Ken Auletta - Googled - The End of the World as We Know It

A meeting on October 9, 2007, did not quite follow this pattern. Brin and Page were to meet with an engineering team to review their proposal for an upgrade of AdWords 1.0. Since its introduction in early 2002, some parts of AdWords had been substantially upgraded while others had not. Small businesses complained that the system was too complicated. Larger customers, such as eBay or Amazon, complained that they wanted new features, including an ability to organize their accounts by products and to break out expenditures by country. To make these functions work, Google needed to enlarge its computers that retained data and enhance the speed of the advertising auctions. To demonstrate their commitment to a new architecture, the founders decided to skip 2.0 and christened this effort AdWords 3.0. The purpose of this session was to receive, and review, the new product teams’ recommendations.

Everyone around the conference table sat on gray-mesh ergonomic swivel chairs. Page was wearing his usual jeans, and a gray T-shirt under a black sports jacket; he sat in the middle of the table, a coffee cup in hand. Brin arrived a few minutes late in jeans and a black crewneck sweater, and plopped in the seat beside Page. More nattily attired in a blue V-necked sweater over a light blue dress shirt and gray slacks, Schmidt sat at the head of the table with a translucent container of salad and a Diet Coke. Schmidt opened the meeting by calling on the team leader, vice president of engineering Sridhar Ramaswamy, to describe the teams’ recommendations.

The upgraded system they proposed, said Ramaswamy, would be less complicated for advertisers, would produce search results faster, and would be “scalable” in that it would allow for the retention of more data. But, he cautioned, it was not quite the gut renovation that had been requested; it would be too expensive and require diverting too many engineers to both speed up AdWords, as Page had urged them to, and to make the sweeping computer changes needed to accommodate Page’s database growth projections.

To a nonengineer, the hour-long discussion was often incomprehensible-“three-tier architecture,” “middle-tier API,” “UI tier,” “end-to-end solutions,” “no latency,” “Java script bindings for third parties,” “10 percent CTR,” “SQL base.” But no translator was required to observe that Page and Brin were unhappy. At first, the founders were stonily silent, sliding lower in their chairs, and occasionally leaning over to whisper to each other. Intermittently, Page looked away from the engineers; Brin, appearing alternately distracted and irritated, would rise and stretch his legs on the empty chair beside his. Schmidt began with technical questions to the product team, but then he switched roles and tried to draw out Page and Brin, saying, “Larry, say what’s really bugging you.”

The room was quiet for perhaps ten seconds before Page responded. When he did, he scolded the engineers, saying they were not ambitious enough. Brin concurred, adding that the proposal was “muddled” and un-Google-like in its caution. “I named this 3.0 for a reason,” Page interjected. “We wanted something big. Instead, you proposed something small. Why are you so resistant?”

The engineering team leader held his ground. Ramaswamy said that his entire team concurred that the founders’ proposed changes would be too costly in money, time, and engineering manpower. Page countered that a significantly improved AdWords would make it easier for advertisers and result in greater revenues. “You are polishing up the program. I wanted to have a redesign.”

Schmidt stepped in to summarize their differences. He noted that Brin and Page were focused on the outcome, while the product team focused first on the process, and concluded that the engineering improvements would prove too “disruptive” to achieve the goal.

Brin said that neither he nor Page wanted to add patches to the system, something Microsoft has been criticized for when they stuff more code into their already bloated operating system. “I’m just worried that we designed the wrong thing,” Brin said. “And you’re telling me you’re not designing the optimum system. I think that’s a mistake… I’m trying to give you permission to be bolder.”

Schmidt achieved a cease-fire by asking the product team to make its slide presentation. It demonstrated how the new product would actually work for advertisers, allowing them to manage their accounts. The discussion now went round and round with Schmidt finally stepping in to summarize the technical changes that would be made, the engineering challenges, the different approaches proposed by the team and by the founders. As he spoke, I kept wondering: When Terry Semel was CEO of Yahoo, or John Sculley CEO of Apple, could either of those nonengineers understand what their engineers were saying? Could they challenge them? (I would ask this question of Semel, who said the Yahoo founders, Jerry Yang and David Filo, both engineers, often accompanied him to similar meetings. Besides, he said, “I’d make people describe things in English!”) Semel brought good judgment and people skills to Yahoo when he arrived in 2000, but the question begs to be asked: Did Yahoo slip technologically because the CEO could not wrap his brain around the technology? Was that why Apple slipped technologically after Steve Jobs had been fired? (While Jobs does not possess an engineering degree, he seems not to need a translator.)

Schmidt, sensing that a resolution was not possible at this Google meeting, told the product team to report back with a detailed design “which is responsive to Larry and Sergey’s criticism,” one that laid out “what it takes to build a good product,” and what it would cost in time and money. He took care to balance this rebuke with praise: “But this is very well done. I love it when people show me the flaws in our products.”

Neither founder was happy after the meeting. “I hope they try to do something a little more ambitious,” Brin said two days later. He compared the project to renovating a house. “Once you get into it, you know it’s going to take some time and effort, so you may as well do as good a job as you can. We prefer not to do too many small things when we know where we’d like to get to.” Page was disappointed in what he described as the engineering team’s “self-imposed, bureaucratic response.” He sounded harsh, and a few seconds later he softened his words: “It’s hard when you’re so focused to see the big picture. It’s sort of easy for us. We just say, ‘If you’re going to make changes at that rate, we’re going to go out of business. It’s just not OK. It’s all of our revenue. We’re obviously doing some things wrong. We need some sort of reasonable plan to fix these things in our lifetime. Our lifetime means years, not multiple years.’” Ultimately, Ramaswamy and his team came back with an AdWords 3.0 proposal that went more than halfway toward the one proposed by the founders; Google has been rolling out the new system in stages.

THE MEETING DEMONSTRATED that the ethos that had launched the Google rocket-to shoot for the moon, not the tops of trees-was intact, no matter how much the company had grown. Page and Brin’s passion for technology was apparent, as was the way they push engineers to act boldly. At meetings they feed off each other, punishing engineers and product managers who think they have devised a “new” solution when, the founders say, they have merely devised a “cute” solution, not a fundamental one. Or as Schmidt said, “They think about what should be, and they assume it is possible.”

Page describes his and Brin’s role as supplying the “big picture,” and by way of illustrating what he sees as a management rather than a technological innovation, he cites the work of Gordon Moore at Intel and his Moore ’s law. “People think it’s this wild statement about how the universe is, but it’s actually a management innovation. Moore ’s law was a statement saying, ‘We’re going to double the performance [of integrated circuits or computer chips] every eighteen months, and let’s get organized to do it.’ They spent billions of dollars doing that. If you didn’t have Moore ’s law, you wouldn’t have that advancement. It’s actually causal in another way.” The management pressure to double performance helps assure it.