Harvard researcher rejects Google's 'personal attack'

image captionGoogle is under investigation by the European Commission

A researcher who has accused Google of bias says the internet giant is now waging a campaign to discredit him.

Ben Edelman, an associate professor at Harvard Business School, published a study this week claiming Google boosts its own products in search results.

The Californian company responded by suggesting he was actually working on behalf of its rival, Microsoft.

But Mr Edelman told the BBC that Google was launching "personal attacks" to distract people from its own behaviour.

"I don't mind personal attacks, to be honest, because I think it shows they can't argue against the research," said the associate professor.

"That's what they've done to most of my recent research."

His study, "Measuring Bias in Organic Web Search", was conducted in August last year alongside Harvard colleague Benjamin Lockwood and published on Wednesday.

It found that Google gave undue prominence to its own sites across a number of popular search terms, such as e-mail, video and chat.

Under a search for "e-mail", for example, the study found that Google's top result is its own Gmail service - despite the fact that almost twice as many people searching the site subsequently clicked on the second result, Yahoo Mail.

The company responded angrily to the study, issuing a statement that accused him and his methodology of deliberate bias.

"The report is highly biased, ignoring contrary examples like "search engine", "book flights" or "directions", and failing to account for other reasons why certain sites rank highly," it said in a statement.

It was also unimpressed by the author's credentials.

"Mr Edelman is a longtime paid consultant for Microsoft, so it's no surprise that he would construct a highly biased test that his sponsor would pass and that Google would fail," it said.

"Google never artificially favours our own services in our organic web search results, and we perform extensive user testing to ensure that search results are ranked in a way that provides users with the most useful answer," it added.

Antitrust investigations

The issue strikes at the heart of Google's multibillion dollar search business, which it has always said is based on a neutral, automated system that merely reflects how popular a site is online.

That claim is already being questioned by regulators in Europe and America, following complaints from some services who feel they have been deliberately excluded.

In November, the European Commission announced that it was launching an investigation into whether the company was abusing its dominance.

Earlier last year, meanwhile, the attorney general of Texas said he was looking at claims that Google manipulates results to boost its own products and therefore make more profit.

Mr Edelman said his work was a thorn in Google's side because it backed up those criticisms.

"I can see why it would hit home for them - they're under antitrust scrutiny on at least two continents," said Mr Edelman. "Their approach of asking people to trust them seems to be wearing thin."

Mr Edelman told the BBC that the study could have been improved with more information and time.

He said its data could have been more recent and the search terms could have been independently chosen. But, he added, the numbers were sound and that he stood by his conclusion.

Not everyone agrees with Mr Edelman's analysis, however.

Danny Sullivan, the editor of the Search Engine Land website, has defended some of Google's practices in the past. He said the Harvard study did outline some "odd" behaviour but said the figures could also be used to show that Google was less biased than might be expected.

"Statistics can easily be turned to whatever you want them to be," he wrote. "I feel like Edelman is turning his study into the most negative view possible."

Responding to Google's claims of bias, Mr Edelman said that he had been a paid consultant for Microsoft in the past, a fact openly disclosed on his website.

However, he added, the research - which also looked at Microsoft's Bing service and found it was less skewed than Google - was not deliberately designed to produce such a result.

He added that he had also given free advice to Google in the past as well as other companies, including the BBC.

He also said that the company supported him when he criticised the behaviour and bad practices of Yahoo, which was Google's main rival at the time.

"When Google didn't have as much market power as they do now, they consulted with me on some projects," he said. "How quickly Google forgets."

More on this story

Related Internet Links

The BBC is not responsible for the content of external sites.