Google’s Bing Sting Could Backfire

Google has leveled accusations that Microsoft has been using its search results to improve Bing’s listings.

Google came to this conclusion based on the results of a sting operation its engineers launched after noticing unusual misspellings were cropping up in both Google and Bing search results.

Google’s suspicions were aroused by a search for “tarsorrhaphy” — a rarely performed surgical procedure on the eyelids, according to a company blog post. In the summer of 2010, the search team at Google was looking at the search results for an unusually misspelled query: “torsorophy.” Google returned the correct spelling along with the appropriate search results.

At that time, Google said, Bing had no results for the misspelling, but it subsequently began returning Google’s results to its own users — without the spelling correction.

That was in the summer. By the last two weeks of December, Google began conducting its sting, inserting so-called honeypot results in about 100 “synthetic queries” — that is, queries one would not expect a user to type, such as “hiybbprqag.” For each synthetic query, Google inserted as its top result a real Web page that had nothing to do with the query. In multiple instances, these results were duplicated on Bing, Google said in the post.

In its initial public responses to the accusations, Microsoft hedged its answers, giving neither a straightforward yes or no in response to Google’s accusation.

Neither Microsoft nor Google responded to the E-Commerce Times’ requests for comment in time for publication.

Yes or No?

Microsoft provided a more in-depth explanation to Google’s charge in a blog post byHarry Shum, Bing’s corporate vice president.

Bing uses more than 1,000 different signals and features in its ranking algorithm, including clickstream data it gets from some customers that opt in to sharing anonymous data.

Google’s revelation is a “spy-novelesque stunt,” Shum said, that doesn’t accurately portray how Bing uses this opt-in customer data.

Some Nerve

The accusations have left the search community buzzing about which company — Google or Microsoft — is displaying the most chutzpah.

“Does this really surprise anybody?” asked David Dalka, a digital business strategy consultant.

“All of the social networks copy features,” he told the E-Commerce Times. “Instead of hiring business leaders to design unique products, Microsoft passively recruits engineers directly from Google. It wouldn’t surprise me at all if some of the actual code was the same.”

That Microsoft feels compelled to use Google’s search results to improve its own suggests that Bing is lagging in ingenuity, Lawrence Knorr, a professor with the Harrisburg University of Science and Technology, told the E-Commerce Times — even if Microsoft is couching it in terms of using customers’ data.

“Apparently, through an opt-in feature — Suggested Sites — in the IE Browser, data is being fed to Bing any time a user of IE utilizes Google,” he said. “These results are then used to influence Bing search results. This is akin to a parrot imitating a voice it hears without knowing what the words actually mean.”

On the other hand, Rob Enderle was taken aback by Google’s nerve in even raising the issue, considering it uses similar practices — and has committed far worse transgressions with the use of search data over the years.

“Google regularly scans this same material to improve their own service,” he said.

Furthermore, Google may regret highlighting this practice so publicly, as it could invite regulatory scrutiny, suggested Enderle. “The government could conceivably decide that search information needs to be protected from all vendors, if only to level the playing field among the providers.”

How the Sausage Gets Made

The controversy also provides a public glimpse into the behind-the-scenes sausage-making that is search.

“First of all, this kind of monitoring goes on on both sides,” Michael Hussey, CEO of search engine PeekYou.com, told the E-Commerce Times. “Competitors constantly monitor each other to see how they compare or stack up with search terms. That is only prudent.”

As for the copying, though it was news to him, Hussey thought Google might be overstating the offense.

“I am assuming that 999.999 percent of Bing’s search results are from its own algorithms,” he said. “If they are copying — and we don’t know that — it would be for very long-tail searches.”

That would make sense, Hussey said, as it is always better to show a user something rather than nothing for a search, no matter how outlandish the request is.