The dust-up between the two companies that process virtually all of North America's search requests grabbed the spotlight Tuesday at an event sponsored by Microsoft about the future of Internet searches. Microsoft's practices have even wider implications now that its technology powers Yahoo Inc. searches in the U.S., Canada, Mexico, Australia, and Brazil as part of a 10-year partnership that grew out of the companies' inability to mount a serious challenge to Google on their own.
Google's attempt to embarrass Microsoft at an event devoted to innovation served as the latest reminder of the tensions between the technology heavyweights. While Microsoft has been pecking at Google in search, Google has been chipping away at Microsoft's advantage in computer software with its own suite of competing products.
"We just want everyone to know the truth about how Microsoft operates as a search engine, which is by taking the hard work of others and presenting it as their own," said Amit Singhal, a Google fellow who oversees the company's closely guarded search formulas. He made his comments in a phone interview.
Microsoft did nothing more than adjust its results after monitoring Internet Explorer users' search requests and clicking activity on Google as well as its own site, Bing, according to Harry Shum, a corporate vice president for Bing. In a blog post, Blum derided Google for engaging in a "a spy-novelesque stunt."
Google Inc. set out to expose Microsoft's tactics last year, said Matt Cutts, the head of Google's Web spam team. That's when it appeared Bing was showing search results that seemed a little too close to Google's own -- especially for obscure, misspelled queries.
The similarities raised suspicions that Microsoft's IE Web browser and various other tools were feeding information back that helped Microsoft's engineers make Bing's results more Google-like.
Google laid a trap to prove it. The company made a list of gibberish or obscure search terms and manually linked them to unrelated websites. Then, 20 Google engineers took home laptops loaded with Internet Explorer, searched Google.com for those terms and clicked on the artificial results. Soon after, searching for the same odd terms on Bing would call up the same odd results.
Cutts likened the trap to a mapmaker drawing a fake street or the Yellow Pages adding a fake name to its directory to flush out copycats.
The "Bing Sting" was first reported on the Search Engine Land blog before emerging as a hot topic during a panel discussion that included Cutts and Shum. The San Francisco event was streamed over the Internet.
"It's not like we actually copy anything," Shum said. "We learn from customers who are willing to share data with us, just like Google does."
Those data include not only the searches people type into Bing, but also into Google, and what links they click on. The information can be used to fine-tune Bing's own search results. And that sort of "collective intelligence," Shum said, is how the Web is supposed to work.
Google doesn't use people's behavior on Bing the same way, Cutts said during Tuesday's event.
In an interview, Singhal argued it's unfair for Bing to piggyback on Google 's technology.
"It's like a student cheating on his test and saying, 'Yeah, I could see my classmates' test, so I wrote it down,"' Singhal said. "If that's not cheating, what is?"
Cutts and Shum traded jabs about whether people read the fine print when installing the Chrome or IE browser software that explains what Web surfing information is fed back to the company.
When the discussion moved on to the problem of increasing spam pages and low-quality content online, Shum blamed Google for rewarding the owners of such pages with advertising dollars.
Cutts said Google manually blocks spam pages regardless of whether they carry Google ads but wants to find a technology solution for the problem instead of picking off useless sites one by one.
Liedtke contributed from San Francisco.