A new paper co-written by Columbia law professor Tim Wu argues Google does a disservice to users by favoring its own products
There’s been a steady drumbeat of antitrust and anti-competitive accusations against Google GOOG over the past several years—not just in Europe but in the US as well—and one of the major players pounding out that message has been Yelp, the online recommendation network. The company launched another salvo on the weekend, with a paper co-authored by noted legal scholar Tim Wu, a law professor at Columbia University who is also a former Google Fellow and former advisor to the Federal Trade Commission.
In the paper, Yelp YELO and Wu and Harvard business professor Michael Luca try to make the case that Google systematically disadvantages users by giving them search results that highlight Google’s own services, rather than “objective” results for the term they entered. According to Wu—who coined the term “net neutrality,” but in the past has demurred on the subject of whether or not Google is being anti-competitive with some of its behavior—the evidence against the search giant is incontrovertible.
“When the facts change, your thinking should change,” Wu told Re/code. “The main surprising and shocking realization is that Google is not presenting its best product. In fact, it’s presenting a version of the product that’s degraded and intentionally worse for consumers. This is the closest I’ve seen Google come to [being] the Microsoft case.”
In order to arrive at this conclusion, Yelp and Wu looked at Google’s results for a range of local search terms—such as “pediatrician in New York”—including the so-called “One Box” the search giant typically displays containing results that come from Google-owned and operated services. The researchers then compared that to a set of results they got by using a browser plugin which stripped out the in-house results, but otherwise left the search untouched.
The fact that users are being disadvantaged by Google’s “self dealing,” the paper says, can be shown by how much more frequently they click on other results when One Box services are excluded. According to the survey, users clicked on the filtered results 45% more than they did on Google’s results with the One Box included.
In other words, Yelp and its co-authors argue, users obviously preferred the cleaned up results to the ones where Google tried to promote its own services. That’s obvious evidence of harm, the paper says, and therefore presumably Google should be sanctioned, or forced to change its behavior.
The idea that Google gives preference to its own services is at the center of the antitrust actions against the company, both in Europe and in the US. And the reason why Yelp and Wu are making such a big deal out of whether users are being disadvantaged is that consumer harm plays a central role in findings of antitrust or anticompetitive behavior in the US under the Sherman Antitrust Act—the law that the government employed to go after Microsoft in the 1990s.
There’s a common misperception that all you have to do to take action against a company like Google is to show that it has a monopoly in a specific market. But simply having a monopoly isn’t illegal—in the United States at least. What’s illegal is using that monopoly position in an anti-competitive way against others in the industry. And even then, anyone arguing an antitrust case has to show that this behavior actually harmed users in some way.
So the idea that Google is giving people inferior search results is a critical factor. But is that really what’s happening? That’s a difficult case to make. Search results aren’t an exact science by any means, nor is Google in the habit of talking about exactly how its algorithm functions. So any assumptions made by critics like Yelp or Wu have to be based on a kind of reverse engineering of a result they don’t really know anything about, and that process is inevitably going to be flawed.
For example, looking at whether a user is satisfied with a search on the basis of how much they click has a number of potential holes. Among them are the fact that much of Google’s strategy with the One Box approach is to give users as much information as possible about their query without making them click at all. So the fact that many didn’t click might actually mean that they were very satisfied with the results — so satisfied that they didn’t have to click to get more information. The paper also focused on hyper-local results, which are notoriously subjective.
Even if the results are accurate and users were not getting the best possible search results, Mike Masnick at Techdirt raised a good question when Yelp first started talking about its “Focus on the User” campaign last fall: Is the quality of search results really something the government should be issuing edicts about? Presumably, if Google results were that bad, users would flee.
The counter-argument, of course, is that Google has too much market power for anyone to leave, thanks to network effects and the influence of an illegally-acquired monopoly. But is the search market really a monopoly? And even if it is, will sanctioning Google do anything to solve that problem? Both of these questions are difficult to answer.
There’s a case to be made that Google is losing much of its previous dominance when it comes to finding content, in part because of the rise of Facebook and the power that social signals have (which is why Google cut a deal with Twitter for access to its “firehose” of user data). Some antitrust watchers would say Microsoft was in much the same position when the case against the company started in 1998. If the Justice Department had left well enough alone, market forces might have had as much or more of an effect on Microsoft and its market share as the government did.
It’s one thing to argue that Google is a large company that throws its weight around and promotes its own services, but it’s another thing entirely to argue that the company should be sanctioned by the federal government for that—or that doing so would somehow magically result in a better search marketplace for consumers. Neither of those are obvious.