Bing Caught With Their Hands in the Google Cookie Jar Honeypot

February 1st, search engine reporter Danny Sullivan broke the story that Google recently found evidence that Bing has been "copying" Google results. Was Google trying to pick a schoolyard fight with Bing, or were the legitimately on to something? From reading both the Google and Bing sides of the story, I have to say I'm leaning towards Google. But read the evidence yourself before taking my word on it...

May 2009 - Google noticed Bing returning results on obscure misspellings, identical to Google's. Google would correct the misspelling, then return results appropriate to the proper spelling; in Bing the top results would simply appear, without any notification of spelling correction.

Images from Danny Sullivan's original article here.

October 2009 - Google noticed a large jump in the number of instances the results within the top 10 overlapped in both engines. Also noted was an increase in the number of times Bing's #1 listing was the same as it was in Google. Although there were still many instances where results were significantly different, it was clear that Bing made some changes in their algorithm to return more "Google-like" results.

Google began to suspect that some copying was going on, because there seemed to be no other way to explain the new found similarities in rankings, especially in the case of misspellings. Their hypothesis was that somehow, Microsoft was leveraging the use of Internet Explorer, specifically the "suggested sites" feature, or the Bing toolbar to send back information on user actions within Google.

December 2009 - To test this search espionage theory, Google set up what has been cleverly nicknamed as the "Bing Sting," which included the following steps:

  1. Complete approximately 100 "synthetic searches" for a string of gibberish of which nobody would ever actually type in to search.
  2. Note neither Google nor Bing returned any quality results for this string of gibberish.
  3. For the first time ever, create a one-time use code that would manually rank a certain website to this specific gibberish-style keyword, nicknamed a "honeypot page," and push live.
  4. Instruct about 20 engineers to perform the gibberish queries at their homes, using IE with both Suggested Sites and the Bing Bar enabled. These tests began on December 17th.

Low and behold, by December 31, some of the Google "honeypot pages" were beginning to rank on Bing. (Incidentally, the synthetically ranked pages they chose are pretty hilarious. Check out Danny Sullivan's post to see that a Google engineer must have been doing this at lunchtime.)

Bing argues that since only 7 to 9 of the 100 gibberish keywords showed Bing ranking Google's honeypot pages, it doesn't prove any cheating going on. My opinion - if Sandra Lee's recipes rank #1 for the keyword "indoswiftjobinproduction," something fishy is going on.

 Please contact us to learn more!