Cindy was reluctant to stir the PR pot further. Larry and Sergey knew we couldn't bluff a group so famously litigious in protection of their copyrights. "Scientology will never back down," they advised Matt. "Focus on the per-search notification alerts. Figure out a way to let users know that some of our search results have been filtered."
Within a week, engineers Daniel Dulitz and Jen McGrath had come up with a solution. Any time a DMCA notice necessitated the removal of a search result, we explained that a result had been deleted and provided a link to a copy of the complaint on a website run by ChillingEffects.org. That way Google met its legal obligations while still letting users know what information had been removed.
Our user mail rapidly turned more positive. The New York Times ran an article about our innovative approach and noted that publicity about Scientology's complaint had pushed xenu.net to the second-highest spot in the search results for "Scientology"—just below the church's official site.
The xenu.net episode went a long way toward establishing our credibility with the hard-core, libertarian-leaning, free-speech army. I suspect some would have preferred Google go down in a blaze of glory, expending all our resources fighting the Scientologists in a Supreme Court smackdown, but they appeared somewhat appeased by Google's innovative way of "fighting without fighting." The outcome reinforced Larry and Sergey's optimism that there would always be a creative, technology-based solution when we got ourselves in a jam. We just had to be intelligent enough to see it, and if Google employees had anything, it was off-the-charts intelligence. Yeah, we could most definitely innovate our way out of anything.
That hubris would carry a price in years to come.
Chapter 21
Aloha AOL
I RAN INTO SOME Hawaiians," I informed our executive staff in March 2002. "They said Google had the best search technology in the market." Normally I wouldn't bother Larry and Sergey with compliments passed along at a marketing conference, but "Hawaii" was our code name for America OnLine (AOL).* We were scrambling to win their business from Overture, and the AOL vice-president I had spoken with gave every sign of being favorably disposed toward Google.
Overture was at the conference too. Their representative spent most of his time trying to convince our sales director, Tim Armstrong, to hire him. I took that as another good sign.
Our attack on Overture's business had been in the works for months, from the time Larry and Sergey had given the green light to AdWords Select. Launching an ad-syndication network was a major undertaking and new ground for Google. Omid realized it would take a unique combination of knowledge and skill to pull it off. Joan Braddi, the head of our sales team, could do it, but she was bogged down selling search services and managing the day-to-day sales effort. She would be part of the team, but she would need the help of a specialist. Omid knew exactly whom to call.
Alan Louie had worked with Joan and Omid at Netscape and now operated as an independent consultant. Lean, energetic, and given to wearing shades and safari hats, Alan had a degree in physics and had worked as an engineer at the Jet Propulsion Laboratory in Pasadena before moving into sales. He only took on unusual jobs and he only worked part-time. He didn't want direct reports. He didn't want to manage a team. He would handle the project, then leave when it was done. He started on contract at Google in October 2001. His assignment was to close the AOL deal, and he went about it with the dispassionate precision of an assassin stalking a high-profile target.
In addition to Alan and Joan, Omid pulled in Miriam Rivera, the second lawyer to join our legal group. John Barabino, who would head the syndication effort once the AOL deal was completed, became part of the team the day he was hired in February 2002.
Alan understood exactly what was at stake, and he knew the opposition. Overture had been born as GoTo in Pasadena, right down the street from the Jet Propulsion Laboratory. Even though they drew talent from Cal Tech and were plenty smart, they weren't Silicon Valley smart. They weren't really a tech company, and they didn't have the business intelligence Alan saw at Google. On the other hand, they had come up with the idea of marrying ads with search and implemented it before Google had.
When Omid called the AOL business-development unit run by David Colburn, AOL agreed to start talks with Google. Not because they intended to give us their business, but because they could use Google as a bludgeon to beat a better deal out of Overture when their contract came up for renewal.
Alan expected that. AOL was the giant of the Internet, with more traffic than any other site. They were doing deals every day and had enormous leverage. They had a reputation for being aggressive, foul-tempered, bloodthirsty, and brutal—and that was with their partners. Google was a pipsqueak trying to break up AOL's happy marriage to Overture with AdWords Select, an untested, unlaunched product with uncertain revenue potential. AOL was already making millions off Overture ads and had little incentive to put that guaranteed revenue at risk.
None of that fazed Alan or the Google team, who approached AOL as if we were already peers, not supplicants. "They wanted an emperor-to-toe-kisser kind of relationship," Alan recalls. "We came off as, 'Okay, you can be emperor, but we'll be the pope.'" Part of the team's swagger came from their sense that AOL was heading up a blind alley toward a dead end. AOL was a walled garden offering screened, selected content to its subscribers. Google was coming from the Internet, an open system without limits. "We knew it was the overconfidence of youth, but it turned out to be correct," Alan told me. "It's just the confidence of knowing the industry and knowing what's going on."
Still, AOL was not going to come to Google to discuss the deal, so Google went to AOL. Their headquarters in Vienna, Virginia, was a sprawling complex not far from Dulles airport. It was hard enough on Omid's team that they were arriving at AOL's labyrinthine fortress armed only with an idea of what they wanted to sell, but all of them were also weakened by the five-hour flight and intense head colds. Their first stop was a pharmacy, where they loaded up on Cold-Eeze and Kleenex. Blowing their noses, they went to meet AOL's negotiating team.
AOL came on strong. The four staffers from Google, dressed in business-casual attire, were met by a dozen negotiators, product managers, and lawyers in expensive suits—a staffing ratio that AOL maintained throughout the negotiation. "They would always have this random cast of characters coming in and out," recalls Alan. "They'd have these people sitting in and we had no clear idea of what they were doing there."
Overwhelming presence was just one of many negotiating tactics. AOL also created artificial urgency, requiring the Google team to jet cross-country to address an issue that then dragged out over a week. The Googlers huddled in hotels, nursing their colds, while AOL's team went home to their families. If Google's negotiators were physically exhausted, they might cede key points, just to get some rest.
The tactic didn't work. "We've had enough," Alan finally announced to his AOL counterparts. "We're going back. We'll just do this over the phone."
"Great," they responded. "Go back home. You guys rest up over the weekend and just come back Monday."
Alan wasn't about to play that game. "Forget it," he said. "We're going home to rest up and we're inviting you guys out to California on Monday. You guys are much stronger and more able than we are, so you guys should come out to California. You guys are just supermen and we're not. I admit it. I'm a wimp."
AOL didn't fall for that. "Give it a rest," they told Alan. "We'll do it by conference call."
The deal would go nowhere, though, if we couldn't back up our claims about potential ad revenue with real data. While we had proven the value of our search technology with the Yahoo deal, Overture would make the most of our lack of a track record with syndicated ads, sowing FUD (fear, uncertainty, and doubt) about Google in the minds of AOL's execs. Overture was a known quantity and a reliable partner, and knew AOL would never agree to be our guinea pig. For a publicly traded company, the risk was way too high.
The only solution for Googl
e was to sign a syndication partner ahead of AOL. Alan wanted someone who was already a good search partner and friendly toward us. Someone he could sign quickly. That was the real significance of the deal he negotiated with Earthlink in January 2002. Earthlink was our proving ground for AOL.
The Earthlink partnership was the ante that bought Google a seat at AOL's table. Now the real game began. There were no rules. David Colburn was an infamously intense negotiator sometimes called "the Butcher" for demanding—and getting—a pound of flesh from potential partners.* According to Alan, Colburn and the rest of AOL's team were the toughest negotiators he ran into during his time at Google. Their basic technique was to keep asking for more until the other team screamed uncle.
Colburn, Alan said, "would keep changing deal terms that had already been decided. We'd agree on certain things and then he'd just change them randomly—the prepay, the revenue split." These were key components of the deal, as was the amount Google guaranteed to deliver to AOL regardless of how the ads performed.
If we agreed to too large a guarantee, it could bankrupt us. The Google team knew the number would have to be big, because Overture was already providing AOL with a significant revenue stream. If Overture was delivering ten million dollars a year, AOL would build on that. They would argue they were growing at fifty percent a year, so the number Google needed to guarantee was actually fifteen million dollars. We would respond that there was a limit to how much Overture's flawed system could deliver, and that Google's method of ad ranking was so effective it would yield a much richer payoff. Ultimately, we would point out, Overture was training people to ignore ads altogether because their system didn't automatically screen out irrelevant ads as AdWords Select did. "You think it's okay for your ads to get ignored?" Alan asked AOL. "No? So then you have to agree with our system."
No, AOL rejoined, Overture would clean up their system and make it more relevant. It wasn't in AOL's interest to admit that Google worked better, even after they came to believe it was true. "I think intellectually they agreed," Alan speculated, "but they wouldn't agree as part of the negotiation. They would always maintain that our system was worse. Their face to us was, 'never give an inch.'"
Overture didn't sit idly by while we tried to steal their biggest customer. They struck back using one of the most powerful tools at their disposal: Google AdWords. Omid was in a meeting with the AOL team demonstrating our superior product when ads began showing up on searches saying "AOL sucks." The ads clearly hadn't been screened, so it was obvious to AOL that Google didn't have sufficient editorial controls in place. Omid excused himself and frantically called home to have the ads taken down.
AOL said, "This is a problem."
"It was a good move on Overture's part," admits Alan. "I would have done the same thing." The result was that the four members of Sheryl Sandberg's AdWords support team began hand-screening every single ad that was submitted. And AOL asked for a better understanding of Google's approval process.
"Here are Overture's policies," AOL said, dropping a phonebook-sized document on the table with a thud. "These are their editorial policies for what is allowed and what is not. Let's see yours."
"Fine. We'll show you ours," Alan replied. "Let's set up an appointment and do it right." We had no policy manual. But we would have one by the meeting the following Monday. Five minutes after that exchange, Omid was on the phone to Sheryl. "Overture has a binder," he told her. "We have to have a binder! We have to have a binder!"
Sheryl and AdWords staffer Emily White spent the weekend pulling one together. While they were at it, they pulled together an editorial team. AOL wanted to know how many people Google had dedicated to reviewing ads and approving them. By enlisting everyone in advertising operations who had ever looked at an ad, Sheryl stretched four to fifteen. "We didn't lie to them," she asserted, "but we included everyone we possibly could."
And Alan struck back by pointing out the weakness in Overture's system. He had the AOL negotiators search for "flight" to see the ads that came up. The top position wasn't occupied by American Airlines or Travelocity or Expedia. It was owned by a Midwest flight school for wannabe pilots that was willing to bid high for placement, knowing that thousands of people would see the ad but ignore it. These "squatters" paid only if someone clicked, so the exposure cost them nothing, but for Overture and AOL, it meant that the top position on a popular keyword generated almost no revenue. Google's more relevant ads would perform better.
Alan also showed AOL that we were willing to sweeten the pot at our own expense. Before Google entered the syndication market, Overture had owned a near monopoly. They set the terms for partners on revenue splits. For small partners the split might have been fifty-fifty, but for a large partner like AOL it would have been more generous. AOL might have been taking seventy percent of the revenue when someone clicked an Overture-supplied ad on their site. Already, though, as George Mannes had postulated in his article, Overture's margins were slipping downhill. Our negotiating team gave them a helpful shove. We were willing to offer AOL eighty percent, a number we could afford because we kept all of the revenue for ads on Google.com.
AOL noted our strategic goals and our generous gesture and demanded ninety-one percent. "We're going to make your network," AOL told Alan's team.
"No, you're not," Alan responded. Everybody knew he was bluffing. Larry and Sergey were desperate to kick-start our ad-syndication program, and a deal with AOL would leap across the Internet, creating a network effect that would bring in thousands of other sites. AOL had leverage and they used it to push harder and harder.
A key for Google was exclusivity for the placement of ads on AOL's pages. The more places to click on a page, the lower the odds a user would click on one of our syndicated ads. If Google was going to make the huge revenue guarantee AOL demanded, we would need every penny a page could generate. AOL kept offering "non-exclusive exclusivity." They drafted a ten-page section just on that topic, filled with loopholes and inconsistencies that would allow them to keep working with other partners. In the end, AOL kept the ability to run someone else's banner ads, and Google got exclusivity for text links in search results.
Then Overture counterattacked. On April 4, they filed a lawsuit claiming that "bid-for-placement" belonged to them and that AdWords Select violated their intellectual property rights. They were taking us to court.
It's not ideal, when negotiating a contract, to have your rights to the technology you're selling be called into question. AOL immediately seized on our weakened position to push for even more concessions. Now they wanted access to the intellectual property behind our ad-serving system.
"Hey, they're suing you," AOL's team reasoned. "If you go down, we've got to go back to Overture hat in hand, so we want all of your intellectual property related to the ads system. Oh, and we want more money from you as well. Up front." It was the upfront payment that threatened Google's existence. If we miscalculated the revenue flow, we would be caught in a cash crunch that would kill the company.
Like bulldogs, the AOL team tightened their grip each time they sensed an opening, gradually moving closer to the jugular. As the far smaller partner, Google kept giving in.
Alan saw a need for a tactical shift. "I think we should say no and see what they do," he said to the rest of the team. "We can always go back and say yes." Step away from a deal and see what the other side does. If they were willing to walk away, we would run back, beg forgiveness, and throw an extra million dollars at them to show our sincere regret for offending them. It was that or keep acceding to every escalating demand.
For once, though, Larry and Sergey were unwilling to take the risk. "So we just kept agreeing until we screamed," recalls Alan. "Until they came back with something really egregious." AOL asked for all of Google's intellectual property—our code and secret search algorithms—if it ever looked as if we were headed for financial ruin. They had finally gone too far. Larry, Sergey, and Eric were furious at their demand for the very hear
t and soul of Google. No fucking way, they agreed. Their heels dangling over the cliff, they finally refused to take another step back. Alan recognized AOL's request for what it was: the closing gambit.
"That was the signal Colburn was waiting for," Alan told me. "Did he maximize his prize? Did he get everything he could get? As long as we were willing to give, we were the gift that kept on giving." It was a classic negotiation move. The only way to find the limit was to push past it. Once AOL knew they would get nothing more, the deal could be closed. Miriam was relieved. She had spent thirty consecutive days getting by on two to four hours of sleep each night, and her hair had become noticeably thinner. At one point David Drummond, our new VP of corporate development, had grabbed her wrist to keep her from throwing a pen at one of AOL's attorneys. Even by Google standards, the stress had been nontrivial.
We still needed a final number for the guarantee—the amount that we would pay AOL even if no one clicked a single ad. It was a deep pit into which Google would jump, with the faith that AdWords Select would pull us out before it collapsed. That faith relied entirely on revenue projections for the performance of AdWords on AOL. As CEO, Eric asked for three independent models, one from Susan, one from Salar, and one from Alan. They began pulling numbers from the limited data being delivered by Google.com and Earthlink. Neither, however, was AOL, so a lot of their assumptions primarily relied on SWAGs—scientific wild-ass guesses.
Alan was the sales guy. He was born optimistic. He looked at the numbers and saw a baseline. Google and Earthlink didn't have AOL's traffic or reach and their websites were not oriented toward consumer purchasing. People on AOL bought stuff. "I argued that as long as the advertisers were making money, they would create more ads," he told me.
As Susan and Salar saw it, that might be true theoretically, but in reality there were constraints that would prohibit meteoric growth. They plugged smaller numbers into their forecast assumptions. Salar was the most pessimistic; Susan was in the middle.
I'm Feeling Lucky: The Confessions of Google Employee Number 59 Page 35