5.

Google Search

Dad, what happens when you die?”

“I don’t know, son. Nobody knows for sure.”

“Well, why don’t you ask Google?”

Evidently, it’s possible for children to grow up with the impression that Google knows everything. Perhaps that’s to be expected. “Dad, how far is the moon from the Earth?” “What’s the biggest fish in the world?” “Do jetpacks really exist?” All of these questions are efficiently answered with the tap of a touchscreen. No need to visit the library to consult the Encyclopædia Britannica, Guinness World Records, or—well, who knows how a pre-Google parent would have discovered the state of the art in jetpack technology? It wouldn’t have been straightforward.

Google may not be clever enough to know if there’s life after death, but the word “Google” does crop up in conversation more often than either “clever” or “death,” according to researchers at the UK’s University of Lancaster.1 It took just two decades for Google to reach this cultural ubiquity, from its humble beginnings as a student project at Stanford University.

It’s hard to remember just how bad search technology was before Google. In 1998, for instance, if you typed “cars” into Lycos—then a leading search engine—you’d get a results page filled with porn websites.2 Why? Owners of porn websites inserted many mentions of popular search terms like “cars,” perhaps in tiny text or in white on a white background. The Lycos algorithm saw many mentions of “cars” and concluded that the page would be interesting to someone searching for cars. It’s a system that now seems almost laughably simplistic and easy to game. But at the time, nothing better was available.

Larry Page and Sergey Brin were not, initially, interested in designing a better way to search. Their Stanford project had a more academic motivation. In academia, how often a published paper is cited is a measure of how much credibility it has; and if it’s cited by papers that themselves are cited many times, that bestows even more credibility. Page and Brin realized that when you looked at a page on the nascent World Wide Web, you had no way of knowing which other pages linked to it. Web links are analogous to academic citations. If they could find a way to analyze all the links on the Web, they could rank the credibility of each webpage in any given subject.

To do this, Page and Brin first had to download the entire Internet. This caused some consternation. It gobbled up nearly half of Stanford’s bandwidth. Irate webmasters showered the university with complaints that the students’ project was overloading their servers. An online art museum thought Stanford was trying to steal its content and threatened to sue. But as Page and Brin refined their algorithm, it soon became clear that they had stumbled on a new and vastly better way to search the Web.3 Put simply, porn websites with tiny text saying “cars cars cars” don’t get many links from other websites that discuss cars. If you searched Google for “cars,” its analysis of the Web’s network of links would be likely to yield results about . . . cars.

With such an obviously useful product, Page and Brin attracted investors, and Google went from student project to private company. It’s now among the world’s biggest, bringing in annual profits by the tens of billions.4 But for the first few years, Page and Brin burned through money without much idea about how they’d ever make it back. They weren’t alone. This was the time of the dot-com boom and bust—shares in loss-making Internet companies traded at absurd prices, based purely on the hope that eventually viable business models would appear.5

It was 2001 when Google found its viability, and in retrospect it seems obvious: pay-per-click advertising. Advertisers tell Google how much they’ll pay if someone clicks through to their website, having searched for terms they specify. Google displays the highest bidders’ ads alongside its “organic” search results. From an advertiser’s perspective, the appeal is clear: you pay only when you reach people who have just demonstrated an interest in your offering. (Try googling “what happens when you die”; there’s an advertiser willing to pay Google for your click-through: the Mormons.) That’s much more efficient than paying to advertise in a newspaper; even if its readership matches your target demographic, inevitably most people who see your advertisement won’t be interested in what you’re selling. No wonder newspaper advertising revenue has fallen off a cliff.6

The media’s scramble for new business models is one obvious economic impact of Google search. But the invention of functional search technology has created value in many ways. A few years ago, consultants at McKinsey tried to list the most important.7

There are time savings. Studies suggest that googling is about three times as quick as finding information in a library, and that’s before you count the time spent traveling to the library. Likewise, finding a business online is about three times faster than using a traditional, printed directory such as the Yellow Pages. McKinsey put the productivity gains into the hundreds of billions.

Another benefit is price transparency—that’s economist jargon for being able to stand in a store, take out your phone, google a product you’re thinking of buying to see if it’s available more cheaply elsewhere, and use that knowledge to haggle. Annoying for the store, helpful for the customer.

Then there are “long tail” effects. In physical stores, space is at a premium. Online stores can offer more variety—but only when the search engines are good enough to enable customers to find what they need. Online shopping with a search that works means customers with specific desires are likelier to find exactly what they want rather than having to settle for the closest thing available locally. And it means entrepreneurs can launch niche products, more confident that they’ll find a market.

This all sounds like excellent news for consumers and businesses. But there are problems.

One problem is those advertisements. Typically they function as one might expect—if you google “craft beer,” then you’ll get advertisements for craft beer. But if you’ve lost your wallet or need to rebook a flight or need a locksmith, you’re a prime candidate for an AdWords scam. Fly-by-night companies will pay Google handsomely to appear next to certain kinds of searches; they’ll gladly take your money and provide little or no service in return. Some are outright fraudulent; others cleverly tiptoe up to the line without crossing it. Nor is it clear how hard Google is trying to stamp out these dubious tactics.8

Perhaps the bigger issue is that this seems to be Google’s responsibility alone, because the company dominates the search market. Google handles close to 90 percent of searches worldwide; many businesses rely on ranking high in its organic search results.9 And Google constantly tweaks the algorithm that decides them. Google gives general advice about how to do well, but it isn’t transparent about how it ranks results. Indeed, it can’t be: the more Google reveals, the easier it is for scammers to fool its algorithms. We’d be back to searching for cars and getting porn.

You don’t have to look far online (starting with Google, naturally) to find business owners and search strategy consultants gnashing their teeth over the company’s power to make or break them. If Google thinks you’re employing tactics it considers unacceptable, it downgrades your ranking in the search results. One blogger complained that Google is “judge, jury and executioner . . . you get penalized on suspicion of breaking the rules [and] you don’t even know what the rules are, you can only guess.”10 Trying to figure out how to please Google’s algorithm is like trying to appease an omnipotent, capricious, and ultimately unknowable deity.

You may say this isn’t a problem. As long as Google’s top results are useful to searchers, it’s tough luck on those who rank lower. And if those results stop being useful, then some other pair of students from Stanford will spot the gap in the market and dream up a better way. Right?

Maybe—or maybe not. Searching was a competitive business in the late 1990s. But now Google may have an entrenched monopoly. The reason? Among the best ways to improve the usefulness of search results is to analyze which links were ultimately clicked by people who previously performed the same search, as well as what the user has searched for before.11 Google has far more of that data than anyone else. That suggests it may continue to shape our access to knowledge for generations to come.