by Steve Levy
The sudden enrichment of Google’s workforce presented a serious threat to a culture that aspired to a certain public humility. (Matt Cutts expressed the unspoken code this way: “I like to drive fast, so I don’t have any Google-related stuff on my car. I don’t want to cut someone off and as I go zooming off have someone say, ‘What a jerk! Oh, he’s a Googler!’”) Its executives took some steps to stem the toxic behavior seen during the recent tech bubble, where newly coined millionaires paid more attention to lucre than product development. In the Googleplex, one’s personal wealth could now be constantly monitored on the same computer screens that engaged everyone’s attention at every moment. On the day of the IPO, Wayne Rosing, the head of engineering, addressed an all-hands meeting. In his hand he held a baseball bat. He told the Googlers that if he looked in the parking lot in the next few days and saw new BMWs or Porsches, he would use the bat to smash the windshields.
Marissa Mayer told her team that she didn’t want them checking the stock price during the day. When her workers did not respond with full compliance, she instituted another policy: if anyone who worked for her spotted someone else in the group looking at the stock ticker, all he or she had to do was walk over and tap that person on the shoulder. Then that person would have to buy you a share of stock. After a number of involuntary exchanges, people either stopped checking or learned to hide their peeking more effectively.
But Googlers were affected by stock ownership. (They were, after all, human.) Bo Cowgill, a Google statistician, did a series of studies of his colleagues’ behavior, based on their participation in a “prediction market,” a setup that allowed them to make bets on the success of internal projects. He discovered that “daily stock price movements affect the mood, effort level and decision-making of employees.” As you’d expect, increases in stock performance made people happier and more optimistic—but they also led them to regard innovative ideas more warily, indicating that as Googlers became richer, they became more conservative. That was exactly the downside of the IPO that the founders had dreaded.
Regulations stipulated that Google employees could not sell stock for ninety days after the August IPO. By that time, the stock price had risen to $175 a share. During a press interview at the Googleplex that November day, Eric Schmidt did his best to convey that Googlers were going to remain the same down-to-earth geeks they always were. “Somehow there’s this assumption that the people at Google made money and are going to retire on their boats,” he said. “These people don’t sail. Some of them do need to buy a house—they’ve been living in itty-bitty apartments.” He turned to David Krane, a onetime English major who was now a tech millionaire. “Do you sail a lot?” he asked.
“I don’t sail,” said Krane.
“You see what I’m saying?” said Schmidt. “Look around—everybody’s here!”
Indeed, that day the people at Google were improving search quality, selling ads, and figuring out how to work the espresso machine—not sailing. Six years after the IPO, an impressive number of Google’s most important early employees—executives such as Susan Wojcicki and Salar Kamangar and core engineers such as Amit Singhal, Ben Gomes, and Jeff Dean—were still working hard at Google, even though they had the wealth of Saudi princes.
Still, that personal wealth would inevitably change the lifestyles of early Googlers. How could it not?
Not long after the IPO, Marissa Mayer shared a recent revelation with a reporter. Previously if she were in a department store and there was a pair of slacks that cost a hundred dollars, she would ask herself whether or not she should buy the slacks. Now she would just buy them. Mayer would later purchase a house in Palo Alto in addition to a penthouse suite in the San Francisco Four Seasons Hotel, and Oscar de la Renta would tell Vogue that Mayer was “one of his biggest customers.” If you spent time with other early Googlers, it would sometimes slip out that they owned lavish homes in posh Atherton, California, vacation retreats in Hawaii, pied-à-terre brownstones in New York City, and other indications of brimming bank accounts. Eric Schmidt, who was already a tech magnate when he joined Google, owned several airplanes and a yacht. Larry Page would buy his own $60 million pleasure boat. (Not all Googlers eschewed sailing, it seems.) The key was keeping it on the down low. When someone failed to maintain that discipline, colleagues would note it.
Even the Google masseuse noticed the impact of money, especially when it came to the divide between early employees holding valuable options and those who came later. “While one was looking at local movie times on his monitor, the other was booking a flight to Belize for the weekend,” she said in a book she wrote. “Don’t think everyone wasn’t aware of the rift.”
Schmidt came to see the IPO as a necessary rite of passage for Google. “I don’t understand, and will never understand, some of the specific trade-offs that were made, and to be honest, we don’t need to. We were never going to do it the way anybody else did,” he said that November. Framed on the wall in the room where he spoke was a poster-sized certificate from Morgan Stanley congratulating Google on the sale of 22,534,678 shares of stock in its August 2004 initial public offering, at the opening price of $85 a share. On the glass covering the poster someone had stuck a yellow Post-it note saying, “SHOULD HAVE BEEN $135.”
Five years later, the Post-it note was still there.
3
“People don’t want to be managed.”
“Yes, they do want to be managed.”
By the time of its IPO, Google had grown to the size where a company usually sets aside its loose structures and adopts well-established management structures. But Google, in Page’s words, was “not a conventional company.” Page and Brin wanted it all: a company with thousands of engineers that ran smoothly while still indulging the creative impulses of its people. Every time the head count doubled, the question came up again: could Google’s bottom-up style of management actually scale? Page and Brin never doubted it. They envisioned an organizational map of Google as looking like a huge sheet covered with polka dots: small teams, flat organizations. The sheet would just get bigger, that was all.
Both Page and Brin believed that the company should run like the Internet itself: fast-moving, bottom up, going to work every day to make yesterday obsolete. “We were born in the Internet time,” says Megan Smith, “so our company’s like our products in some weird way.”
Google, however, had been through an early ordeal that showed that this flat-org ideal was unattainable. In 2001, Google had more than four hundred employees, reaching the point where it was impossible to pretend that it was an intimate company where everyone knew everyone else. Worse for Page and Brin, despite their best efforts, a layer of middle management was creeping in. Worse still, some of the newcomers were experienced product managers from companies such as Microsoft, whose training made them un-Googley—and those newcomers had difficulty adopting the often heretical approaches of the founders.
Brin and Page came up with a solution: Google would no longer have managers. At least not in engineering. Instead, they figured, the engineers could self-organize. That approach worked well in the nascent days of Google. If something needed fixing, people would figure out on their own what was wrong, and what was broken would be fixed. Other people would identify interesting problems in computing, and from those insights new products would arise. At the time Google had just hired Wayne Rosing to head engineering. Brin and Page figured that everyone could just report to him. The engineers would arrange themselves in pods of three, work on projects, and check in with Wayne.
That struck some of Google’s executives as madness. Stacy Sullivan, the head of HR, begged Page and Brin not to go through with it. “You can’t just self-organize!” she told them. “People need someone to go to when they have problems!”
The newly arrived Schmidt and the company’s unofficial executive coach, Bill Campbell, weren’t happy with the idea, either. Campbell would go back and forth with Page on the issue. “People don’t want to be ma
naged,” Page would insist, and Campbell would say, “Yes, they do want to be managed.” One night Campbell stopped the verbal Ping-Pong and said, “Okay, let’s start calling people in and ask them.” It was about 8 P.M., and there were still plenty of engineers in the offices, pecking away at God knows what. One by one, Campbell and Page summoned them in, and one by one Page asked them, “Do you want to be managed?”
As Campbell would later recall, “Everyone said yeah.” Page wanted to know why. They told him they wanted somebody to learn from. When they disagreed with colleagues and discussions reached an impasse, they needed someone who could break the ties.
Nonetheless, Page and Brin were determined to go through with the plan. They called an all-hands meeting and announced it to a baffled workforce. For a few people it meant leaving the company. Others scrambled to find new roles. On the other hand, the move was welcomed by the engineers, who had been chafing at the creeping management restraints. For example, Eric Veach, who at the time was trying to invent the auction-based AdWords, later said that losing a manager had liberated him to make his breakthrough.
Ultimately, however, the plan petered out. After the initial turmoil, there was a quiet backslide where Google’s managerial class reassembled and regained a place in the structure. You just couldn’t have more than a hundred engineers reporting to Wayne Rosing. Google was taking on new engineers at a furious rate, and, brilliant as they were, the new people needed some guidance to figure out what to do. “I don’t remember Larry and Sergey saying that they were wrong and that we were right, but they agreed we could start to hire managers again, as long as the managers were good culture fits and technical enough and could be highly respected by the engineers,” says Sullivan.
Another organizational crisis at Google centered more specifically on product managers—the people who led the small teams of engineers. At Google, teams would typically have a tech lead (the smartest engineer) and a product manager. But it didn’t seem Googley to have lesser minds tell the brainiest engineers what to do. Unlike other tech companies, as late as 2001, Google didn’t have a top executive focusing on product management, and Schmidt kept suggesting candidates. Not convinced that the job should exist at all, Page and Brin kept rejecting them. Then Schmidt heard that Jonathan Rosenberg, a former executive at Excite@Home, had once impressed the founders during an interview for a VP of marketing job he’d turned down.
Schmidt begged Rosenberg to come in and talk. To get a sense of Rosenberg’s skills, he was asked to audition by delivering a test briefing. At one point in his canned presentation, Rosenberg stared at a spreadsheet calculation in his PowerPoint deck and corrected a subtle mathematical error. Everyone was blown away. (In fact, Rosenberg, knowing that Sergey Brin was supposed to be some sort of math Olympian, had planted the mistake and faked his spontaneous discovery.) Schmidt showed Rosenberg Google’s astounding financials, convincing Rosenberg that the job was the opportunity of a lifetime.
But his first year was awful. Larry Page would sit in meetings and second-guess every move Rosenberg made. “I would come to the staff meeting with my structured agenda, the market research we needed to do, the one- and two-year road maps that we needed to develop, and Larry would basically mock them and me,” Rosenberg later said.
For Rosenberg, whose leadership style was based on aggression and self-confidence, it was a crushing experience. Google’s executive shrink, Bill Campbell, suggested that Rosenberg ask Page what he thought Rosenberg should do. Page said that instead of working with schedules and plans, Rosenberg should just listen to the engineers. They were the ones with the ideas that mattered. After taking the advice to heart, Rosenberg finally got it: engineers rule. Page wasn’t out to get him—he and Brin were just adamant about not wanting product managers telling engineers what to do. He reevaluated his view of Page. Larry wasn’t ignorant of management processes; he was simply not an effective communicator. It wasn’t until a couple of years later that Rosenberg got the acknowledgment from Page he really sought. Page was showing his mother around Google one day, and he introduced her to Rosenberg. “What does he do?” she asked Larry. “Well, at first I wasn’t sure,” he told her. “But I’ve decided that now he’s the reason I sometimes have free time.”
Even as he sorted out his role, Rosenberg had another problem, a difficulty in getting product managers hired. His usual modus operandi was to go to places like Stanford and Harvard and bag the Baker Scholar or the R. J. Miller Scholar. But Page would meet such people and send them home jobless. “They would talk about paradigm shifts and competitive advantage and all that shit Larry wasn’t interested in,” Rosenberg says. “They weren’t technical.” It was Marissa Mayer who told him the obvious—Page wasn’t looking for project managers who were smart enough to understand engineers—he wanted them to be engineers.
Mayer suggested that Google look for computer science majors who saw themselves not just as engineers but as future CEOs. Her idea was to assemble a legion of “associate product managers.” Google would get them straight out of school, young people with no preconceptions derived from working elsewhere. Their careers would coevolve with Google. “We value insight over experience,” says Mayer. “We take people who we think have the right raw skills and insights and put them into roles with a lot of responsibility. And while that happens with APMs, it also happens all across the company. People here might not really be accomplished or have a long career before coming to Google, but they have the right data instincts about their area.”
It took months to find the first APM, a Stanford grad named Brian Rakowski. But what would he do? Mayer decided to put him in charge of launching an important product, a web-based email system. On Rakowski’s first day, she met him in the Ping-Pong conference room. “You’re going to be working on Gmail,” she said. Rakowski was speechless. “I was twenty-two years old,” he says. “I was shocked that they were going to let someone that young and inexperienced do that job.”
A lot was riding on Rakowski—if he flamed out, the program would be tarnished. The engineers on the project didn’t welcome a kid coming in as their PM. They asked Mayer if they could interview him first. She reminded them that he was already hired. The situation was defused a bit when the engineers checked out Rakowski’s web page. There was a picture of him taken after he’d had dental surgery, and his cheeks were puffed up like some sickly bunny’s. This was an indication that the kid had a certain wit and humility. But they still had to submit him to a technical gauntlet, just to make sure that his Stanford CS degree wasn’t just some anomaly. Even then, the job presented Herculean challenges. A product manager at Google didn’t give orders. His (or her) job was to charm the engineers into a certain way of thinking. It was a Mensa form of cat herding.
The way to do that, of course, was by hard numbers. Information was the great leveler at Google. “Because the APMs work with people who are so much more senior and more experienced, they don’t have the authority to say, ‘Because I said so.’ They need to gather the data, lobby the team, and win them over by data,” says Mayer. That process made the managerial weakness of the APM an asset for Google, by making sure that data was at the center of decision making. (Google further cemented this hierarchy by creating a position called UTL, or über tech lead—a wizard-level engineer on a bigger team who really calls the shots.) If an APM had an idea, he or she could order up a 1 percent A/B experiment (in which one out of a hundred users gets a version of the product with the suggested change), then go to the über tech lead and the team and say, “Users with this new experience are doing 11 percent more page views and clicking on ads 8 percent more.” With ammunition like that, a decision to include the new feature in the product wouldn’t be based on a power struggle but on a mathematical calculation. Nothing personal. It was data.
The APM program was a huge success. Google provided its young managers support in the form of regular meetings with Mayer and her staff and even regular sessions with executive coaches. Ultimately, th
e program helped Google maintain its team approach while still focusing on engineering as opposed to the kind of more elusive un-Googley skills that an MBA brings. (One might also note that Google, in its management practices and hiring preference for freethinkers, has achieved a complete turnaround from the ethic posed in William H. Whyte’s 1956 classic The Organization Man, which describes the perfect corporate employee as “obtrusive in no particular, excessive in no zeal”—the polar opposite of a Googler.)
Executives at Google were still worried about keeping the company’s teams lean. “Google still does try to keep things small and have teams that are really motivated, who feel they own the project,” says Urs Hölzle. But when a team begins to get too big, Google breaks the project into smaller pieces to keep the teams smaller—it refers to this practice as “load balancing,” as if its people were servers in a data center.
Another form of corporate load balancing assures that engineers’ dreams won’t mess with the bottom line. Around 2005, Google determined a simple formula to distribute its engineering talent: 70–20–10. Seventy percent of its engineers would work in either search or ads. Twenty percent would focus on key products such as applications. The remaining 10 percent would work on wild cards, which often emerged from the 20 percent time where people could choose their own projects. For all the talk about its other, well-publicized fraction—the 20 percent of free time that supposedly gestated Google’s big innovations—70–20–10 became Google’s magic allocation algorithm.
As the years passed and Google’s management system became formalized, a corporate amnesia seemed to envelop Brin and Page’s 2001 kill-the-managers caper. Ask Larry Page about it, and he professes only a vague memory. “We were two years old as a company,” he says. “You try different things, and we learned things that worked and things that didn’t.”