by Steve Levy
Pregibon says that in a certain sense, what Google did with advertising wasn’t much different from what AT&T had done in the era of Ma Bell. “Google makes its money in volume. It gets a quarter or fifty cents whenever someone clicks on an ad. AT&T did the same thing—it had hundreds of millions of phone calls a day, and it would make a dime, fifteen cents on every phone call.” But though both businesses were driven by data, there was a monstrous difference in how they approached it. “AT&T was a hundred-year-old company, and it collected its billing data originally to send out bills! Later it realized that the data was useful for understanding the network, traffic, fraud detection, marketing, and other things. It backed into the importance of data underlying its basic business.”
Google, on the other hand, had been diving for data from day one. Brin and Page began with data mining. That shaped Google’s mind-set from the start. That’s why Google populated not just its search business but its ad business with scientists like Pregibon. If one were to be cynical about his job, you’d say his mission is to get people to click on ads. But Pregibon believed that his role was doing science. He was tackling deep, interesting questions. “It wasn’t presupposed that that is where I would have ended up, but that’s kind of what happened,” he says.
Eventually Google got so adept at understanding what Wojcicki refers to as “the physics of clicks” that it was able to predict not only how many clicks an ad would probably draw but how many sales those clicks would deliver to the advertiser. Google developed a product available to advertisers (like other tools to analyze the success of ads, it was free) called conversion optimizer that shared this information with customers.
To keep making consistently accurate predictions on click-through rates and conversions, Google needed to know everything. “We are trying to understand the mechanisms behind the metrics,” says Qing Wu, a decision support analyst at Google. His specialty was forecasting. He could predict patterns of queries from season to season, in different parts of the day, and the climate. “We have the temperature data, we have the weather data, and we have the queries data so we can do correlation and statistical modeling.” To make sure that his predictions were on track, Qing Wu and his colleagues made use of dozens of onscreen dashboards with information flowing through them, a Bloomberg of the Googlesphere. “With a dashboard you can monitor the queries, the amount of money you make, how many advertisers we have, how many keywords they’re bidding on, what the ROI is for each advertiser.” It’s like the census data, he would say, only Google does much better analyzing its information than the government does with the census results. Google did predictions so well that anomalies shook him. “We wonder if there’s something wrong with us. Are we losing market share?” One year, some weird results came from Belgium at Eastertime, and “we all kind of panicked.” (Turns out it was too warm, and more people than usual stayed home and clicked on Google ads.)
Qing Wu calls Google “the barometer of the world.” Indeed, analyzing the clicks of Google users was like sitting beside a window with a panorama on the world. You saw the changes of seasons—clicks gravitating toward skiing and heavy clothes in the winter, bikinis and sunscreens in the summer—and could track who was up and down in popular culture. Most of us remember news events from television or the newspapers; Googlers analyzing click-through rates recalled them as spikes in their graphs. “One of the big things a couple of years ago was the SARS epidemic,” says Diane Tang. “There was a big spike during the 2008 election. A big spike with Janet Jackson after the Super Bowl.” One Googler studied Google’s data on the day of a massive blackout; there was almost a perfect correlation of Google use with the restoration of electricity.
Varian himself once even did a study that compared Google traffic in individual countries to the state of their respective economies. Not surprisingly, Varian says, high GDP tracks closely with how much people use Google. His paper was titled “International Googlenomics.”
The piles of money that Google made from its golden geese of AdWords and AdSense enabled the company to fund a dizzying array of projects, initiatives, and creature comforts that make it a unique competitor and a most desirable company to work for. “Larry and Sergey think that engineering and computer science can make a big difference in the world,” says CFO Patrick Pichette. “And to have the freedom to do it without having a gun to your head every quarter on financial matters is an immense luxury.” Google’s ad products were a gold-threaded safety net underneath every daring innovation. And Google’s success was hiding in plain sight.
PART THREE
DON’T BE EVIL
How Google Built Its Culture
1
“Make sure it looks like a dorm room.”
One day in 2005, Marissa Mayer was trying to explain why the looniness of Google was actually the crazy-like-a-fox variety and not the kind calling for straitjackets. Responding to the company’s ventures beyond search, outsiders had been charging that Google was out of control, tossing balls into the air like a drunken juggler. And that was before Google decided to remake the energy industry, the medical information infrastructure, the book world, radio, television, and telecommunications. She conceded that to an outsider, Google’s new-business process might indeed look strange. Google spun out projects like buckshot, blasting a spray and using tools and measurements to see what it hit. And sometimes it did try ideas that seemed ill suited or just plain odd. Finally she burst out with her version of the corporate Rosebud. “You can’t understand Google,” she said, “unless you know that both Larry and Sergey were Montessori kids.”
“Montessori” refers to schools based on the educational philosophy of Maria Montessori, an Italian physician born in 1870 who believed that children should be allowed the freedom to pursue what interested them.
“It’s really ingrained in their personalities,” she said. “To ask their own questions, do their own things. To disrespect authority. Do something because it makes sense, not because some authority figure told you. In Montessori school you go paint because you have something to express or you just want to do it that afternoon, not because the teacher said so. This is really baked into how Larry and Sergey approach problems. They’re always asking ‘Why should it be like that?’ It’s the way their brains were programmed early on.”
Both Brin and Page were certainly smart enough and sufficiently self-aware to understand the disrupting impact of unconventional behavior, but it’s as if somewhere along the line—Montessori?—they made independent decisions to act on impulse—even if the results sometimes were, as Mayer says, “mildly socially mortifying.”
Larry—do you realize you just questioned the physical constant to [famed inventor] Dean Kamen? Are you sure you’re right about that?
Sergey—you just asked Colin Powell whether he made the right moves in Desert Storm. Seriously, you’re talking to Colin Powell!
Then there was the time in St. James’s Palace, when they were having dinner with the queen’s husband, Prince Philip. The pomp was intense, a multicourse formal menu. The waiters brought out soufflés along with tiny glasses of passion fruit juice to adorn them, like a syrup. Mayer did what was expected—she mashed down her soufflé and poured the juice over it; otherwise it would have been too dry. She looked on in horror as Larry Page picked up the glass and downed it like a tequila shot. Sergey did the same. Prince Philip looked stunned. Later Marissa explained that the juice was to be regarded as a syrup to flavor the soufflé. She recalls their response with a mixture of awe and repulsion: “Who says?”
“Their attitude is just like, ‘We’re Montessori kids,’” said Mayer. “We’ve been trained and programmed to question authority.”
Thus it wasn’t surprising to see that attitude as the foundation of Google’s culture. “Why aren’t there dogs at work?” asked Marissa, parroting the never-ending Nerdish Inquisition conducted by her bosses. “Why aren’t there toys at work? Why aren’t snacks free? Why? Why? Why?”
“I th
ink there’s some truth to that,” says Larry Page, who spent his preschool and first elementary school years at Okemos Montessori Radmoor School in Michigan. “I’m always asking questions, and Sergey and I both have this.”
Brin wound up in Montessori almost by chance. When he was six, recently emigrated from the Soviet Union, the Paint Branch Montessori School in Adelphi, Maryland, was the closest private school. “We wanted to place Sergey in a private school to ease up his adaptation to the new life, new language, new friends,” wrote his mother, Eugenia Brin, in 2009. “We did not know much about the Montessori method, but it turned out to be rather crucial for Sergey’s development. It provided a basis for independent thinking and a hands-on approach to life.”
“Montessori really teaches you to do things kind of on your own at your own pace and schedule,” says Brin. “It was a pretty fun, playful environment—as is this.”
He was gesturing to his surroundings in an odd little loft in the Googleplex that is restricted to the founders. It was a combination of a rich child’s bedroom and an exhibit hall in the National Air and Space Museum. The floor was covered with an AstroTurf-like carpet. There were sports equipment, game tables, and an astronaut’s suit. A giant Apple display glowed on his desk. The aerie overlooked a savannah of cubicles with shelves lined with gizmos, yurtlike conference rooms, and countless microkitchens equipped with goodie-stuffed fridges and high-end espresso machines. Red physio balls were scattered here and there. The workplace was similar to those in more than a dozen buildings within scooter range here in Mountain View and in Google offices in New York, Kirkland, Moscow, and Zurich. Google offices appeared to be a geek never-never land for unspeakably brainy Lost Boys (and Girls). If you looked closely, though, there were endless bureaucratic structures—data-driven, logically drawn schemata—that kept a $23 billion business humming.
As an indication of this, Brin’s elementary school reverie was interrupted by an unusual occurrence in the Googleplex—a brief power brownout that dimmed the lights. Brin bolted from his chair to his terminal, where he quickly accessed a software dashboard that monitored the building’s electrical system and determined that it was an anomaly. “That’s like the beginner of a Terminator movie!” he said, shrugging it off.
As a corporation, Google was determined to maintain its sense of play, even if it had to work to do it. The high holy day of Google culture is April 1, when imaginations already encouraged to run wild are channeled into elaborate pranks requiring months of work. The effort involves considerable organization, as ideas go through an elaborate approval process to find a place in the company’s ever-increasing roster of seasonal spoofs. The need for some oversight became clear as early as 2000, when Brin sent employees an email announcing that Google had a new valuation (meaning the estimate of its market price had gone up) and would soon reprice its employee stock options—from 25 cents to $4.01. Some people didn’t realize that $4.01 was a reference to the calendar and frantically tried to buy up all the shares that they were entitled to before the price went up. They dug into savings and borrowed from their families. Google eventually had to make people whole.
Google’s external April Fool’s joke that year was an announcement of “MentalPlex,” a search engine that reads your mind, eliminating the need to type in queries. This started an odd succession of self-parodying jokes, where a seemingly outrageous April Fool’s announcement, often involving a step in moving Google toward omniscience, omnipresence, or consciousness, reflected Brin and Page’s actual dreams. (In 2009, there was a complicated announcement of a system called CADIE—Cognitive Autoheuristic Distributed-Intelligence Entity.) As the years went on, more Google divisions felt compelled to devise their own jokes, and by 2010 Wikipedia listed seventeen major April Fool’s initiatives for that year alone.
If April Fool’s was an indulgence of the founders, it must be said that indulgence is spread around at Google. Early in its history, Google instituted a “20 percent rule,” stating that employees can devote one day a week, or the equivalent, to a project of their choosing, as opposed to something imposed by a manager or boss. The idea was Page’s, inspired by similar programs at HP and 3M (supposedly, Post-it notes came from such a spare-time effort). In practice, the self-directed labors often came in addition to a full week’s work. Thus the companywide joke that such endeavors were actually “120 percent projects.” But people participated anyway, and some important products, including Google News, came from the program.
You could even see the company’s work/play paradox in its bathrooms. In some of Google’s loos, even the toilets were toys: high-tech Japanese units with heated seats, cleansing water jets, and a control panel that looked as though it could run a space shuttle. But on the side of the stall—and, for men, at an eye-level wall placement at the urinals—was the work side of Google, a sheet of paper with a small lesson in improved coding. A typical “Testing on the Toilet” instructional dealt with the intricacies of load testing or C++ microbenchmarking. Not a second was wasted in fulfilling Google’s lofty—and work-intensive—mission.
It’s almost as if Larry and Sergey were thinking of Maria Montessori’s claim “Discipline must come through liberty…. We do not consider an individual disciplined only when he has been rendered as artificially silent as a mute and as immovable as a paralytic. He is an individual annihilated, not disciplined. We call an individual disciplined when he is master of himself.” Just as it was crucial to Montessori that nothing a teacher does destroy a child’s creative innocence, Brin and Page felt that Google’s leaders should not annihilate an engineer’s impulse to change the world by coding up some kind of moon shot.
“We designed Google,” Urs Hölzle says, “to be the kind of place where the kind of people we wanted to work here would work for free.”
From the very beginning, Page and Brin had an idea of how Google would be different. “Even when we were three people, we had a culture,” says Craig Silverstein, the first person hired by the founders. “Partly it’s just our personalities, and partly it was the vision that we had for the company.”
That culture took shape even as Page and Brin changed Google from a research project to a company and moved off the Stanford campus. Susan Wojcicki, who owned the house that hosted the company after it moved from Stanford, thought that Google’s origins in a residential setting, with all the comforts of home, set a tone for the eventual bounty of amenities the company would offer its employees. “Because they were working out of a house, they realized that a lot of these conveniences are really important to have,” she says. “For example, having a shower is really important. When you’re attracting a really young group that’s mostly come out of college, having these services is pretty important, like having the food around, having a washer and dryer.”
The Google half of her house, separated from her kitchen by a flimsy door, consisted of a garage packed with equipment; two small rooms used as offices by factotum Heather Cairns and Harry “Spider-Man” Cheung; and a back room with several desks where Sergey, Larry, Craig Silverstein, and another engineer worked, with a view of the backyard and hot tub. Their desks were doors on sawhorses, a setup that would become a Google tradition. “Being a house, it didn’t have a lot of core things you want from a business,” Wojcicki says. “It didn’t have a lot of parking, and you can’t park on the street in Menlo Park at night. And also, they needed a cable modem to get Internet access. I thought it was great, because I got free cable out of it.” (The servers were off-site.)
Wojcicki believes the fabled Google perk of free food began the day Sears delivered the refrigerator she ordered. Her intention was to stay around the house that day so she could instruct the deliveryman to install it in her kitchen; it was intended for her and her husband. But she was in the shower when the truck arrived. “Sergey and Larry answered the door and said, ‘Oh, a new refrigerator! Install it here, in the garage!’” By the time Wojcicki realized what had happened, she was the unintentional benefactor of
the first Google snack station.
“We just had to be clear about the rules,” says Wojcicki. When guests came to Google, they had to enter by the garage; using the front door would mean traipsing through her home. There was the occasional weird moment, like the meeting at Intel, where she worked. Her coworkers were talking about this hot new start-up called Google. “They work from my house,” she said, drawing astonished stares. Generally, she loved being the landlady. She could have contractors come by even when she was at work. “I would say, ‘The electrician’s going to come, show him the light that needs to be fixed.’” Her husband traveled a lot, and when she got lonely, she would go to the other side of the house and talk to the Googlers. After a number of late-night sessions when she’d heard Larry and Sergey’s dreams time and again, she quit Intel to join Google herself. Eventually Sergey began dating her sister. (Anne Wojcicki and Sergey would marry in 2007.)
In early 1999, Google moved to its new office space on University Place in Palo Alto, over the bicycle shop. The conference room had a Ping-Pong table, and, maintaining the tradition, the desks were doors on sawhorses. The kitchen was tiny, and food was yet to be catered. Larry and Sergey’s fondness for physio balls was apparent, as the red and blue plastic spheres were scattered about.
There is a special magic in a start-up of barely a dozen people whose entire existence centers around the shared dream of building the next Apple or Microsoft, only better. At the end of the night, when people who had families and homes with furnishings and air conditioners would otherwise have gone home, Google’s young engineers would engage in an iteration of the kind of rambling bull sessions they had experienced in college—only a year or two earlier. “We’re all working, like, a hundred and thirty hours a week and sleeping under our desk and doing all this stuff,” recalls Marissa Mayer. “But at two or three in the morning, the office would degenerate to us all sitting around on the couches and balls, chitchatting about what we’d do if there weren’t only ten or twelve of us,” says Mayer. Amping up the thrill was the fact that Google search was generating feedback and excitement far beyond the few cluttered rooms they occupied. Press notices were coming in. They were getting fan letters from librarians, scholars, schoolkids. This was real data indicating that Google actually could change the world. It was like some amazing logic drug.