In The Plex
But Ivester’s experience showed that Google could accommodate exceptions to its standards. Just as in the case of elite institutions, the stray C or a non-Mensa SAT score could be trumped by an accomplishment that indicated that one was special. “It’s like they did some crazy skiing thing or could do the Rubik’s cube better than anybody,” says early employee Megan Smith. Stacy Sullivan could recall having trouble hiring someone in international sales—until she noted that his résumé cited a foosball championship in Italy. “That’s pretty good,” said Sergey. “We can hire him.” If the guy worked that hard at something, the logic went, he’d probably be pretty good at selling ads. And if you were stuck at the airport with him, you’d have the best foosball conversation ever.
Tales of the stringent Google hiring process gave rise to an entire genre of web literature, generally mini-memoirs about how the author had navigated (and, more often than not, failed at) Google’s arcane hiring obstacle course. Generally, even those with dashed hopes expressed gratitude at a lesson well learned and a great meal at Charlie’s Café.
As years went by, the company streamlined the process. After a period when candidates would go through a series of as many as twenty interviews, Google whittled down the number. Even though the company’s metrics determined that after four interviews the returns diminished, candidates often had closer to eight. “It used to take forever—anywhere between six and twelve months—to get hired by Google. Now it takes, on average, something like forty-six to sixty days from start to finish,” says Laszlo Bock, Google’s director of People Operations. (He describes his role as “HR with math.”) If a hiring council wanted to go beyond eight interviews, it had to seek Bock’s approval.
Still, even in its short form, landing a job at Google put you through a process that made a Harvard application look easy. The interviews, loaded with brainteasers that would challenge Gauss and computer-coding versions of Jeopardy, were only the first stage. The recruiters would comb the data, and, if they were high on the candidate, they compiled a detailed packet including all the interview feedback, academics, references, and so on. It could run twenty to forty pages. Then the application went to a hiring council made up of people with some expertise in the area—but not those who would directly manage the new employee. Otherwise, the temptation would be too great to give an offer to a substandard employee because “every manager wants some help rather than no help,” says Peter Norvig. The council then did its own analysis. “We read through about eight pieces of feedback—each is more than a page—that discuss analytical ability, overall intelligence, technical skills, cultural fit, résumé, and sort of an overall summary piece,” says Marissa Mayer. If the council gave thumbs up, an executive management group reexamined the packet to make the penultimate call.
The last word always went to Larry Page, who insisted on signing off on every employee hired at Google. For every hire, he was given a compressed version of the packet, generated by custom-made software that allowed Page to quickly see the salient data but also empowered him to probe into the gritty details should he choose. “It’s sort of a nested electronic index of everything,” he says. Page would get a set every week and usually returned them with his approvals—or in some cases bounces—in three or four days.
Page didn’t think it unusual or a control-freak quirk that his personal seal was required for every hire. “It helps me to know what’s really going on,” he says. “I can get a pretty good feel for that in a short amount of time. I occasionally do a spot check, to ask what is the real quality of person we’re hiring.” On the day he spoke about this, in early 2010, he cited his most recent session, a few days earlier. “It only took me about fifteen to twenty minutes to do, and we hired probably over a hundred people.”
2
“I look at people here as missionaries—not mercenaries.”
It was Bill Campbell’s idea to gather a few key Googlers together and hammer out a set of the young company’s corporate values. He had no idea that it would be the source of a motto that would become a controversial self-definition of the company—a combination guiding light and curse.
Campbell was a Silicon Valley legend: if movers and shakers there were assigned human PageRanks according to important links, he’d be a rare 10. A former football coach at Columbia who had run Apple’s software company in the 1980s, Campbell was the chairman of the software company Intuit. He was also best friends with Steve Jobs; in the Valley that was like being “1” on God’s speed dialer. In early 2001, John Doerr had brought him into Google for an unofficial but critical role as an executive coach. A burly, profane straight shooter who mixed his macho with bearlike hugs and verbal wet kisses, Campbell improbably connected not only with Brin but with the not-so-huggy Page. Even more than Eric Schmidt, he became the father figure in Google’s corporate family and was instrumental in easing the tensions in Schmidt’s bumpy process of establishing his role in Google’s ruling troika.
Brin and Page’s idealistic views of a corporate culture impressed Campbell, but he worried that as the company grew, those values would be diluted, misinterpreted, or ignored as more layers wedged between the founders and a workforce of thousands. At Intuit, a group of employees had compiled a set of corporate values that could be shared both inside and outside the company. Campbell convinced the executives at Google that they should do something similar.
On July 19, 2001, Stacy Sullivan, who had come to Google to run human resources, pulled together a group for that purpose. They gathered in Charlie’s, about fifteen of them from various parts of the company, including David Krane from communications, Paul Buchheit and Amit Patel from engineering, and Joan Braddi, VP of search services. Marissa Mayer was there, as was Salar Kamangar. And Campbell. Page and Brin were not in attendance. Charlie made smoothies. It was an unusual meeting.
Sullivan explained the format. People would identify Google’s values, and she would write down the good ones with a marker on a giant pad she’d set up on an easel. Some of them were straight from the conventional playbooks of management and self-realization, such as “Play hard but keep the puck down.” That was a riff on the twice-weekly roller hockey games that the Googlers played in the parking lot—since no one wore padding, there were frequent reminders not to emasculate anyone with a hard rubber disk. (Minor injuries were nonetheless common.) Another one stipulated, “Google will strive to honor all its commitments.”
As Sullivan scrawled these nostrums on the big pad, Paul Buchheit was thinking, This is lame. Jawboning about citizenship and values seemed like the kind of thing you do at a big company. He’d seen enough of that at his previous job at Intel. At one point the chipmaker had given employees little cards with a list of values you could attach to your badge. If something objectionable came up you were to look at your little corporate values card and say, “This violates value number five.” Lame. “That whole thing rubbed me the wrong way,” Buchheit later recalled. “So I suggested something that would make people feel uncomfortable but also be interesting. It popped into my mind that ‘Don’t be evil’ would be a catchy and interesting statement. And people laughed. But I said, ‘No, really.’”
The slogan made Stacy Sullivan uncomfortable. It was so negative. “Can’t we phrase it as ‘Do the right thing’ or something more positive?” she asked. Marissa and Salar agreed with her. But the geeks—Buchheit and Patel—wouldn’t budge. “Don’t be evil” pretty much said it all, as far as they were concerned. They fought off every attempt to drop it from the list.
“They liked it the way it was,” Sullivan would later say with a sigh. “It was very important to engineering that they were not going to be like Microsoft, they were not going to be an evil company.”
When the meeting ended, “Don’t be evil” was just one of a number of broad statements on an otherwise timid list of values. But Amit Patel felt that when it came to corporate values, that phrase really said it all; follow that commandment, and the rest should flow. Patel, remember, wa
s one of Google’s first engineers. He had been an early keeper of the logs and had focused on the way they could be used to demonstrate Google’s value as a barometer of public interest. Now he had a new crusade. He would imprint the phrase into Google’s corporate subconscious. Making use of the whiteboards that were ubiquitous in the hallways and conference rooms in the Googleplex, he scrawled the phrase over and over in his distinct calligraphic style, a sans-serif, Tolkien-esque script. Amit Patel became Google’s Kilroy.
“He wrote everywhere he could,” says David Krane. “It became this atmospheric, pervasive reminder.”
“It was just an informal sort of reminder that we’re all here to do the right thing,” says Cindy McCaffrey, head of PR at the time. “Everybody felt good about it, especially the engineers. It meant, ‘Look, out in the world, there are all kinds of companies doing evil things and we have an opportunity here to always do the right thing.’”
It had a powerful effect within the company. Even in the kingdom of data, there was one thing that you could go on by gut: what was evil and what was not. The concept could impinge on your consciousness in small ways. You might be in a microkitchen eyeing someone else’s leftovers in the fridge and then see the little note saying “Don’t be evil.” And, says David Krane, “You realize, it can mean, ‘Don’t take someone’s food that looks appealing.’” But it also applied to much bigger things, like maintaining a stiff line between advertising and search results, or protecting a user’s personal information, or—much later—resisting the oppressive measures of the Chinese government.
For months, “Don’t be evil” was like a secret handshake among Googlers. An idea would come up in a meeting with a whiff of anticompetitiveness to it, and someone would remark that it sounded … evil. End of idea. “Don’t be evil” was a shortcut to remind everyone that Google was better than other companies. Since the slogan was internal, no outsiders were talking about it. But then Eric Schmidt revealed Google’s internal motto to a reporter from Wired. To McCaffrey, that was the moment when “Don’t be evil” got out of control and became a hammer to clobber Google’s every move. “We lost it, and I could never grasp it back,” she says. “Everybody would’ve been happy if it could’ve been this sort of silent code or little undercurrent that we secretly harbored instead of this thing that set us up for a lot of ridiculous criticism.” Elliot Schrage, who was in charge of communications and policy for Google from 2005 to 2008, concluded that “Don’t be evil” might originally have benefited the company but became “a millstone around my neck” as Google’s growth took it to controversial regions of the world.
Nonetheless, most people at Google continued to take pride in being associated with that risky admonition. “It’s easy to take a cheap shot at them with that as a motto,” says John Doerr. “But I think it’s served them well.” Doerr believes that the meme is so deeply implanted in the Google ethos because the rule became internalized. You wouldn’t hear it much in the boardroom, he says, because “it doesn’t need to be said—it’s implicit.”
Alan Eustace, Google’s director of engineering, believed that the motto simply reflected what’s in the souls of Googlers: “I look at people here as missionaries—not mercenaries,” he says.
In any case, the founders themselves embraced “Don’t be evil” as a summation of their own hopes for the company. That was what Google was about: two young men who wanted to do good, gravitated to a new phenomenon (the Internet) that promised to be a history-making force for good, developed a solution that would gather the world’s information, level the Tower of Babel, and link millions of processors into a global prosthesis for knowledge. And if the technology they created would make the world a better place, so would their company; Google would be a shining beacon for the way corporations should operate: an employee-centric, data-driven leadership pampering a stunningly bright workforce that, for its own part, lavished all its wit and wizardry on empowering users and enriching advertising customers. From those practices, the profits would roll in. Ill intentions, flimflammery, and greed had no role in the process. If temptation sounded its siren call, one could remain on the straight path by invoking Amit Patel’s florid calligraphy on the whiteboards of the Googleplex: “Don’t be evil.” Page and Brin were good, and so must be the entity they founded.
Which explains why Larry Page and Sergey Brin made the expression into the centerpiece of the initial public offering that transformed Google from a clever Internet start-up into a corporate phenomenon.
Neither Page nor Brin wanted to go public. The idea of hewing to the complicated reporting protocols of a public corporation was anathema to the secretive Page. And during the time when going public became a virtual imperative—in early 2004—the “hiding strategy” was still in effect. Talking to a reporter at that time, Page and Brin insisted that an IPO was not a foregone conclusion. “I think there’s always the opportunity to screw it up, be it private or public,” Brin said. “Perhaps I’m naïve, but I think that we could maintain Google being private or public.”
But by that point, going public was inevitable. The previous year, Google had begun its path toward the public offering that was destined to be a Silicon Valley milestone. The venture capitalists, as one would expect, were forceful advocates; selling stock on the marketplace was their means of realizing the huge payback their investment had reaped on paper. And Page and Brin’s arguments against the move had become progressively weaker. The attractions of remaining private were stealth and control. But regulations required a company with more than five hundred shareholders to reveal financial information anyway. In 2004, Google crossed that line. In any case, many of Google’s employees deserved the opportunity to convert some of their own options to cash. It was almost sadistic to deny them.
Google would go public. But Larry and Sergey would do it their way. The process played out as a slow-motion conflict. It was the values of Google squaring off against the values of Wall Street, which embodied everything its founders despised about tradition-bound, irrational corporate America.
The first order of business was making sure that outside shareholders (who almost by definition would not be as smart as Googlers) would never be able to overrule LSE’s decisions. Their model was Warren Buffett’s Berkshire Hathaway, which was the most prominent example of the dual class ownership structure. In Google’s case, ordinary investors would buy class A stock, which counted as one vote per share. Class B stock, restricted to founders, directors, and owners, would have a weight of ten votes a share. The vast majority of class B stock was owned by Page, Brin, and Schmidt. This way Brin and Page could maintain control even if their combined shares fell well short of 50 percent of the total.
Google warned investors that it intended to ignore short-term gains in favor of enduring value. Google also said that the information it provided every quarter would adhere to the minimum required by the law and generally be much less than other companies provided. In short, if you bought Google, you were taking a flier on its leaders. These specifications did not please VCs John Doerr and Mike Moritz—in theory, they would make shares less valuable to investors—but they accepted them. Even with those restrictions, Google’s IPO would easily reap over a billion-dollar profit for each VC’s fund.
Brin and Page decided that the IPO would be conducted by auction. Their impulses were both egalitarian and financial. In a typical IPO, the opening price is set much lower than the market would dictate. Opportunities to buy shares are available only to insiders—people connected to the investment banks organizing the offering or the company itself. Within hours the shares reach their true market price, often many times what the insiders paid for them. (When Netscape had gone public in 1995, shares had opened at $28 and been priced at $75 by the end of day.) That was unfair to the general public but also penalized the company, which wound up receiving less than the true value of its shares. In theory, an auction would eliminate those problems. The concept had been used previously, but never for a publi
c offering the size of Google’s, which would certainly sell over a billion dollars’ worth of stock.
Eric Schmidt later said the tipping point for this decision was a letter “from a little old lady” who was griping in advance that whenever she tried to invest in an IPO, stockbrokers would get there first and she’d be shut out. But the real lure was the math of it. Logically, it seemed to be a better approach, and that meant a lot at a data-driven operation like Google. “We’re an auction company—we’re going to run an auction,” Schmidt would later say, as if it were no big deal.
In 2003, Google started hiring people whose skill set was geared toward guiding a company through the IPO process. One was Lise Buyer, a former investment banker who’d moved to the world of Silicon Valley venture capital. She worked with Google’s chief finance officer, George Reyes.
In early 2004, Google began contacting investment banks. The company felt it had the leverage to make sure that the ones involved would be on its wavelength. It began by limiting the field to investment banks that had sent it queries about an IPO before October 20, 2003. Those banks each received a solicitation asking for detailed answers to twenty questions on how they would handle the offering. It was Google’s equivalent of the show-your-stuff essay that prospective students file with a college. Some bankers were offended at having to explain themselves and immediately got on the phone to try to talk their way out of committing themselves on paper. No waivers were granted.