A Key Question in Google’s Trial: How Formidable Is Its Data Advantage?
The federal government’s first monopoly trial of the modern internet era is less than a week old, but already a central character has emerged: data. Its role, its use and its power are key issues in the Justice Department’s case against Google.
The government claims that Google bribed and bullied smartphone producers like Apple and Samsung and the browser maker Mozilla to be their featured search engine, funneling far more data to Google and cutting off competitors.
Data, the government says, drives the flywheel of Google’s success. Each search query adds data, which improves search results, attracting more users who generate still more data and advertising revenue. And Google’s ever-growing data advantage, the government asserts, is an insurmountable barrier for rivals.
Data is “oxygen for a search engine,” Kenneth Dintzer, the Justice Department’s lead lawyer, declared in his opening statement on Tuesday.
The government’s case is not that Google violated the law in becoming a search giant. Instead, the government claims that after Google became dominant, the company broke the law with its tactics to defend its monopoly. Contracts with industry partners to be their default search engine were the weapon — exclusive deals that froze out rivals, the government claims. So Google is now protected from competition behind a fortress built with data.
Google replies that the government’s case is an artifice of misleading theory unsupported by the facts. The government has chosen to “ignore inescapable truths,” John Schmidtlein, Google’s lead lawyer, asserted in his opening statement.
Those truths, according to Google, are that the company holds its leading position in search because of its technical innovation. It competes with others for default-placement contracts and wins mainly because Google is the best search engine. Those contracts, Google argues, help reduce prices for smartphones and benefit consumers.
The government, Google insists, is overstating the importance of data. In a brief filed this month, the company stated, “Google does not deny that user data can improve search quality, but Google will show that there are diminishing returns to scale.”
The trial resumes this week with the Justice Department continuing to present its case. The first witness scheduled to testify on Monday is Brian Higgins, an executive at Verizon who oversees mobile device and customer marketing. The trial is scheduled to run for 10 weeks. A ruling from Judge Amit P. Mehta will come next year.
Google holds 90 percent of the search engine market in the United States, while Microsoft’s Bing is a far-distant rival, with less than 5 percent. The difference, Google says, is explained by the smarts of its engineers, not the size of its trove of rich data.
To make its point, Google will call an expert witness, Edward Fox, a computer scientist at Virginia Tech. Professor Fox has conducted a “data reduction experiment” on Google’s behalf to estimate how much Google’s search quality would decline if it used far less data — roughly the amount available to Bing. The result, according to Google’s filing, was that the data difference explains only part of the gap in search quality between Google and Microsoft.
Google’s public messaging on that issue has been consistent over the years. But the government claims that the messaging has been deceitful. In his opening statement, Mr. Dintzer said Google had “misled the public about the importance of data.”
To try to show the deception, the government introduced emails among senior Google employees, sparring over that point, as evidence last week. Hal Varian, Google’s chief economist, was questioned about defaults and data as the Justice Department’s first witness.
At issue were comments that Mr. Varianmade in a 2009 interview with the technology news site CNET.
In the article, Mr. Varian said, “The scale arguments are pretty bogus.”
To explain, he added in the article: “It’s not the quantity or quality of the ingredients that make a difference. It’s the recipes.” It was a deft analogy, with the ingredients being the data and recipes being the clever algorithms written by Google’s engineers. It reached a wider audience when Mr. Varian’s explanation was picked up in a Time magazine article.
But in an email to Mr. Varian shortly after, Udi Manber, a senior engineer on Google’s search team, took issue with the economist’s description. “It’s absolutely not true that scale is not important,” Mr. Manber wrote. “We make very good use of everything we get.”
In an email string with other Google employees, Mr. Manber wrote, “I know it reads well, but unfortunately it’s factually wrong.”
A lot has changed since then. The government introduced the emails to cast doubt on the credibility of what Google is saying in court. They amount to a few snippets at the beginning of a lengthy nonjury trial that will generate piles and piles of evidence, testimony and rebuttal for Judge Mehta to weigh and consider.
But what is already clear is that a debate over data in search — whether it has market-determining power or not — is a pivotal issue, and one that both sides will most likely be compelled to return to repeatedly in the trial.