- About CEI
- Support CEI
Google is many things -- but not an illegal monopoly
Google is many things -- but not an illegal monopoly
October 25, 2012
Originally published in CNET
The Internet market is notoriously dynamic. Its giants rise and fall far faster than their brick-and-mortar counterparts. This dynamism perplexes and worries many -- especially regulators in Washington, D.C.
Perhaps no Internet leader faces as much scrutiny from government as Google, which has been the subject of a Federal Trade Commission antitrust probe for over a year. As thisinvestigation comes to a close, the government is reportedly leaning toward suing Google before year's end. Naturally, its rivals are lobbying the feds to come down hard on the search giant.
Yet Google's critics haven't put forward a serious legal case against the company. The world's top search firm may be many things -- some of which aren't pretty -- but an illegal monopoly, it is not. If the feds haul Google to court, they'll send Silicon Valley a powerful message: Washington is open for business and happy to meddle with the Internet economy.
The search bias fallacy
Perhaps the most commonly recited complaint against Google is that its search results are biased toward the company's own offerings -- such as its Product Search and Maps -- to the detriment of its competitors. Some specialized search engines, such as Foundem (a price comparison service) complain they're inexplicably ranked poorly by Google search, even as Google's own specialized search engine appears prominently in the results.
Before unpacking this argument, it's worth noting every attempt to show Google suffers from rampant "search bias" has fallen flat.
A 2011 study by Harvard Professor Ben Edelman -- perhaps the most prominent analysis of search bias to date--purported to find that Google linked to its own properties in its search results more than three times as often as do other search engines. But more recently, a vastly more extensive study (PDF) by George Mason University's Joshua Wright found Edelman's methodology to be "designed to maximize the incidence of search bias." So Wright analyzed a bigger data set and found Google rarely ranks its own content higher than rival search engines do. It turns out that Bing, Google's chief rival, actually displays alleged search bias nearly twice as often as Google.
Admittedly, measuring bias in search is tricky, as Google founders Larry Page and Sergey Brinexplained in a 1998 Stanford working paper. If a search engine is covertly preferring its top advertisers in the "organic" rankings, users may never find out. But nobody is accusing Google of selling search priority to the highest bidders. Rather, critics like Edelman merely claim Google's search engine treats the company's products somewhat more favorably than rival search engines do.
Does this mean Google is biased? Not really -- at least, no more so than its competitors. There is, after all, no "unbiased" baseline against which to compare every search engine. If anything, search is all about competing biases. Search engines compete by striving to give users better results than they can get from rivals.
A search engine's job is to answer each query by generating links that best match the user's search term. This task is enormously challenging; deciding which Web site best "answers" a particular search query is as subjective a question as they come.
How do search engines decide which results to generate for each query and how to order them? For one thing, they may measure each site's popularity. If most users who search for "email" go on to Yahoo Mail, for instance, it could mean Yahoo Mail is the best result.
But if searching the Web were simply a popularity contest, all search engines would return nearly-identical results. They certainly wouldn't employ teams of computer scientists to constantly tweak their search algorithms.
In reality, however, a Web site's popularity is only one of hundreds of signals that search engines use to generate results. An awesome new gaming Web site that few users have yet discovered may merit a higher search ranking than older -- but more widely trafficked -- gaming sites. A search engine's challenge is to sort through the trillion Web sites out there and decide how best to weigh the many factors behind each site's ranking.
Back to search bias: Why, in 2011, did a Google search for "email" return Gmail first -- and Yahoo Mail second -- even though far more users actually selected the latter option?
Maybe Google was out to cripple Yahoo by forcing its own inferior products on users. More likely, however, Google simply thought Gmail was a better answer than Yahoo Mail. After all, Google spent years building a product it hoped would revolutionize email by offering unique features (such as gigabytes of free storage.)
Google's critics obsess over the inner workings of the company's search algorithm, known as Google's "secret sauce." Conspiracy theories abound regarding Google's alleged manipulation of search results. Why would Google demote my Web site, the argument usually goes, unless it's out to get me?
But what really matters is whether Google's algorithm is doing its job: generating results that users believe best answer their searches. This effort entails frequent tweaks, and occasionally causes some Web sites' rankings to drop dramatically. For instance, in April 2012, Google unveiled the "Penguin" update to punish low-quality websites. And in September, Google began demoting Web sites that frequently generate copyright complaints. Even Google's own properties aren't immune from these demotions; in early 2012, whenChrome violated Google's paid links policy, its search ranking plummeted.
Google undoubtedly makes the wrong call sometimes, to be sure; not every tweak works out for the better. But these mistakes are inevitable when search engines are coded by human beings.
Where's the harm?
The secrecy surrounding Google's search algorithm makes it impossible to know just how it decides to rank each Web site. (It also makes it tough for spammers to game Google's system, though not for lack of trying.)
Google's critics say it can't be trusted to overcome the irresistible temptation to "abuse" its search dominance. They point to Page and Brin's 1998 academic paper, in which they admit "search engine bias is particularly insidious" and tough to detect. Thus, critics argue, the only way to ensure Google acts in users' interests is by government bureaucrats peering over its shoulder (PDF) -- and examining the recipe for its secret sauce.
For all we know, Google may well routinely manipulate search results in a self-serving manner. If it does, would users notice the degraded results? Would they flee to other search engines such as Bing and AOL, or specialized search sites such as Amazon or Twitter?
Absolutely, Google argues. "Competition," it often says, "is one a click away."
Bolstering this argument is Microsoft's recent blind search engine "taste test," in which 5 million visitors preferred Bing to Google by a two-to-one margin.
But even if Google critics are right, and the company could abuse its dominance, where is the evidence that it has? Where is the evidence that Google's efforts to improve search quality have harmed consumers?
Without such evidence, there is no serious antitrust case against Google. It's perfectly legal for a company to compete by serving its users, no matter how high its market share or how much its rivals suffer as a result. As the Supreme Court has explained, a company with monopoly power is generally free to compete however it wishes, so long as it has "legitimate competitive reasons" for its actions. What could be more legitimate than trying to deliver better search results?
There's also no evidence that Google's conduct has choked off rivals. In fact, as The Daily Caller reported in September 2012, many of Google's "victims" appear to be thriving. Yelp's CEO reported massive growth in use of its mobile app during an August earnings call. And in July, TripAdvisor's CEO told investors not to worry about Google because his company was "entrenched in the fabric of travel planning.
" So much for Google "tend[ing] to destroy competition itself," to borrow a phrase from a Supreme Court opinion explaining the purpose of antitrust laws.
To Google's critics, however, any evidence of preferential treatment proves Google's guilt. Their argument boils down to the audacious claim that it's illegal for Google's search engine to show preferential treatment to Google products. As punishment, they want the government to forbid Google from deciding (PDF) that consumers actually might prefer its products and thus rank them prominently in Google search.
But if Google is so powerful that it can trounce rivals and spread its dominance throughout the Web virtually unchallenged, why did its much-hyped Wave and Buzz products fall flat? Why does Google+ have a mere 100 million active users versus Facebook's staggering 1 billionusers? And why are venture capitalists investing in startups more than they ever have since the dot-com bubble's peak in spring 2001?
If Google is evil, so is everybody
In fairness to Google's critics, they are right about one thing: Google is no saint. In fact, it's committed some major screw-ups over the years. For instance, the company's Street Viewcars collected wireless "payload data" from some unsecured networks, landing Google in hot water with governments around the world. In early 2010, when Google launched Buzz (the predecessor to Google+) some users' private contacts became publicly visible. Within days, Google apologized for the mistake and fixed it--but still ended up paying $8.5 million in a class action settlement. And last August, Google paid an unprecedented $500 million fine to settle claims its employees had helped Canadian pharmacies illegally peddle medicine to U.S. consumers.
Yet, troubling as these actions may be, they do not make Google an illegal monopoly. If they did, Google's rivals would be in hot water, too. Nearly every major multinational corporation has run afoul of at least one law or regulations--often several times. Just as Google is subject to a 20-year consent decree with the FTC over alleged privacy violations, so are Facebook, Twitter,Microsoft, and MySpace (to name just a few examples).
As much as Google is disliked in some circles, America's antitrust laws are designed not to punish companies for growing too big or too unpopular, but to ensure no company stifles competition itself. The thriving Internet sector -- a bright spot in America's otherwise lackluster economy -- shows no signs of suffering from too little competition. While all markets are decidedly imperfect, antitrust litigation is ill-equipped to distinguish beneficial innovation from the harmful kind, as scholars Geoffrey Manne and Joshua Wright explained in a 2011 Harvard Journal of Law and Public Policy article.
Nobel Prize-winning economist Ronald Coase once quipped (PDF) that when "an economist finds ... a business practice ... he does not understand, he looks for a monopoly explanation." This sums up the fallacy underlying the case against Google. Its behavior may be frustrating, its employees fallible, and its products inconsistent -- but it's also an American success story that has changed the world for the better, following in the footsteps of Ford, Sears, General Electric, Apple, Amazon, and even Microsoft.
In Silicon Valley, brilliant ideas breed success, and entrepreneurs expand our economic pie by creating wealth. Washington, by contrast, revolves around influence-peddling, special interest lobbying, and zero-sum rent-seeking games. If we want the Internet to remain dynamic, vibrant, and unpredictable, the government needs to leave it alone.