Controversial Competitor Study Suggests Google's Algorithm Has Search Bias

Yet another study charging Google with search results favoring it or it's clients? You don't say!

The study purports to demonstrate that consumers prefer search results including listings from other sources rather than Google-only selections. This does, conveniently, mirror charges against Google filed by the European Union earlier this year (we will refrain from raising questions of competence based on its handling of the Greek crisis ...). This suggests Google rivals are piling on in hopes of having the Europeans do for them what they can not competitively do themselves.

But it also hearkens back to an earlier, simpler era when 'eyeballs,' 'stickiness' and other spurious measures of attention were deemed evidence of commercial viability. That further research concluded such behaviors were, au contraire, the result of frustration with websites' poor design and overall kludginess appear not to have occurred to this latest set of researchers.

Google may well be fudging research results to drive traffic to its own sources or those of its advertisers and business partners. And those questions may well deserve further scrutiny by the appropriate authorities. But having real evidence rather than rhetorical extrapolations paid for by competitors to support that contention would be a good place to start. JL

Dino Grandoni reports in the New York Times:

Lots of clicks can be sign of irritation rather than satisfaction. It is also difficult to discern whether users were more engaged with Google during the test because the results were more relevant, or merely because the page had a different look.

Google entices people to search by promising links to the best that the web has to offer. But research released Monday, led by top academics but paid for by one of Google’s rivals, suggests that Google sometimes alters results to play up its own content despite people’s preferences.

In the study, researchers from Harvard and Columbia presented 2,690 web users with two versions of Google. One version showed search results for local businesses as users usually see them, with links to the businesses along with ratings as posted to a Google site. The other version showed links to businesses along with ratings from rival sites like Yelp, the online review website, which paid for the study.

The people studied were 45 percent more likely to click on links if Yelp and other competitors were included — a sign, researchers say, that users prefer more diverse search results.

The study could renew calls for government regulators — in particular, the Federal Trade Commission — to reopen an investigation into Google for unfairly promoting its own services. The results may also provide new ammunition to officials in Europe who have accused the company of antitrust violations.

“The idea that you can build a better version of Google search engine pretty easily if you don’t exclude competitors to me was a pretty startling finding,” said Tim Wu, a co-author on the study, who was paid by Yelp to conduct the study.

Yelp has become one of Google’s most vocal competitors, and it promoted the study on Monday with an accompanying website and YouTube video to explain the findings. But Yelp’s biggest promotional asset may be Mr. Wu, a Columbia law professor known for coining the term “net neutrality,” the phrase widely used to describe Internet service in which all data is treated equally.

Mr. Wu has defended Google’s competitive practices in the past, but said he was swayed to do the study after speaking with a Yelp executive last fall and looking at internal data Yelp collected on Google.

Yelp also paid his co-author, Michael Luca, to conduct the study, and the company flew Mr. Wu to present the findings at Oxford on Sunday in England.

Google questioned the results on Monday.

“This isn’t new,” Google said in a statement. “Yelp’s been making these arguments to regulators, and demanding higher placement in search results, for the past five years. This latest study is based on a flawed methodology that focuses on results for just a handful of cherry-picked queries. At Google we focus on trying to provide the best results for our users.”

Some outside experts questioned the findings, too. Lots of clicks can be sign of irritation, they said, rather than satisfaction, like if users keep navigating from Google to Yelp and back again. It is also difficult to discern whether users were more engaged with Google during the test because the results were more relevant, or merely because the page had a different look.

“Over all, it comes across more to me as a public relations exercise rather than precise science,” said Danny Sullivan, a search engine analyst and the founder of Search Engine Land, a site about the search industry. “However, I do think Google could easily include links to other review sites which would benefit both its users and competitors.”

In April, European Union officials accused Google of unfairly manipulating search results, the first time the company has faced such charges.

The company has been given until mid-August to review documents and make a defense. A loss could mean billions of dollars in fines for Google.

Google avoided a similar fate in the United States. In 2013, the F.T.C., the federal agency that regulates unfair business practices, voted not to bring charges against Google after an investigation, finding that the updates to the search engine were made to help users, rather than hurt rivals.

A spokesman for the F.T.C. declined to comment on the study.

While Yelp published the study “for an antitrust audience,” according to Luther Lowe, vice president for public policy at Yelp, the company had no comment on whether the F.T.C. should reopen the case against the company.

“There are antitrust authorities around the world looking into Google,” Mr. Lowe said.