Jack Menzel is a Google’s Social Search product manager. I am at SMX West in San Jose, sponsoring the event with www.majesticseo.com and I managed to take some time out to hear him talk.
He started by showing us the phrase “SMX”. He showed that whilst everyone in the room associated t with the conference, there are many other associations – and it is really only those connected to the people in the room that understand “SMX” to be a conference on search.
Language can be imprecise, but presonalization can help to fill in the gaps, providing it can be done in a privacy sensitive way. Google’s interpretaion of privacy was a sense of transparency and control for the users and – presumably – the used.
Jack noted that search – old school search – was ultimately a very efficient word conting machine. Whilst sophisticated, the algorithms ultimately did not understand anything. With personalization, however, Google understands context as well. Knowing that you have put up pictures of Rome, for example, gives Google knowledge that you have been to Rome – which can affect the result you see in personalized search.
The concept of search circles – social circles – is one that Google has used significantly in Google+. For example, without context, a search for “Ben Smith” is unlikely to be meaningful for you if you have a friend called Ben Smith, but knowing who you know, easily helps Google to hone in on the Ben Smiths that have relevance to you.
As an aside to Jack’s presentation, here is a schema that I quickly knocked up to illustrate this.
You >>>>>> Friend A >>>>>>>>>>>>>>>>> Ben Smith
>>>>>>>>>> Friend B >>>>>>>>>>
(If two or more of your friend link to a particular Ben Smith, then this is a string social signal.)
There is no such thing as “non-canonicalized search results” in social. Turning off personal search does not mean you will see what other people see.
Google has been strongly encouraging the use of the +1 button, to help make your content more shareable and they are also encouraging people to share ion conversations when you see a friend or colleague recommending a search result in personalized results.