Battle of the Titans: How the Fight to the Death Between Apple and Google is Transforming our Lives. Fred Vogelstein
other small acquisitions. In those the founders rarely stayed, quickly discovering that actually working at Google was frustrating. Google often bought companies just to test out a new technology and/or hire talented engineers, but without a clear game plan. Page didn’t want Rubin to become frustrated like that, and he specifically tasked executives such as Alan Eustace—who helped Page negotiate the purchase of Android—to make sure Rubin felt that he had the access to the people and resources he needed. Google immediately opened its wallet to the tune of $10 million to help Rubin buy necessary software licenses. Schmidt personally helped negotiate some of them. To ensure the secrecy of their project, the Android team was allowed to keep its software code separate from the rest of Google, and inaccessible to anyone without Rubin’s permission. Page gave Rubin the rare privilege of being able to hire his own staff, instead of going through Google’s famously rigorous and lengthy hiring process.
But all this attention didn’t spare Rubin from having to navigate Google’s wacky politics. For starters, it wasn’t clear to him for a while who Google’s ultimate boss was. Eric Schmidt was the CEO and played a critical role in helping Google deal with its hypergrowth back then. He was also the public face of the company, which he did well and which Page and Brin had much less interest in doing. He had been a CEO before—at Novell—and an executive at Sun Microsystems for fourteen years before that. But Schmidt, who joined Google in 2001, was not a founder as were Page and Brin, which made his true role slightly murkier.
Officially, the three33 ran Google as a triumvirate, but Google employees debated about how much power Schmidt actually had—whether Brin and Page called the shots, with Schmidt filling a largely ceremonial role, providing “adult supervision” in Silicon Valley parlance. Schmidt didn’t help with this confusion by describing his job the way a chief operating officer would, not a CEO. In an interview with me in 2004 he said,
My primary responsibility is making the trains run on time, so I try to make sure that the meetings happen, that all of the functions of a properly running company are in place and people are paying attention. Larry and Sergey have driven the top-level strategy and much of the technology strategy. I contribute by organizing the strategy process, but it’s really their strategy and their technology strategy. And if there is a disagreement among the three of us … we’ll have a significant conversation, and somebody will eventually say yes. A few months later somebody, one of the three, will say, ‘Well, maybe the other guy was actually right.’ So there’s a very healthy respect now between the three of us and it’s a wonderful thing. We’re best friends and we’re very good colleagues.
Rubin also noted to colleagues that it seemed to him as if Page and Schmidt didn’t completely agree on what Android should become. Schmidt wanted Android to be software only, and for a while he wondered if it should just be low-level software, without fancy graphics or animations. This was Rubin’s original vision: give phone makers and carriers code that runs all phones and applications the same way, but allows them to decide things such as what the opening screen would look like, and what kinds of graphic flourishes each phone would have. Page, however, was more interested in having Google build a phone. “I remember talking to Andy about this,” one Android executive told me. “He said he always made sure never to demonstrate an Android feature to Page without a prototype of the actual hardware it would run on.”
Then there were legal issues34. Most of Android was open-source software, meaning no one owned it, and the code could be modified by anyone in any way. But not all of it was open source, and Google negotiated licenses for those portions for tens of millions of dollars. Rubin hoped a big chunk of licensed code would come from Sun Microsystems, makers of Java. Sun had spent ten years building Java as an alternative to Microsoft’s Windows. It typically gave the software away for free, on the condition the user didn’t modify it in a major way. Rubin used it for the operating system on the Sidekick, and it was, back then, a widely used language by engineers coming out of top universities. But Android wanted to modify Java more than Sun would allow. No amount of money seemed able to move Sun from this view. Payments as high as $35 million were discussed. This created two problems for Rubin: Without the Java code, Rubin had to spend months of extra time creating a work-around. Second, it infuriated Sun, which believed Google had copied portions of Java to build the work-around. It ultimately became the focus of a messy lawsuit that went to trial in 2012. Google was not held liable, but Sun, now owned by Oracle, is appealing that decision.
Finally, Rubin had the enormous task of just doing what he’d promised to do: build a mobile phone operating system that carriers and manufacturers would want to use and that software developers in addition to Google would want to write programs for. There certainly was precedent for it. It’s what Bill Gates did to transform the PC industry and become the world’s richest man. Most of us now just assume that any PC we buy will run Microsoft Windows or Apple OS X, and that it will have an Intel processor running on a certain kind of circuit board that connects to every printer, mouse, keyboard, monitor, and almost every other electronic device. But in the 1980s, the PC industry was just like the mobile industry in 2005. Not until Gates came along and used MS-DOS and Windows to create a platform for developers to write to did the PC application business take off. “I remember at the time telling Andy, ‘This is going to be really hard, really, really difficult. I don’t want you to get discouraged, but I think the chances are low on this,’” said Alan Eustace, Google’s head of engineering and Rubin’s boss at the time. “And then he and I would laugh about that because he was a true believer. It wasn’t that I was a skeptic. I supported the project all the time. It’s just that we both knew it was going to be hard.”
Some of the issues in building Android were similar to the ones Apple faced. Few people had put an operating system as sophisticated as Android on a phone chip. Meanwhile, all the testing had to be done on simulators because the actual chips and displays Rubin wanted to put on the Dream phone weren’t going to be manufactured for another year. But Google was in an even worse position than Apple to take on these challenges. At Apple the iPhone had nearly brought the company to its knees, but at least Apple was used to building things that consumers wanted to buy. Google had no such experience. Google made money selling advertising. Everything else it built—web software—it gave away for free. It had no fancy industrial-design division comparable to Apple’s. Indeed, the idea of a finished product of any sort was anathema to Googlers. To them the beauty of building web software was that it was never finished. When a feature was mostly done, Google would release it, then refine it over time based on consumer usage with updates to their servers.
Google also viewed marketing with the kind of contempt only an engineer could muster. If a product was good, word of mouth on the web would get people to use it. If it was not good, people would not use it. The idea that Google might be selling more than just a cool phone, but amorphous feelings of satisfaction and self-confidence—the way Jobs sold Apple’s devices—seemed silly. This thinking was firmly rooted35 in Google’s DNA as a corporation. In Google’s early years executives had hired the famed consultant Sergio Zyman—the former head of marketing for Coca-Cola—to draw up a plan to get the world excited about their new company. After he spent months working on a plan, the founders rejected the whole marketing concept and did not renew Zyman’s contract. They believed—correctly—that Google’s search engine would sell itself. Google didn’t even have a director-of-marketing position until 2001.
Rubin and the Android team believed they could compensate for these deficits by partnering with wireless carriers and phone makers. That was the whole point of Android, after all: everyone would do what they did best. Google would write software, manufacturers would make phones, and carriers would supply bandwidth and sales and marketing heft. HTC and T-Mobile were committed to the project. They had helped Rubin build the Sidekick when he was at Danger.
Rubin’s problem was that T-Mobile wasn’t a big enough carrier in the United States to get Android on enough phones, and the two other big US wireless carriers, AT&T and Verizon, were deeply suspicious of anyone from Google interested in a business deal. For all Android’s promise and Rubin’s ability to sell that promise, by the end of 2006 the rest of Google was starting to scare people, telecom companies