Facebook recently released the beta of its new Facebook Graph search. In the past, Facebook searches returned specific objects – a friend’s name finds their profile, a brand name finds the brand’s pages or groups, and so on. Graph Search now allows users to search with natural search queries like you’d normally use on Google. For example, “People in Austin who like bikram yoga,” or “Coffee shops in Sydney my friends have been to.”

Graph Search’s current limitations are fixable

It’s interesting that Facebook is releasing the beta without surfacing results from text. In its current iteration, the Graph relies strongly on people tagging their photos and videos completely, with all needed information about location, context, etc.

Personally, it’s unlikely I’ll have been so comprehensive with my own photos. When posting a photo, I already know all of the attached details someone might tag – i.e. where I was, whose birthday I was celebrating, what mountain I was on, etc. – so I’m unlikely to include that information in my post. In researching for this blog post, it seems I’m not the only one with this perspective. Writes Jared Newman, TIME blogger,

“Unless you and your pals are putting lots of data in, you’re probably not going to get a lot of data out. I know for sure that I haven’t put much effort into connecting my real life story to Facebook, and as I poke around my network, I see that many of my friends haven’t either… Collectively, we haven’t invested in making Graph Search as useful as it could be.”

I’m much more likely to give rich info about what I’m doing in my status updates. Assuming Facebook knows this, status updates may simply be more difficult to sort/ put an algorithm behind; I estimate though, that teams at the social network are working on just that.

And we can expect sharing these needed contextual tags to become more frictionless over time. As more people connect to Facebook with mobile devices, for example, automatic geotagging of posts is just one way we might enable tags without any effort on the part of the user.

Graph Search’s potential benefits

If successful, Graph Search should make searching on Facebook much more helpful. This weekend for example, I went to New Orleans for a friend’s wedding. My fiancé and I posted all weekend long about costume parties (yes, a la Carnaval), “New Orleans,” and “Bourbon Street.” Several friends also at the wedding posted about road tripping and flights over the Mississippi. But today when I search for “New Orleans” in regular Facebook search, all I get is the Saints, the Hornets, and the city itself. Based on the current search functionality, the only way to go from here is up!

Chatting with some colleagues, we noted that Facebook could become more similar to a dating site. “So I can search for people who like things I like,’ then refine the search based on relationship, gender, and hometown?” asked a coworker. But this ability to narrow down a network by specific criteria would have much more uses than romance.

The network has always allowed users to connect with new people and perhaps make new friends, but largely through tools like “People you may know.” By finding people whom the people you know also tend to know, it’s possible new relationships formed. Graph Searches for people with specific interests that match your own could make this networking even easier, more popular, and targeted. My colleague Ian Greenleigh elaborates on the usefulness of targeted Graph Searches over simple social graph searches:

“Graph Search’s social graph based functionality isn’t as useful as it may seem, since recommendations from friends are in many cases completely mismatched and less compelling. For example, the results from a search for ‘music my friends like’ would likely be very hit and miss, since one’s Facebook friends reflect a wide array of individuals that are connected to you primarily because they know you, not because they are like you. I’m Facebook friends with family members whose musical tastes could not be further from mine, so their recommendation of the new Kenny G album means nothing to me.

On the other hand, Graph Search’s ability to harness the interest graph is where it gets interesting. If I search for ‘music liked by people who like Tool, My Morning Jacket, and Robert Earl Keen,’ the results I get will reflect the tastes of people like me. They will be relevant, interesting, and ultimately, more compelling to me as a Facebook user looking for new music.”

Will Graph Search revive some long lost Facebook behaviors?

From the Facebook Newsroom post, “When Facebook first launched, the main way most people used the site was to browse around, learn about people, and make new connections.” I didn’t remember that this is how we first used Facebook – but I do remember that there was no “News Feed.” I joined in 2004, and back then, there was no news feed, no chat, not even messaging! You’d just go to friends’ pages to see what they were up to. While searching around to see how I used Facebook back then, I found this:

That’s right – we used to navigate to a friend’s profile (before it was called a “Wall”), post something (see the post above on March 2), wait for the friend to respond, and THEN we could reply back (see follow up on April 7)!

These days, the News Feed “pushes” to us what others have posted already, and we can simply and immediately reply to their thoughts. Facebook is like a news ticker for our personal networks, streaming everything our friends are currently doing with no real organization.

Graph Search could help bring back some of the original Facebook activities that may’ve been more common before the news feed, such as looking for what people were doing last week, what experience they have had with a certain place, etc. In other words, Facebook may become more specific again in some ways. Just as we once had to decide whose profile to visit and what to check up on, we’ll be able to search for specific topics of interest.

So while Graph Search’s usefulness may be a tad limited today, one can see how the functionality can become much more useful over time. It’ll be interesting to watch how the technology develops – and whether it changes the way people connect on the social network.

  • http://twitter.com/txTDM Tara DeMarco

    Hi Krithika, thanks for your comment. I absolutely agree with your first paragraph — this is a chance for brands to get much more visibility (and more meaningful visibility) on Facebook if they take the time to post useful content and tag intelligently.

    Not sure about your second paragraph, but if I had to guess, I’d say Facebook probably does want to integrate Connect data into Graph Search. That way Graph Search becomes more like a regular search engine — if you’re looking for a TV, and you search “TVs liked by people who like Blu-Ray,” you’d find actual product pages on say, BestBuy.com, that people on the site Liked using Connect. Just my guess.

    Thanks for reading!

  • http://twitter.com/krithika123 Krithika Rosenthal

    So, hypothetically speaking if and when this is rolled out to everyone, brands that truly optimize relevant content could see some success? What I mean is if a travel brand posts photo of Paris, which is appropriately tagged and dated, it could show up in results when say I search for photos of Paris? And if the search is based in friend likes then the fan base and engagement also become a component?

    How about all the likes via FB Connect across sites, do you think those will also show up in search results? For instance if I like a pair of shoes on a site and my friend searches for “shoes my friend likes” would there be a greater likelihood of this being in the result even though it is not a post by the brand that I liked?