Integrating social search into search results is tricky, says Mayer, because people view search as a private activity. Most people just aren't that comfortable letting their whole network of friends know what they have been searching for. Google knows that they must respect that privacy, so they would want the user to explicitly approve any friend connections that would be used to add social elements to the search results. (In other words, you won't just log on one day and find your Google search results re-ranked based on what your MySpace friends are doing.)
When asked how Google is planning on implementing social search, Mayer mentions a few different ideas they have which include labeling, identifying users like you, social network integration, and a social-influenced PageRank.
Labeling: With labeling, Google users could annotate the search results and those notes could then be shared with friends on their social network or with others like them. She mentions that this has worked to some extent in Google Co-Op in certain areas, like health, but overall annotation is not a model that works well in its current state. However, the benefit of annotation is that it avoids the privacy issues because someone who is labeling search results presumably does not mind that others would see those notations.
Users Like You: Another option might involve Google taking a page from Amazon's book, and adding "others like you searched for ..." or "other people who did this search also did searches..." to Google's search results. Although useful, these related queries don't truly integrate results from your friends, nor do they influence the search result rankings, so they are not the best example of pure social search.
Social Network Integration: To identify your friends and allow them to influence search results, Google may even try social network integration with search. Using aggregate statistics on your friends' searches would allow privacy to be maintained, but you would also be able to see trends that are important to you. Initially Google would leverage the Google user base and the connections that exist within it. However, 3rd party social network integration may come in time as well. Mayer uses the example of how you could see that several of your Facebook friends had searched for a particular topic one day - a stat that would be provided without user names. If a large number of your friends are searching for something, it's likely that you may be interested in that topic, too.
Social-Influenced PageRank: With today's version of PageRank, it's the link structure of the web that determines the most authoritative pages. However, Google believes that people would naturally give more authority to pages their friends visit. To bring in this influence, Google could take web history and then allow that data to influence rankings, so that pages that your friends visit would rank higher in search results. Today, Google web history is still an opt-in option and if it was going to be used to influence rankings, that would hopefully be an opt-in choice as well, but Mayer does not go into that level of detail.
So, what is the future of search? Mayer responds, "I think one way it will be better is in understanding more about you and understanding more about your social context: Who your friends are, what you like to do, where you are. It’s hard to imagine that the search engine ten years from now isn’t advised by those things."
]]>I know Google's everone's darling, the belle of the ball, as it were. But I think they're behind on this one.
Y!s been working on stuff for a while internally, and what do you think the whole MS/FB collaboration is about? I actually think MS is ahead of everyone on Social Search and if they get Y!, it's going to catapult them ahead even farther as they're effectively paying to buy everyone's social graph. (MS has the vision and technology, but to succeed they require people's connections, which they haven't been able to aggregate on their own sites, so they rent from FB)
Y!'s sitting on piles of social data they've collected over the years, a much larger,richer set of data than Google's. Google's been able to use the data they've collected very effectively to this point, but Y! can beat them in Social Search, with or without MS. Y! has the technology and the data to do it. Google still has a lot to learn in this space. Delver looks cool, ( it's like Spock, but useful ) but they're starting from scratch.