Hershberg.org

They certainly picked a questionable day to make the announcement, but as of Wednesday, Google Social Search is available to all Google users and it’s definitely a big deal.  Marshall Kirkpatrick probably summed up the reason why best by saying:

This is a very big step. What’s your portal to the Internet: Google’s algorithmic search of the Web at large, or your social circle of people on Facebook? That’s the battle for the future that Google and Facebook are waging now, and Google Social Search is a big move.

Of course, the prerequisite for using Social Search is the creation of a Google Profile.  These pages can include various pieces of information about a person, including companies worked for, schools attended and most importantly, links to social profiles created on other sites across the web.  Google users who have added at least some of this information to their profile pages can now use Social Search to see a customized set of search results consisting of content created or shared by people in their social networks, including Twitter, Flickr, and Google-owned services like Gmail and Google Chat.  I spent some time with it earlier this week and while Google’s record of my “social circle” (defined as friends and contacts) seemed pretty outdated (it included people I had stopped following on Twitter months ago), it’s a major step towards meeting recent calls for a service that allows our questions to be answered by our friends.

Needless to say, the ultimate success or failure of Social Search hinges on Google’s ability to get a significant number of users to create their own profiles — something it has struggled with in the past.  Absent that information, it doesn’t have the ability to make sense of someone’s social circle.  But what I think is perhaps less obvious, yet equally as important, is that Google’s chances of owning real-time search also depend on its ability to have users create a critical mass of Google Profiles. I think Google knows this and that we’ll see them make Profiles a priority in the future.

The Relationship Between Google Profiles and Real-Time Search

I’ve been very critical of Google since it launched its real-time search feature in December.  For a company that obsesses over always offering a simple and clear user experience, the “Latest Results” section of a reply page seems surprisingly out of synch with the rest of the page’s content – almost as though it were implemented as an afterthought.  Additionally, I’m often confused as to when Google is going to show real-time results for a given topic and even more importantly, why and how its decided whose content to feature.  Generally speaking, I’ve found Google’s real-time results, particularly tweets, to be irrelevant and pretty mindless.

So, I was very interested in Technology Review’s interview with Google Fellow, Amit Singhal, the person responsible for Google’s ranking systems and who led development for real-time search.  When asked about methods for determining tweet relevance, Singhal likened it to PageRank, the process Google uses for ranking web pages, saying:

“You earn reputation, and then you give reputation. If lots of people follow you, and then you follow someone–then even though this [new person] does not have lots of followers,” his tweet is deemed valuable because his followers are themselves followed widely.”

I’ve spent quite a bit of time thinking about this statement over the past couple of weeks and it makes so little sense that one of two things must be going on here.  Google is either absolutely clueless as to how the process it developed for ranking web pages can be properly applied to tweets (possible) or Google is being as opaque about what it plans to do here as it is about pretty much every other aspect of its business (likely).

To begin with, when determining the relevance of a tweet, the absolute number of followers that user has is essentially a worthless metric.  It’s basically the equivalent of saying that web pages with lots of links are inherently authoritative, regardless of the subject.

Singhal’s stated approach to ranking tweets also assumes that everyone is a generalist and that they can speak about any topic with the same amount of expertise when that’s obviously not the case.  For instance, while I would put a lot of stock into what Paul Krugman, who has ~350K followers on Twitter, has to say about economics and politics, I wouldn’t turn to him for an opinion on music.

Google must find a way to account for the fact that tweets from the same user about different topics don’t necessarily have equal amounts of value.  Looking at number of followers alone does not come close to accomplishing that.

Separately, I can’t understand why people (and companies) continue to make a distinction between reputation and authority on Twitter and reputation and authority in the real world — they are one and the same.  If someone is a proven expert in her field or has firsthand experience with a given topic, shouldn’t we value what she has to say about it regardless of how many people follow her on Twitter? 

Prior to signing agreements to have tweets included in Google and Bing search results, Twitter was basically an island, which made it very difficult, if not sometimes impossible, to match someone’s Twitter account with the rest of their social profiles.  The combination of Twitter’s data feed and Google Profiles changes all that.  Think about all the information about each individual user that Google now has the potential to thread together:

• LinkedIn profile: Includes work history, schools attended, current location, links to other personal websites and recommendations from friends and co-workers.
• Feed reader: Provides insight into the types of content I’m interested in consuming.
• Geo-location information: Provides verification that I checked-in at a specific place or event and could therefore place higher priority on my firsthand experience.  A huge improvement over the current hashtag “solution.”
• Disqus profile: Blogs I moderate and those I comment on.
• Last.fm, Pandora, Lala, etc profiles: History of types of music and specific artists I listen to.
• Photo albums: Provides insight into the places I’ve been.
• Etc, etc…

Google is the only company I can think of that is likely to give enough incentive for users to provide this information (access to Social Search) AND that has the ability to fully leverage it.  As a result, Google will be able to index and then properly filter and rank tweets based on an aggregated and (nearly) complete picture of each user’s profile — not just the number of followers someone happens to have.  That is how real authority and experience should be determined.  And if that’s the case, there’s no way that even Twitter can rank tweets as well as Google can — it simply doesn’t know enough about each of its users.

Google’s history with social media absolutely sucks and past attempts to get users to create Google Profiles have resulted in very little success.  But I am not ready to count Google out just yet. If it can properly communicate the benefits of creating a profile – which will require significantly more promotion than serving an ad to people googling “me” (literally) – I think Social Search will get some traction.  And Google knows the profile information users share won’t just improve their personal search results, they’ll help Google improve a major component of its real-time index as well.

A couple of related notes:

The point has been made that Google’s Social Search is dependent on many of the services mentioned above making all of their data available to Google.  I agree.  Still, Google does not need agreements with any of these companies to index public profile pages on each of these sites.  That is apparently what it’s currently doing with Facebook.  When aggregated together, the information contained on those public pages will generally provide enough insight to give Google a more complete picture of the person behind each Twitter account.

Also, I don’t believe that Twitter search and real-time search are synonymous with each other, though this post probably reads that way.  That is because Amit Singhal was discussing how Google ranks tweets exclusively in the context of the real-time web.  As Google improves its ability to determine the relevance of tweets, it’s likely that they will be featured in broader search results as well.  

Follow me on Twitter: @hershberg