Inside the Googleplex
Aug 30th 2007 | SAN FRANCISCO
From The Economist print edition
It is rare for a company to dominate its industry while claiming not to be motivated by money. Google does. But it has yet to face a crisis
IN AMERICA a phenomenon might claim to have entered mainstream culture only after it has been satirised on “The Simpsons”. Google has had that honour, and in a telling way. Marge Simpson types her name into Google's search engine and is amazed to get 629,000 results. (“And all this time I thought ‘googling yourself' meant the other thing.”) She then looks up her house on Google Maps, goes to “satellite view” and zooms in. To her horror, she sees Homer lying naked in a hammock outside. “Everyone can see you; get inside,” she yells out of the window, and the fumbling proceeds from there.
And that, in a nutshell, sums up Google today: it dominates the internet and guides people everywhere, such as Marge, to the information they want. But it also increasingly frightens some users by making them feel that their privacy has been intruded upon (though Marge, technically, could not have seen Homer in real time, since Google's satellite pictures are not live). And it is making enemies in its own and adjacent industries. The grand moment of Marge googling herself, for example, was instantly available not only through Fox, the firm that created the animated television show, but also on YouTube, a video site owned by Google, after fans uploaded it in violation of copyright.
Google evokes ambivalent feelings. Some users now keep their photos, blogs, videos, calendars, e-mail, news feeds, maps, contacts, social networks, documents, spreadsheets, presentations, and credit-card information—in short, much of their lives—on Google's computers. And Google has plans to add medical records, location-aware services and much else. It may even buy radio spectrum in America so that it can offer all these services over wireless-internet connections.
Google could soon, if it wanted, compile dossiers on specific individuals. This presents “perhaps the most difficult privacy issues in all of human history,” says Edward Felten, a privacy expert at Princeton University. Speaking for many, John Battelle, the author of a book on Google and an early admirer, recently wrote on his blog that “I've found myself more and more wary” of Google “out of some primal, lizard-brain fear of giving too much control of my data to one source.”
Google itself has been genuinely taken aback by such sentiments. The Silicon Valley company, which trumpeted its corporate motto, “Don't be evil”, before its stockmarket listing in 2004, considers itself a force for good in the world, even in defiance of commercial logic. Its founders, Larry Page and Sergey Brin, and Eric Schmidt, its chief executive, have said explicitly and repeatedly that their biggest motivation is not to maximise profits but to improve the world.
Too many sermons
Such talk can make outsiders wince. Book and newspaper publishers, media companies such as Viacom, businesses which depend on Google's search rankings and a lengthening queue of others are tired of moralising sermons. Some feel their own livelihoods are threatened and are suing Google. Even some employees (called “Googlers”) or former employees (“Xooglers”) are cynical. Google is “arrogant” because it feels “invincible”, says a Xoogler who left to run a start-up firm. The internal attitude towards customers, rivals and partners is “you can't stop us” and “we will crush you”, he says. That “kinder, gentler” image is “mythology” and, he reckons, Google gets away with it only because of its impressively high share price.
That share price has quintupled since 2004, making Google worth $160 billion. The company has not yet had its tenth birthday. Yet Piper Jaffray, an investment bank, expects it to have revenues of $16 billion and profits of $4.3 billion this year. With so much money pouring in sceptics say it is easy to ignore shareholders and talk about doing good instead of doing well. But what happens when earnings fall short of Wall Street expectations or some other disaster strikes? Yahoo! and other rivals have gone through such crises and been humbled. Google has not.
Fifty cents at a time
Google's success still comes from one main source: the small text ads placed next to its search results and on other web pages. The advertisers pay only when consumers click on those ads. “All that money comes 50 cents at a time,” says Hal Varian, Google's chief economist. For this success to continue, several things need to happen.
First, Google's share of web searches must remain stable. Thanks to its brand, this looks manageable. Google's share has steadily increased over the years. It was about 64% in America in July, according to Hitwise. That is almost three times the volume of its nearest rival, Yahoo!. In parts of Europe, India and Latin America, Google's share is even higher. Only in South Korea, Japan, China, Russia and the Czech Republic does it trail local incumbents.
Second, Google must maintain or improve the efficiency with which it puts ads next to searches. And here its dominance is most impressive. In a recent analysis by Alan Rimm-Kaufman, a marketing consultant, it took a whopping 73% of the budgets of companies that advertise on search engines (versus 21% and 6%, respectively, for Yahoo! and Microsoft). It charged more for each click, thanks to its bigger network of advertisers and more competitive online auctions. And it had far higher “click-through rates”, because it made these ads more relevant and useful, so that web users click on them more often.
Perhaps most tellingly, advertisers do better with Google. Mr Rimm-Kaufman found that Google's ads “converted” more often into actual sales, which tended to be larger than those originating from Yahoo! or Microsoft. This is astonishing, given that Yahoo! has just spent a year on an all-out effort, codenamed Panama, to close precisely these gaps.
But even lucrative “pay-per-click” has limits, so Google is moving into other areas. It is trying (pending an antitrust inquiry) to buy DoubleClick, a firm that specialises in the other big online-advertising market, so-called “branded” display or banner ads (for which each view, rather than each click, is charged for). And Google now brokers ads on traditional radio stations, television channels and in newspapers of the dead-tree sort.
Sceptics point out that with each such expansion, Google reduces its profit margins, because it must share more of the revenues with others. If a web surfer clicks on a text ad placed by Google on a third-party blog, for instance, Google must share the revenue with the blogger. If Google places ads in newspapers or on radio stations, it must share the revenues with the publisher or broadcaster.
Yet Google does not look at it that way. Its costs are mostly fixed, so any incremental revenue is profit. It makes good sense for Google to push into television and other markets, says Mr Varian. Even if Google gets only one cent for each viewer (compared with an average of 50 cents for each click on the web), that cent carries no variable cost and is thus pure profit.
The machinery that represents the fixed costs is Google's secret sauce. Google has built, in effect, the world's largest supercomputer. It consists of vast clusters of servers, spread out in enormous datacentres around the world. The details are Google's best-guarded secret. But the result, explains Bill Coughran, a top engineer at Google, is to provide a “cloud” of computing power that is flexible enough “automatically to move load around between datacentres”. If, for example, there is unexpected demand for Gmail, Google's e-mail service, the system instantly allocates more processors and storage to it, without the need for human intervention.
This infrastructure means that Google can launch any new service at negligible cost or risk. If it fails, fine; if it succeeds, the cloud makes room for it. Thus Google can redefine its goals almost on a whim. Its official strategy recently became “search, ads, and apps”—the addition being the apps (ie, software applications). Sure enough, after a string of acquisitions, Google now offers a complete alternative to Microsoft's entrenched Office suite of programs, all accessible through any web browser. A new technology, called Google Gears, will make these applications usable even when there is no internet connection. And Google is hawking these applications not only to consumers but also to companies. Ultimately it does so because, thanks to its supercomputer, it can.
With Google's cashflow and infrastructure, the freedom to do anything it fancies gives rise to constant rumours. Often, these are outrageous. It used to be conventional wisdom that Google would build cheap personal computers for poor countries. This turned out to be nonsense, because Google does not want to make hardware. Now there is talk of a “Gphone” handset. This is also unlikely because Google is more interested in software and services, and does not want to alienate allies in the handset industry—including Apple, which shares board directors with Google and uses Google software on its iPhone.
Sometimes the rumours are both outrageous and true. Google is experimenting with new ways of bringing broadband connections to consumers, by blanketing parts of Silicon Valley with Wi-Fi networks. It is planning to enter an auction for valuable radio spectrum in America, and thinking of radically new business models to make money from wireless data and voice networks, perhaps a free service supported by ads.
If it goes wrong, how?
Beyond its attempts to expand into new markets, the big question is how Google will respond if its stunning success is interrupted. “It's axiomatic that companies eventually have crises,” says Mr Schmidt. And history suggests that “tech companies that are dominant have trouble from within, not from competitors.” In Google's case, he says, “I worry about the scaling of the company.” Google has been hiring “Nooglers” (new Googlers) at a breathtaking rate. In June 2004 it had 2,292 staff; this June the number had reached 13,786.
Its ability to get all these people has been a competitive weapon, since Google can afford to hire talent pre-emptively, making it unavailable to Microsoft and Yahoo!. Google tends to win talent wars because its brand is sexier and its perks are fantastically lavish. Googlers commute on discreet shuttle buses (equipped with wireless broadband and running on biodiesel, naturally) to and from the head office, or “Googleplex”, which is a photogenic playground of lava lamps, volleyball courts, swimming pools, free and good restaurants, massage rooms and so forth.
Yet for some on the inside, it can look different. One former executive, now suing Google over her treatment, says that the firm's personnel department is “collapsing” and that “absolute chaos” reigns. When she was hired, nobody knew when or where she was supposed to work, and the balloons that all Nooglers get delivered to their desks ended up God knows where. She started receiving detailed e-mails “enforcing” Google's outward informality by reminding her that high heels and jewellery were inappropriate. Before the corporate ski trip, it was explained that “if you wear fur, they will kill you.”
Google is a paradise only for some, she argues. Employees who predate the IPO resemble aristocracy. Engineers get the most kudos, people with other functions decidedly less so. Bright kids just out of college tend to love it, because the Googleplex in effect replaces their university campus—with a dating scene, a laundry service and no reason to leave at weekends. Older Googlers with families tend to like it less, because “everybody, even young mums, works seven days a week.”
Another Xoogler, who held a senior position, says that by trying to create a “Utopia” of untrammelled creativity, Google ended up with “dystopia”. As is its wont, Google has composed a rigorous algorithmic approach to hiring, based on grade-point averages, college rankings and endless logic puzzles on whiteboards. This “genetic engineering of their workforce,” he says, means that “everybody there is a rocket scientist, so everybody is also insecure” and the back-stabbing and politics are reminiscent of an average university's English department.
Then there is the question of what all these people are supposed to do. “We kind of like the chaos,” says Laszlo Bock, the personnel boss. “Creativity comes out of people bumping into each other and not knowing where to go.” The most famous expression of this is the “20% time”. In theory, all Googlers, down to receptionists, can spend one-fifth of their time exploring any new idea. Good stuff has indeed come out of this, including Google News, Gmail, and even those commuter shuttles and their Wi-Fi systems. But it is not clear that the company as a whole is more innovative as a result, as it claims. It still has only one proven revenue source and most big innovations, such as YouTube, Google Earth and the productivity applications, have come through acquisitions.
In practice, the 20% time works out to be 120% time, says another Xoogler, “since nobody really gets around to those projects for all their other work.” The chances of ideas being executed, he adds, “are basically zero.” What happens to the many Googlers whose ideas are rejected? Once their share options are fully vested they consider leaving. The same phenomenon changed Microsoft in the 1980s, when allegedly T-shirts popped up saying FYIFV (“Fuck you, I'm fully vested”). Already some are going to even “cooler” start-ups, such as Facebook or Twitter.
This week George Reyes, Google's finance chief, said he would retire. At 53, he is a multi-millionaire. Mr Reyes has maintained the company's policy of not providing guidance to Wall Street on future earnings, although his comments on growth prospects have moved its share price.
As Nick Leeson was to Barings...
Besides the slow risk of calcification that comes with growth, there is also the risk that Nooglers will dilute Google's un-evil values. Worse, Google might inadvertently pick up a rogue employee, as the late Barings Bank notoriously did with Nick Leeson. Indeed, Google is fast becoming something like a bank, but one that keeps information rather than money. This applies equally to its rivals, but Google is accumulating treasure fastest. Peter Fleischer, Google's privacy boss, argues that the risk of a malicious or negligent employee leaking or compromising the data, and thus the privacy of users, is minimal because only a “tiny” number of engineers have access to the databases and everything they do is recorded.
But the privacy problem is much subtler than that. As Google compiles more information about individuals, it faces numerous trade-offs. At one extreme it could use a person's search history and advertising responses in combination with, say, his location and the itinerary in his calendar, to serve increasingly useful and welcome search results and ads. This would also allow Google to make money from its many new services. But it could scare users away. As a warning, Privacy International, a human-rights watchdog in London, has berated Google, charging that its attitude to privacy “at its most blatant is hostile, and at its most benign is ambivalent”.
At the other extreme, Google could decide not to make money from some services—in effect, to provide them as a public benefit—and to destroy data about its users. This would make its services less useful but also less intrusive and dangerous.
In reality, the balance must be struck somewhere in between. Messrs Schmidt, Page and Brin have had many meetings on the subject and have made several changes in recent months. First, says Mr Fleischer, Google has committed itself to “anonymising” the search logs on its servers after 18 months—roughly as banks cross out parts of a credit-card number, say. This would mean that search histories cannot be traced to any specific computer. Second, Google says that the bits of software called “cookies”, which store individual preferences on users' own computers, will expire every two years.
Not everybody is impressed. The server logs will still exist for 18 months. And the cookies of “active” users will be automatically renewed upon expiry. This includes everybody who searches on Google, which in effect means most internet users. Then there is the matter of all that other information, such as e-mail and documents, that users might keep in Google's “cloud”. Mr Schmidt points out that such users by definition “opt in”, since they log in. They can opt out at any time.
As things stand today, Google has little to worry about. Most users continue to google with carefree abandon. The company faces lawsuits, but those are more of a nuisance than a threat. It dominates its rivals in the areas that matter, the server cloud is ready for new tasks and the cash keeps flowing. In such a situation, anybody can claim to be holier than money. The test comes when the good times end. At that point, shareholders will demand trade-offs in their favour and consumers might stop believing that Google only ever means well.
Posts
