Finding relevance through Application Focus signals
So do you think Google is watching what else youre doing on your computer and seeking to personalize your search experience based on open document, emails or other desktop applications? It is a interesting question
. So lets get out our tin-foil hats once again and go for a ride.
I have read about such mythical beasts in a few patents over the last while and thought it was an interesting idea, worth at least floating here. Basically, things you may be reading, writing or listening to on your computer (or Android phone?) are used as signals for customizing your search results. Sound like madness? Then youve come to the right place as we specialize in it around here.
Consider I am listening to Rock Lobster on my media player and decide to search b52 by monitoring application focus the search would likely re-rank documents relating to the music group not the famed airplane. The same if related phrases are being used in an instant messaging interface, in an email
well, you get the idea.
Application focus as a ranking factor
I started thinking more after reading a few of the comments from the folks developing search for the Android mobile system that mentions;
we've integrated search across the platform so that applications can even share search capabilities with each other.
Then we wonder about what extent they communicate with other applications? And then, obviously, what personalization could be occurring in the search results. It was time to play connect the dots, slipping further into the madness.
Google on my desktop
Heres a look at the search result for link building to get this;
Those top bits are related to my Google Desktop installation, (which itself was an experiment to connecting the dots). Essentially it is mixing results from my other applications with the normal SERP (Search Engine Results Page) as explained here;
The Google Integration feature shows you combined personal search results and web results in your web browser. GDT Support
The suggested results then recommended not only local documents, but searched the URLs in my recent search history (from any browser) which is much different than my search history as normally associated with personalized search (via Google account).
Clicking that leads to;
As you can see it looks at local files, emails, web history and even chats
(what is other I wonder?) Anyways, plenty of potential signals there.
The Google Empire
Now lets take stock for the moment and look where they can get these cross application focus type signals;
Potential for Application Focus signals;
Google PC (total access)
Google Chrome (applications, browser and full surfing history)
Google Android and upcoming G1 (surfing history, email, services/applications)
Google Desktop (near total access)
Related personalization data;
Google Services Account such as Gmail, Docs etc.. (PS via cookie, demographic info)
Google ToolBar (full surfing history, bookmarks)
This is a formidable list, even if the gPC hasnt been gang busters of late
Think of the cumulative data that theyd have. And getting rid of search history still isnt the easiest. Looking at the info on the Chrome browser, using Incognito mode, the search history might still be tracked;
If you sign into your Google Account on http://www.google.com while in incognito mode, your subsequent web searches are recorded in your Google Web History. In this case, if you want to make sure your searches are not stored in your Google Account, you'll need to pause your Google Web History tracking. -
I just love these guys
indexing the worlds information right?
My friends, personalization is becoming far more common and takes many forms. Data that can be mined from changing application focus may be part of a new age of information discovery. It is at very least concepts worthy of strategic and lateral thinking for us search geeks.
What does any of this have to do with u?
This means that there are more ways to collect personalization signals through application focus than ever before. Hell, I wouldnt be surprised to see a Google TV or some way to get their fingers into the data cookie jar of the original time sink.
Point being that not only are these potential personalized re-ranking data access points, but also larger over all ranking signals. Sure, theyd likely not make for a strong signal, but playing with the data for tweaking seems a sensible path. Behavioural metrics make for some very interesting analysis not only on a singular level, but categorization and segmentation (clusters) as well.
Its something to watch if youre a Search marketer and even for those Social Media Types that can create buzz and spawn email, instant messenger, social activity (because we havent forgot OpenSocial right?). Who says ranking signals and information retrieval needs to be confined to a search engines index? Changing the application focus means access to data typically beyond the reach of search engineers not so long ago.
As behavioural data and personalization efforts increasingly play a factor in how search is evolving a premium on value and reach is paramount.
You may now take off the tin foil
but keep it handy.
NOTE: we're playing with some data on SERP flux/personalization... stay tuned for more soon ;0)