facebook"s growth was the contact importer it had launched at the time of open registration. New users entered their email username and pa.s.sword, and Facebook helped them send anyone on their email list an invitation to join. Hotmail was the largest source by far of such user referrals. But it interpreted many email invitations coming in from Facebook as spam-unwanted commercial messages. On some days Hotmail blocked the use of the contact importer altogether. Facebook"s user growth then dropped as much as 70 percent, says Moskovitz. So in the midst of the ad talks, uber-negotiator Van Natta, Moskovitz, and D"Angelo trooped up to Microsoft headquarters in Redmond, Washington, to iron out the conflict. "This was absolutely not something we could walk away from," says Moskovitz. After a day or so of talks, Van Natta got Microsoft to stop interfering with the imports even though Facebook conceded almost nothing in return.

In a cla.s.sic brinksmanship move, Microsoft"s executives told Van Natta they wouldn"t budge on letting Facebook sell some of the U.S. ads. Van Natta refused to release Microsoft from its minimum payments for ads next to photos. Microsoft ad chief Johnson replied that keeping the U.S. inventory was critically important to him. In order to keep it, he said, he was willing to lose the international deal. "Fine," said Van Natta. "We"re going with Google."

Johnson returned to Microsoft"s headquarters in Redmond, Washington. But Hank Vigil, the company"s top dealmaker, stayed in Palo Alto to continue talks with Van Natta. He made a breakthrough. Van Natta had told the Microsoft team that his job depended on a successful outcome of the talks. Now he waffled. Vigil proposed that Microsoft would let Facebook use 15 percent of the U.S. ad inventory if it made certain concessions. On Sat.u.r.day morning Vigil set up a conference call with Kevin Johnson and Van Natta to expand on the offer. Johnson said he"d agree to release the inventory if Microsoft got the international deal and if Facebook eliminated the photos ad minimums and agreed to use Microsoft"s search engine inside Facebook. Johnson instructed Vigil to take his team to Facebook"s offices on Monday morning and not leave until he had a deal.

By 11 A.M A.M Monday, all the players were ensconced in a conference room on the second floor of Facebook"s office on University Avenue in Palo Alto. Sitting at a big gla.s.s table, with sunshine streaming in through the walls of gla.s.s, the two companies" teams conducted the top-secret negotiation that would transform Facebook"s reputation. In another Facebook building, a smaller group from Google met for part of this time, discussing its own possible deal. Monday, all the players were ensconced in a conference room on the second floor of Facebook"s office on University Avenue in Palo Alto. Sitting at a big gla.s.s table, with sunshine streaming in through the walls of gla.s.s, the two companies" teams conducted the top-secret negotiation that would transform Facebook"s reputation. In another Facebook building, a smaller group from Google met for part of this time, discussing its own possible deal.

For the next twelve hours the Microsoft and Facebook teams went back and forth on issues large and small. Microsoft got an agreement to move toward providing search technology inside Facebook, yet another blow in its near-feudal joust with Google. Facebook demanded Microsoft not display ad banners either at the top of the screen or on the lower left, just on the lower right side. (Google"s acquiescence on this point gave Facebook ammunition.) Unlike the year-earlier U.S. deal, there would no longer be any up-front guarantee of how many ads Microsoft would display nor how much it had to pay Facebook. The two companies would instead share the revenues from any ads sold. Facebook successfully pushed for a higher percentage than is usual in such deals. And the social network got its all-important flexibility to experiment and innovate on new ad formats on 15 percent of the U.S. display ads.

The Facebook team-Van Natta, Rose, Yu, and General Counsel Rudy Gadre-occasionally ducked around the corner to huddle with Zuckerberg, whose desk was only steps away. The CEO was much more involved than in past such negotiations. Whenever things bogged down, Van Natta loosened Microsoft up again with a vague allusion to Google. He implied, but didn"t exactly say, that Google was ready to do all the things Microsoft wouldn"t. It was close to true, anyway.

At about 11 P.M. P.M. it was apparent a deal was in sight, though many details remained to be resolved. Everyone"s energy was flagging. Just then, a blast of thumping house music invaded the quiet conference room. Several negotiators stepped out into the loftlike office to see what was going on. They found a Facebook programmer at a DJ stand, with the music turned all the way up. This was the signal to Facebook"s engineers that a hackathon was about to begin. These were the all-night sessions, legendary in Facebook"s engineering culture, when many of the site"s most interesting innovations emerged. Unlike a typical hackathon, though, this so-called "convertathon" had a specific agenda-to change Facebook"s underlying software code to make it easier to translate into languages other than English. Translation of the site was set to begin in a few months in order to further bolster already torrid international growth. it was apparent a deal was in sight, though many details remained to be resolved. Everyone"s energy was flagging. Just then, a blast of thumping house music invaded the quiet conference room. Several negotiators stepped out into the loftlike office to see what was going on. They found a Facebook programmer at a DJ stand, with the music turned all the way up. This was the signal to Facebook"s engineers that a hackathon was about to begin. These were the all-night sessions, legendary in Facebook"s engineering culture, when many of the site"s most interesting innovations emerged. Unlike a typical hackathon, though, this so-called "convertathon" had a specific agenda-to change Facebook"s underlying software code to make it easier to translate into languages other than English. Translation of the site was set to begin in a few months in order to further bolster already torrid international growth.

Back in the conference room, heads began to bob and feet to tap. It was ludicrous, but energizing. A Microsoft negotiator got up and stood cordially in line with Facebook programmers, waiting to dig into the bins of take-out Chinese food. Negotiations picked back up once the music was turned down. By 3 A.M A.M, they had a deal. Essentially Facebook got everything it wanted. The negotiators left the engineers to their labors and went to bed.

The issue of an accompanying investment had not been a topic during the gla.s.s-room negotiations, but the next morning, Van Natta raised it bluntly. Says his then-deputy Rose: "We said to them, "Look, if you want to use the investment opportunity to cement the relationship, we want you to lead the round." We said, "We might be talking to your compet.i.tors."" Microsoft had continued to make clear that if Facebook was willing to sell, it was interested in buying. But Zuckerberg was not about to sell. Van Natta was prodding Microsoft to buy a small chunk. Ballmer had in effect already agreed that Facebook was worth $15 billion, so the valuation wasn"t much in dispute. At that astronomical level even a small percentage of the company would net Facebook many millions that it could use to underwrite its money-losing operations. The Hong Kong billionaire Li Ka-shing, often called "Asia"s Warren Buffett," had earlier approached Facebook to invest and had been negotiating hard during these very same days and had agreed to invest at that valuation. "It was a frenzied period," says Yu. Everybody was acting as if Facebook could become a financial colossus, even though at the moment the only thing huge about it was its membership growth rate.

After some frenzied back-and-forth occurring mostly over the course of a single day, Microsoft agreed to invest $240 million at a $15 billion valuation for 1.6 percent of Facebook, alongside Li Ka-shing, who would put in $60 million for 0.4 percent. Microsoft"s executives were happy. "It was all about the search war with Google," says one. "A $240 million investment that helped us fight them was definitely worth it." The pressure was great to conclude the deal, so Microsoft had little time for any detailed financial due diligence. But it was critical to Microsoft that another investor partic.i.p.ate alongside them in the round. Microsoft had to be able to demonstrate it wasn"t paying an inflated price in order to achieve an ad deal. Otherwise, were Facebook later to be determined to be worth less than $15 billion, accounting rules would require Microsoft to write down as a loss on its books the proportional difference between $15 billion and the actual valuation. So Li"s stake, while small, was crucial.

Microsoft did not get a particularly attractive deal in legal terms, either. In order to move quickly for this so-called Series D round, it agreed to abide by the same doc.u.ments that had applied to investors in the Series C when several VCs had invested in mid-2006. The convertible preferred shares it bought had what is called a "1X nonpartic.i.p.ation liquidation preference," which means if Facebook were ever sold outright, Microsoft would get back either its actual cash outlay of $240 million or 1.6 percent of the purchase price, whichever was larger. But it could do nothing to stop a subsequent investment round at a lower valuation. If there is eventually a public offering of Facebook"s stock, Microsoft will be forced to convert its preferred shares into common stock proportionate to its ownership, no matter how much the company is then worth, whether more or less than $15 billion. Microsoft was willing to accept all these conditions because its primary goal was the completion of the advertising deal. But at the last minute it demanded one important condition: Facebook could not take any investment money from Google. And if it ever contemplated an outright sale to the search nemesis, Microsoft had to get advance notice.

The deal was announced on Wednesday, October 24, and prompted an outcry of amazement. The Wall Street Journal Wall Street Journal called Facebook "the newest Internet darling" and said the deal was "reminiscent of the Internet bubble that ended in 2000." The called Facebook "the newest Internet darling" and said the deal was "reminiscent of the Internet bubble that ended in 2000." The Los Angeles Times Los Angeles Times called the $15 billion figure "staggering." "It tips the scales in terms of totally ridiculous valuations," wrote the influential TechDirt blog. This was by far the highest valuation ever given to a private technology company, and one with no profits to boot! Either Microsoft"s Steve Ballmer was insane, or Facebook mattered more than anyone had realized. But if the f8 platform event five months earlier had firmly put Facebook once and for all onto the technology industry map, this investment did the same thing for Facebook on Wall Street. Microsoft"s stock jumped markedly. The ad deal that precipitated the investment was barely noticed in the hubbub over the valuation. called the $15 billion figure "staggering." "It tips the scales in terms of totally ridiculous valuations," wrote the influential TechDirt blog. This was by far the highest valuation ever given to a private technology company, and one with no profits to boot! Either Microsoft"s Steve Ballmer was insane, or Facebook mattered more than anyone had realized. But if the f8 platform event five months earlier had firmly put Facebook once and for all onto the technology industry map, this investment did the same thing for Facebook on Wall Street. Microsoft"s stock jumped markedly. The ad deal that precipitated the investment was barely noticed in the hubbub over the valuation.

Facebook"s timing on the deal could not have been better. Only two weeks earlier, the stock market peaked at a level it has not approached since. In 2008 the world fell into the worst recession of the postwar period. But Zuckerberg had a lot of crucial cash in hand to help him through the down times. In addition to the initial $300 million it raised in Series D, Li Ka-shing invested an additional $60 million several months later, and three Munich-based venture capitalists, the Samwer brothers, invested $15 million around the same time, bringing the total raised in the Series D round to $375 million. Zuckerberg has a simple explanation for how Facebook achieved such an amazing financial result. "Peter [Thiel] helped us time it," he says simply. "He was like "Now would be a good time to raise money.""

Now that Microsoft was no longer an obstacle to Facebook selling ads on its own site, Zuckerberg and company lost no time launching a new sort of ad on the service. Only two weeks after closing the Microsoft deal, Facebook hosted its first-ever big event for the advertising community in New York on November 6. The announcement had several parts. Any commercial ent.i.ty could now create a "page" on Facebook for free, which would have many of the characteristics of an individual"s profile, including the ability to host applications. The "sponsored page" model had outlived its usefulness. The company"s strategy was to get as many companies into its system as possible, on the presumption that once they were operating there they would find cause to advertise or otherwise spend money, even if their page itself were free.

A user could become a "fan" of one of these pages rather than a "friend," as they did with an individual. Activities of users on these new commercial pages would be broadcast to their Facebook friends" News Feeds. (I soon became a fan of the New York Times New York Times page, for example, and my friends saw an announcement of that in their News Feeds.) Barely mentioned was the fact that a service called Beacon would also enable forty-four companies, and more later, to extend a similar alert system onto their external websites. Activity on these external websites could go into friends" Facebook News Feeds as well. page, for example, and my friends saw an announcement of that in their News Feeds.) Barely mentioned was the fact that a service called Beacon would also enable forty-four companies, and more later, to extend a similar alert system onto their external websites. Activity on these external websites could go into friends" Facebook News Feeds as well.

The meat of the Facebook Ads announcement, at least in the minds of those who planned it, was that Facebook would launch a new kind of self-service advertising that enabled any company, even a tiny one or an individual, to go online and design and purchase an ad on Facebook that they could target very exactly to their intended audience. In effect, the kind of custom targeting that Moskovitz had pioneered three years earlier-for example, when Interscope Records targeted ads for Gwen Stefani"s "Hollaback Girl" to cheerleaders-was now coming to the ma.s.s market. One a.s.sumption was that the owners of the new pages would be heavy users of these ads as a way to promote their presence on Facebook. Another component of the new self-service ads was what the company called "Social Ads," which would pair a paid commercial message with a Facebook user"s endors.e.m.e.nt.

At the announcement Zuckerberg attracted considerable attention-and derision-with his grandiose introduction. It was the first time he had ever given a big promotional talk outside the confines of Silicon Valley. Hearing him speak, you might think he had gone almost overnight from despising ads to wanting to own the worldwide ad industry. "Once every hundred years," he began, "media changes. The last hundred years have been defined by the ma.s.s media. In the next hundred years, information won"t be just pushed out to people. It will be shared among the millions of connections people have... Nothing influences people more than a recommendation from a trusted friend... A trusted referral is the Holy Grail of advertising." Unfortunately, since Facebook"s original intention was to promote the new self-service ads, which were more about making targeting available to the ma.s.ses than about trusted recommendations, Zuckerberg"s intro was misleading to begin with.

The feature that would come to define Facebook Ads, and to turn November 6 into a day of infamy, was Beacon and the way it worked outside Facebook"s walls. Beacon was a poorly designed alert service. It wasn"t even an advertising product, since it generated no revenue. It was built by Facebook"s platform team, not the ad group. But while it was intended for activities like playing a game or adding a recipe to an online recipe box, it also could be used to announce purchases you made on partner sites. And Facebook had lined up a bunch of commercial partners for it. If you, say, rented a movie at Netflix, bought a pair of shoes at Zappos.com or a movie ticket on Fandango, you could give the website permission to broadcast that fact to your friends back in Facebook via an item into your News Feed. But Beacon was a last-minute add-on to the Ads launch and had barely been tested with users. Its implications were overlooked by Zuckerberg and his executives in days leading up to the launch.

And it had a major design flaw. When you, say, bought your shoes at Zappos, you weren"t asked to explicitly approve sending that fact to your friends inside Facebook. Instead, you were shown a little drop-down menu that asked if you wanted to not send the information. If you didn"t proactively stop the alert, it would proceed. In Web lingo, that"s called "opt-out" rather than "opt-in." And the opt-out menu only displayed for a few seconds before disappearing. Many users seemed to miss it altogether.

After the launch, stories began emerging in the press of users who had unintentionally launched word of their commercial actions back into Facebook with unfortunate consequences. One Ma.s.sachusetts man bought a ring, whereupon an item appeared in his wife"s News Feed: "Sean Lane bought 14k White Gold 1/5 ct Diamond Eternity Flower Ring from overstock.com." Within two hours Lane"s surprised wife, Shannon, sent him an instant message: "Who is this ring for?" In fact it was to be her surprise In fact it was to be her surprise Christmas present, according to a story in the Christmas present, according to a story in the Washington Post. Washington Post. Lane told the Lane told the Post Post that he was "crestfallen" that his surprise was spoiled (and also possibly that Shannon"s News Feed item linked to an Overstock Web page showing he got 51 percent off). Another relationship was disrupted when a New York man"s girlfriend saw he had purchased a ticket on Fandango to a movie he was scheduled to see with her the following week. A number of users who did lots of shopping at Beacon-affiliated sites found that the entire contents of their Christmas gift list had been published to friends in Facebook. that he was "crestfallen" that his surprise was spoiled (and also possibly that Shannon"s News Feed item linked to an Overstock Web page showing he got 51 percent off). Another relationship was disrupted when a New York man"s girlfriend saw he had purchased a ticket on Fandango to a movie he was scheduled to see with her the following week. A number of users who did lots of shopping at Beacon-affiliated sites found that the entire contents of their Christmas gift list had been published to friends in Facebook.

Beacon felt invasive, and misused personal information. It seemed to many that Facebook wanted to hijack data about its users in order to make a buck. After things started going wrong, many in the press looked back to Zuckerberg"s hubris at the introduction as a sort of explanation-Facebook was all about power, and Zuckerberg didn"t care what happened to his users. This was a fundamental misreading of the young CEO, but Facebook had become so large so fast that journalists were only beginning to understand it.

The backlash built quickly. As with any Facebook controversy, the viral distribution tools of Facebook itself were well used against it. The liberal political group MoveOn.org stepped in to lead the Beacon protest. It took out ads on Facebook (using the new self-service tool) that asked, "Is Facebook Invading Your Privacy?" It invited users to join a protest group, and 68,000 did. In reality the percentage of users protesting was relatively tiny-0.1 percent versus over 10 percent at the height of the News Feed fracas. But MoveOn got a lot of attention. It and other activist groups were also filing formal complaints with the Federal Trade Commission. Some were preparing lawsuits.

And now whatever happened at Facebook was big news. It had 57 million users and Microsoft"s money behind it. The press wanted Zuckerberg to apologize and turn off Beacon. Many writers argued that Facebook"s stunning new valuation had made it suddenly, desperately eager to prove it could be profitable. One story that indicates how far Facebook"s image had fallen was written by Fortune Fortune"s Josh Quittner. t.i.tled "RIP Facebook?" it argued the company was "coming undone." Quittner compared twenty-three-year-old Zuckerberg"s rash decision making Quittner compared twenty-three-year-old Zuckerberg"s rash decision making in the Beacon episode to "watching an unattended child play with a pack of matches in a wooden house." in the Beacon episode to "watching an unattended child play with a pack of matches in a wooden house."

Beacon was the worst and most damaging controversy Facebook has ever faced, for several reasons. First, unlike with the News Feed, the company made a serious product design mistake. Beacon really did result in data being misused. It thus violated Zuckerberg"s principles about the importance of privacy and user control of information. But the damage was compounded because for more than three weeks Zuckerberg did nothing to respond to the complaints. As his silence continued, the controversy grew angrier. He was watching user statistics as he always did and saw that Beacon was not affecting behavior inside the service. But that fact belied the genuinely painful experiences of a small number of users as well as the legitimate outrage of the press.

There is an edge of sanctimony to Zuckerberg that at times like this can serve him poorly. But the irony was that he had endlessly resisted, up until now, anything that resembled an intrusive ad or message on Facebook. Here was someone who had said for years that he wanted to do what best served his users, now suddenly acting as if he knew better than they did.

Now, in retrospect, Zuckerberg acknowledges he"d become c.o.c.ky. "We didn"t react quickly enough," he says ruefully, "because we were just so used to people complaining about things and then us eventually being right. We were like "Hey, whatever, they"ll eventually get over it." Then it was like "Hey, no, we actually messed this one up.""

Finally, on November 29, more than three weeks after Beacon debuted, Facebook redesigned it to be a fully opt-in system. No message about you would now be sent without your explicit permission. MoveOn issued a cautious victory statement. Then a week later, Zuckerberg made his first public statement on the mess, with a deeply contrite blog post on Facebook"s site t.i.tled "Thoughts on Beacon."

"We"ve made a lot of mistakes building this feature, but we"ve made even more with how we"ve handled them," he began. "We simply did a bad job with this release, and I apologize for it... We took too long to decide on the right solution... Facebook has succeeded so far in part because it gives people control over what and how they share information... In order to be a good feature, Beacon also needs to do the same." He also announced that Facebook was now making it possible to turn off Beacon completely, something MoveOn had asked for.

Beacon put a black eye on Facebook that still hasn"t fully disappeared. Internally, the Facebook Ads launch was dubbed beforehand with the project name "Pandemic." It really did turn into a disease that was hard to wipe out-a disease of negative perception that lingered long after the product was modified. Membership growth slowed discernibly in the aftermath of all the negative press coverage, though it picked up again by early 2008. Dan Rose, the former Amazon executive who heads marketing for Facebook"s advertising efforts and was deeply involved in the Facebook Ads launch, says the controversy was "devastating" for the company. "By the time we fixed Beacon, the meme was already out there that people didn"t control the way their information flowed," he says. "We just really screwed it up. It took a long time for our brand to grow past that."

But Beacon did illuminate where Facebook hoped to go in the future-to become a social hub where information about your behavior across the Web was aggregated for friends to see. If you buy something or make a comment on a blog or indicate you like something, Facebook"s goal is that eventually it should be possible to let your friends in Facebook know that. In fact, the Beacon program didn"t end until late 2009, along with the settlement of a lawsuit about it. Meanwhile, objections to it had died away. Zuckerberg now says he may have simply launched Beacon too soon: "One of the things that was bad about Beacon was that people just weren"t ready yet to share their information off of Facebook." The company in 2008 launched a much more widely deployed technology called Facebook Connect for people to share what they do at partner websites. There has been virtually no protest, largely because Connect gives users sufficient control over the information about them that it sends to their friends.

Shortly after the Beacon frenzy died down, board member Jim Breyer had a stern conversation with Zuckerberg. "We blew it," he said. "We should have apologized right away. This, to me, Mark, is an example of why it"s so important for us to get a new chief operating officer into the company." Owen Van Natta was a great deal guy, unsurpa.s.sed at business development, but not the firm, steady, second hand on the tiller that Breyer thought a company with a still-learning twenty-three-year-old CEO needed. And the company also needed someone well-versed in the complexities of the online advertising business. Zuckerberg took a couple of weeks to think about it, then told Breyer he agreed. He would tell Van Natta of their decision in early January and begin a search.

At a Christmas party in mid-December, Zuckerberg got into a conversation with Sheryl Sandberg. She was a senior executive at Google who had built the search company"s self-service ad business into one of the economic powerhouses of the Web. The two of them ended up standing in a corner for over an hour as Zuckerberg queried her about how to manage a growing tech organization. They agreed to get together sometime for dinner.

In the meantime, after a difficult conversation with Van Natta, Zuckerberg began meeting potential candidates for the COO job. One of them was Dan Rosensweig, the former chief operating officer at Yahoo who had only a little more than a year earlier avidly pursued the purchase of Facebook with Yahoo CEO Terry Semel, and who with his wife had hosted the party where Zuckerberg met Sandberg. Another was Jeff Weiner, another top Yahoo executive widely known for his judgment and managerial smarts.

Sandberg, who had been at Google since 2001 and made many millions from her stock options there, had decided she was ready to leave. She had already been offered a great job at a big East Coast media company, and was mulling it seriously. She spent an afternoon talking to Roger McNamee, an industry sage and one of Silicon Valley"s best-known investors. She wanted his advice about the media job. "It"s a really good idea. You should do this," McNamee told her, then hesitated. "But what you really should do is go work with Mark Zuckerberg at Facebook." McNamee had been informally advising Zuckerberg and knew he was looking for a new COO. Sandberg and Zuckerberg had coincidentally been emailing about a dinner the following week. Sandberg hadn"t thought of it as a recruiting dinner, though Zuckerberg"s aide-de-camp and in-house recruiter Cohler had talked to her repeatedly for over a year about Facebook. "When are you coming to work with us?" he asked every time he saw her.

By the time Sandberg arrived at the quiet little Silicon Valley restaurant, McNamee had spoken to Zuckerberg and made a case for Sandberg. At dinner the two talked and talked. The restaurant closed at 10 P.M. P.M., then Zuckerberg went back to Sandberg"s home to keep talking. She"s the mother of two small children and she usually goes to bed by 9:30 P.M P.M. By midnight she had to kick him out so she could sleep.

Dinners like that continued. Zuckerberg was in no hurry. He wanted to get to know this person whom he might be working with for the next ten to twenty years. This time he wanted to hire somebody for the long haul. Sandberg says the meetings with Zuckerberg, which he estimates took a total of fifty hours, were "endless." "He never left!" she says in an interview. "Put that in your book. He just would not leave my house."

Sandberg is an elegant, slightly hyper, light-spirited forty-year-old with a round face whose bobbed black hair reaches just past her shoulders. Prior to her six years at Google she served in the powerful role of chief of staff to Lawrence Summers when he was secretary of the Treasury in the Clinton administration. She met him as a student at Harvard-yes, that school again-where she majored in economics. She wrote her thesis on the economic factors that lead women to remain in situations where they are abused by their husbands. (Zuckerberg has always sought to hire academic stars, despite having dropped out himself.) She speaks at an astonishingly rapid pace, yet without omitting inflection, in a kind of musical torrent of words. She"s stylishly dressed when I come to see her, in new knee-high black Prada boots with black slacks and a cashmere sweater, and her polish contrasts dramatically with Zuckerberg"s plainness-and with just about everybody else at Facebook.

Sandberg was eager to keep these meetings secret from others in the close-knit Silicon Valley community. One time she and her husband, Dave Goldberg, a top Yahoo employee, joined Zuckerberg and his girlfriend, Priscilla Chan, for dinner at an obscure restaurant near the San Francisco airport where no one would recognize them.

Zuckerberg asked a lot of questions, and Sandberg replied in kind. The topics they covered in their discussions ranged from where Facebook would be in five years to Sandberg"s experiences in government to theories of management to personal history. He was vetting her, but she needed to be convinced as well. Shortly before they started meeting, the Harvard-oriented magazine 02138 02138 had published a long expose about the convoluted campus origins of Facebook. It accepted the Winkelvoss arguments at face value and suggested Zuckerberg was probably an intellectual thief. "We"ll never know what really happened in the Harvard dorms four years ago," the article concluded. "The question remains: Whose idea was it?" Sandberg was concerned when she read it and queried her friend McNamee, who a.s.sured her of Zuckerberg"s honesty. had published a long expose about the convoluted campus origins of Facebook. It accepted the Winkelvoss arguments at face value and suggested Zuckerberg was probably an intellectual thief. "We"ll never know what really happened in the Harvard dorms four years ago," the article concluded. "The question remains: Whose idea was it?" Sandberg was concerned when she read it and queried her friend McNamee, who a.s.sured her of Zuckerberg"s honesty.

Late in January both of them were heading to the World Economic Forum in Davos, Switzerland. Sandberg invited Zuckerberg to join her for the flight from San Francisco to Zurich on Google One, Google One, as the 767 owned by its co-founders Larry Page and Sergei Brin is known. The two talked conspiratorially the entire flight, a fact not unnoticed by some of her Google colleagues. as the 767 owned by its co-founders Larry Page and Sergei Brin is known. The two talked conspiratorially the entire flight, a fact not unnoticed by some of her Google colleagues.

As the discussions got more serious, Sandberg called her good friend Don Graham at the Washington Post Co. to ask his opinion of Zuckerberg and Facebook. Graham had been among the many who tried to hire her in 2000 when she left the Treasury Department. (Others included the New York Times Company and the nonprofit AIDS Vaccine Initiative.) It turned out that Zuckerberg had also called Graham to ask about her. The Post CEO gave them both strong endors.e.m.e.nts. "What a sensational hire that was. Wow," says Graham now about Sandberg at Facebook.

Jim Breyer also spoke at length with Sandberg and with others who were in contention for the COO job. She was one of the few who didn"t say, one way or another, that she would like to keep open the possibility of being Facebook"s CEO some day. That was a deal breaker. "Mark"s our long-term CEO," says board member Breyer. "We were looking for a great business partner who was comfortable with that."

Aside from her willingness to be number two, the Harvard connection, the Graham connection, her role developing Google"s ad business, and her experience as a manager, there was an additional thing that Zuckerberg found intriguing about Sandberg. "We spent a lot of time talking about her experience in government," he says. "In a lot of ways Facebook is more like a government than a traditional company. We have this large community of people, and more than other technology companies we"re really setting policies." Beacon, of course, was an example of very poor policy setting.

He hired her, and she started at Facebook at the end of March 2008. If Microsoft"s investment proclaimed to the world that Facebook was a formidable economic force, hiring this Internet superstar declared it would be a well-managed one.

For all the vetting and planning, on the day that Sandberg arrived she had some trepidation. What would it really be like working for this twenty-three-year-old day in and day out? On that first day, Sandberg, Zuckerberg, and the so-called M team-the group of eight or so most senior executives-were discussing a ratings system that was going to be used in human resources. The question arose: How does one best set up a ratings system? Sandberg had overseen many ratings systems at Google, so she spoke up. "You always have five categories: two on the top, two on the bottom, and one in the middle," she said briskly. Someone asked her why. "Well, three is too few, seven is too many, and six is an even number. You need one in the middle to anchor it. Everyone understands five categories," she a.s.serted. Shortly thereafter the meeting ended. Zuckerberg walked out alongside Sandberg.

"I"m really sorry," he said.

"For what?"

"Well, I rolled my eyes."

"I didn"t even notice."

"Well," Zuckerberg said, "I"m bringing you in here and I know I need to empower you and make sure everyone knows I believe in you, and I shouldn"t be rolling my eyes."

She was impressed Zuckerberg would call himself out for such a minor infraction. "I said to myself, "This is going to work,"" she recalls. And the candid back-and-forth between them has continued. They meet privately several times a week. For the first few minutes of every Friday"s meeting they give each other direct feedback. Before Sandberg started she told Zuckerberg she wanted regular feedback from him. But he insisted it should go both ways.

From the moment she arrived, Sandberg was the company"s top advertising champion and salesperson. She had immense experience with advertisers from Google and a deep appreciation of the importance and potential of ads on the Net. According to some at Facebook, in her first weeks there was hardly anyone else at the company about whom the same could be said. Despite the Microsoft deal, despite Facebook Ads, despite the clear need to build up revenue as the service burgeoned, there remained a profound corporate ambivalence toward advertising as the means for Facebook to become a real business. That ambivalence was rooted in the CEO, who firmly believed that the product and the user experience comes first. Sandberg had some work to do.

13.

Making Money.

"What business are we in?"

How would Facebook turn its social success into a lasting, moneymaking business? It was a question that could elicit a surprisingly broad range of answers even among senior executives at Facebook when Sheryl Sandberg arrived. Zuckerberg didn"t have a good answer, though that didn"t bother him much. But Sandberg, who is a very methodical manager, was intent on creating alignment among Facebook"s leadership. She had come to the company to turn it into an advertising powerhouse. She needed all her staff and peers on the M team to work in synch. There was no question in her mind that Facebook represented one of the great advertising environments of all time.

The matter was hardly academic, because Facebook needed the money. It was burning through the $375 million it had raised from Microsoft, Li Ka-shing, and the Samwer brothers faster than anybody had expected. Some of Zuckerberg"s allies in management had already concluded it had been an error not to accept a lower valuation, which would have allowed Facebook to raise a lot more money because so many more investors would have been willing to buy. The company had been hiring quickly and was by now paying about five hundred employees and adding servers to its data centers by the hundreds. Soon Facebook would also have to build new data centers outside the United States to accommodate its international growth. It had built a fancy new cafeteria for employees in a separate building a block or so from its main buildings, with chefs hired from Google and fabulous food-all served free. Plans were afoot to move out of the twelve buildings in which staff was now scattered throughout downtown Palo Alto and move into one big new s.p.a.ce.

After Sandberg had been at the company about five weeks, she decided to host a series of meetings to get Facebook"s management to focus on the ad opportunity. Zuckerberg wasn"t going to be around, because he was embarking on a monthlong around-the-world trip, now that he"d completed his search for a number two. He"d wanted to take a break for a while. Now was his chance. He traveled alone, carrying only a backpack, to Berlin, Istanbul, India, and j.a.pan, among other places. In India he made a brief pilgrimage-by dusty local bus-to the ashram high in the Himalayas where Steve Jobs and Baba Ram Da.s.s, among others, have sought enlightenment.

Colleagues believe Zuckerberg timed the trip deliberately, to give Sandberg a bit of runway to establish her authority inside the company without his interference. But it is symbolically apt that her meetings about how Facebook could best turn its vast user base into a powerful business occurred with Zuckerberg-he of the ambivalence towards ads-out of town. Never before had executives from across the organization come together to brainstorm on what people in the Internet business peculiarly call "monetization"-how to turn all those Facebook users into money.

The meetings ran from 6 P.M. P.M. to 9 to 9 P.M. P.M., with dinner brought in, once or twice a week. The first one included a small number of the top ad-related leaders of the company: Mike Murphy, who headed ad sales; Chamath Palihapitiya, who was in charge of growth and international; Tim Kendall, who oversaw the online self-service ad business; Dan Rose, who managed the Microsoft ad partnership; Kent Schoen, head of advertising products; Kang-Xing Jin, the engineer responsible for advertising software (and Zuckerberg"s close friend since Harvard days); and Matt Cohler, Zuckerberg"s tousle-haired "consigliere." On the whiteboard Sandberg wrote, in big letters, "What business are we in?"

These were bull sessions at first, giving everyone a chance to express their views. As they continued, the meetings grew steadily in size. Word spread that you shouldn"t miss these conversations. Pretty soon the entire M team and a larger swath of ad people were making it, a total of fifteen to twenty on a typical evening.

At the time, Facebook"s monetization strategies were varied. Microsoft was selling banner ads, of course, but by the end of 2007, despite the new international deal, Microsoft accounted for less than 25 percent of overall revenue. Facebook wanted that figure down even further, so that it could control its own destiny. The self-service online ads launched at the same time as the disastrous Beacon were now growing rapidly. Facebook also had what it called "sponsored stories"-ads inserted into users" News Feeds that looked like an alert you"d get from a friend, except that it was from Coca-Cola or another company. Virtual gifts, a fast-growing but still tiny share of revenue, were little graphic icons people paid for. For your friend"s birthday, for example, you could buy a little picture of a cupcake with a candle for a dollar. And finally there was the Facebook Marketplace, a cla.s.sified ads system, which had only recently debuted to a lukewarm response from users.

On the whiteboard Sandberg listed the options. Facebook could be in the advertising business. It could sell data about its users. It could sell avatars and other virtual goods to those users. Or it could enable transactions and take a small cut, like PayPal. Staffers researched various markets and brought carefully compiled charts to the next meeting, showing the size of each market, its likely growth rate, the big players, and what Facebook could do uniquely well. After weeks of this, at the final meeting Sandberg went deliberately around the room and asked each person what percentage of Facebook"s revenue would ultimately come from each category. Virtually everyone said 70 percent or more would be advertising in some form.

They all knew Zuckerberg only approved projects that fit into his long-range plan for Facebook. "Mark is very focused on the long run," says one partic.i.p.ant in the meetings. "He doesn"t want to waste resources on anything unless it contributes to the long run. If you don"t know what business you"re in, then anything you do to make money is a waste, because it might not last." While Zuckerberg had been forced by circ.u.mstances to accept advertising, he did so only so he could pay the bills. Whenever anyone asked about his priorities, he was unequivocal-growth and continued improvement in the customer experience were more important than monetization. Long-term financial success depended on continued growth, he believed, and even his grand declarations at the Facebook Ads launch just meant the company would start seeking new approaches. And the Beacon fiasco had shaken everyone"s confidence.

In order to articulate a business strategy that would fit solidly and inarguably into Zuckerberg"s long-term frame, the conferees at the Sandberg sessions went further than merely saying ads were it. They arrived at a crucial distinction to clarify and differentiate Facebook"s opportunity. Whereas Google-Sandberg"s corporate alma mater and the undisputed king of Internet advertising up to now-helped people find the things they had already decided they wanted to buy, Facebook would help them decide what they wanted. When you search on something in Google, it presents you an ad that is a response to the words you typed into the search box. Very often it"s relevant to you and that process makes many billions of dollars for Google. But the ads you typically click on there are the ones that respond to what you already know you"re looking for. In advertising-speak, Google"s AdWords search advertising "fulfills demand."

Facebook"s, by contrast, would generate demand, the group concluded. That"s what the brand advertising that has long dominated television does, and that"s where most ad dollars are spent. A brand ad is intended to implant a new idea into your brain-hey, you should want to spend money on this thing. But such ads have never worked well on Google. You may find a Canon camera via a Google search ad if you type the keywords "digital camera" in the search field, but the company has never found a good way to convince you that you should want a digital camera. (Google"s efforts to find such methods are what have led it to emphasize, for example, its Gmail service, in which its software watches words in your emails and displays messages it thinks you might respond to.) For all Google"s success, it operates almost entirely within a relatively small sector of the overall advertising industry. Only 20 percent-at most-of the world"s $600 billion in annual advertising spending is spent on ads aimed at people who already know what they want, Sandberg"s researchers discovered. The remaining 80 percent, or $480 billion a year, was up for grabs as more and more ad spending shifted to the Internet.

The long-term prospects for advertising on Facebook looked bright to this group. The Internet is pulling consumers away from TV, newspapers, and magazines. And Facebook is taking a disproportionate amount of that Internet time. It is now where Net users spend the most online time by far in the United States and most other countries. That, says advertising marketer Dan Rose, combined with Facebook"s unparalleled ability to target ads based on information about its users, should enable it to attract more and more demand-generation advertising as time goes on. Says Rose: "There is an imbalance between where the dollars are spent and where the audience is spending its time. Those dollars are going to move online over the next ten years." Rose was so effective in the sessions that afterward Sandberg gave him the new t.i.tle of vice president for business development and monetization.

Sandberg"s eight or so business-model sessions concluded just as Zuckerberg was returning from his round-the-world vacation. He was impressed with the group"s conclusions. "Now Mark understands that we have a business model and this is the long-run thing," says one top ad executive. "So now he"s willing to invest."

Zuckerberg explained his ideas about ads on Facebook in detail. At a subsequent off-site meeting on monetization, he told the group what made Facebook different from other websites was its ability to help users have two-way dialogues with one another or with advertisers. "The basic idea is that ads should be content," he says now. "They need to be essentially just organic information that people are producing on the site. A lot of the information people produce is inherently commercial. And if you look at someone"s profile, almost all the fields that define them are in some way commercial-music, movies, books, products, games. It"s a part of our ident.i.ty as people that we like something, but it also has commercial value."

From these discussions with Zuckerberg emerged something Facebook calls the "engagement ad." It is a modest-looking message from an advertiser on users" home pages that invites them to do something right on the page. It might ask you to comment on a video in hopes that friends will be drawn into the conversation. It might be a product giveaway. Starbucks has offered coupons for free cups of coffee. It might enable you to engage in a dialogue with friends right there on the ad. Or it might enable you to click on the ad to instantly become a fan of a product"s Facebook page.

Soon engagement ads replaced the sponsored story as the main product sold by Facebook"s advertising salespeople. Sponsored stories are not, in Zuckerberg"s terms, "organic information that people are producing on the site." The new engagement ads became a big hit. In the first year alone they generated close to a hundred million dollars of revenue. Facebook charges at least $5 per thousand views for these ads. With 400 million users viewing their home page many times a month, those dollars can add up. Moreover, once an advertiser establishes some sort of connection with a user it gets a tremendous amount of what Facebook calls "derivative value." Executives say that once a brand makes a connection with a consumer that leads to an average of about 200 free additional "impressions"-occasions when people on Facebook see information about that brand.