At a conference on search Tuesday morning, Google search engineer Matt Cutts suggested Bing may be copying Google’s results. After seeing indications that this might be the case, Cutts’ team set up a bunch of fake search results to see if they’d make their way into Bing’s. Some of them did.
But Microsoft said Cutts was misinterpreting what was happening. “We learn from our customers,” Microsoft corporate vice president for core search development Harry Shum replied. “We use customer data to improve the customer search experience.”
Specifically, Bing, like Google, has an algorithm that processes over a thousand variables–called “signals”–to calculate a page’s rank. Among those signals is data about what users have clicked on in previous searches. Bing collects that data through programs that users opt-in to. (And yes, the data is anonymized.)
Bing doesn’t only collect the clickstreams when its users use Bing. They also use the clickstreams generated when their users use other search engines. Like Google. What Google most likely did in the example now bouncing around the Internet (and likely turning up in Bing as results for searches on terms “Bing,” “Steal,” and “Google”), is narrow down factors Bing considers in its results to a single one — the Google one. So it’s true. If you ever happen to be searching for a term that makes no real sense in English, and Google happens to have created a fake search result for that illogical term, Bing will give you those fake results, too. “This does not a conspiracy make,” Bing spokesman Stefan Weitz tells Fast Company.
“This allegation that we’re copying their results wholesale is disingenuous,” he adds. “We’ve always said we use clickstream data, with people’s permission, to improve the quality of search.”
“I commend [Google] for their PR skills, but when you dig into it, it’s a lot less sexy,” Weitz says, adding later, “We’ve probably learned a few things about click fraud perpetrated by our competitor.”