While trying to get my head around the whole “Google caught Bing cheating” I found myself asking more and more questions.
The first question was when Google ran the experiment they say around 20 Google engineers went home and ran the test queries using Internet Explorer with a Bing toolbar installed and the Suggested Sites enabled. They also had to click on the top results in Google ( or the Fake results )
What no one is reporting is how these test queries where actually run ?? I guess one of two ways.
A) They opened IE8, typed in www.google.com then entered the search query ( How many times did they search in Google, 1, 10, 100, 10,000 times) then Clicked on the fake result but we don’t know how many clicks each engineer did on the fake listing?
B) They set the default search engine in IE8 to be Google and started searching for the Fake query :
The first question is really important to me because it actually makes me lean more to Bing doing nothing wrong if it’s actually turns out to be option B and here is why.
The set up that Google actually created, Microsoft states “statistics about your usage of Suggested Sites will also be sent to Microsoft such as the time that websites were visited, which website referred you, and how you got there”
so how does all this work ( NB I haven’t a clue ;)) but if I had to guess :
an IE8 user which has agreed to send data to Bing to improve the search quality enters a query, Bing then check the database to see what they have for that term ( I guess if nothing returned then Bing works hard to improve it’s quality ) so I search for Dave Naylor and Bing monitors which sites I visit :
My search moves to the default search engine I have set in IE
I click on #1 result http://www.davidnaylor.co.uk/ I hit the back button within seconds
returning me to : http://www.google.com/search?q=dave+naylor&sourceid=ie7&rls=com.microsoft:en-gb:IE-Address&ie=&oe=
I scan the results and find the twitter feed for dave naylor and click
http://twitter.com/davenaylor, I read a few tweets and click back, not what I was looking for, click back again
http://www.google.com/search?q=dave+naylor&sourceid=ie7&rls=com.microsoft:en-gb:IE-Address&ie=&oe= I see the Globe and Mail and click that
http://www.theglobeandmail.com/blogs/from-the-sidelines/ If I was looking for the sportswriter David Naylor I would stay here longer and read the articles etc..
In Bing’s eyes I searched in IE and visited 3 sites excluding Google
http://www.davidnaylor.co.uk/ – time on site 1 second
http://twitter.com/davenaylor – time on site 3 seconds
http://www.theglobeandmail.com/blogs/from-the-sidelines/ time on site 180 seconds
If people searching for “Dave Naylor” actually wanted the Globe and not me then a quality signal would have been the time on site which Bing actually says they track. Now I agree that it would look like Bing is using Google to get signals to improve their results but if they are doing it on any search engine which is set to default I would have to say they aren’t copying Google, they are using User click stream. A better test would have been to setup a fake search engine and fake the results see if they appeared in Bing as well 🙂
But What if the answer was A, Bing could have just set a script to monitor anything that gets sent with the Url http://www.google.com/search?q=dave+naylor and then follow the data stream. This is a little more naughty but still not copying Google. The fact that Google created fake data that no one had seen before, then created fake clickstream data and sent it to Bing (that says that the data will be used to personalize your experience, as well as improve the quality of our products and services.)
Either way one thing Google has done is given us a way to game Bing 🙂 and how much monitoring of Bing did Google do and how did they monitor them ?