Mark Zuckerberg

Why Facebook’s “Graph Search” is the Future

Last week, Facebook took the tech world by storm.  And for what seems like a first in a very long time, the news does not revolve around some privacy controversy or critical scandal.  On January 15, Facebook revealed its new “Graph Search” feature at its Menlo Park headquarters in Palo Alto.  Zuckerberg touts Graph Search as a fundamental addition to the Facebook experience, a “third pillar” along with the Newsfeed and Timeline.

Powered by the OpenGraph API, Graph Search is more than just a third pillar, it is the logical next step of Web 2.0.  If implemented properly, Graph Search will harness what can be done with social media data in revolutionary new ways.  But as with all potential groundbreakers, Graph Search has already faced its fair share of critics.  Bringing together the praise and the concern, we can paint a picture of the future of the web we know so well.

Harnessing the “Social Graph” 

First, to clarify, Graph Search is not a direct competitor to Google.  While Google allows users to find organization and clarity in the plethora of information available online, Graph Search allows users to find answers among their social connections.  In doing so, Graph Search harnesses the so-called “social graph,” the immensely complicated web of connections inherently created by Facebook members.

Graph Search thus lets you find who likes what, where, and when.  This new feature is more of a direct competitor to Yelp and Fandango than Google.  In short, Graph Search is an incredible feat of social engineering.  For one, Facebook engineers have actually made sense within the mess of the social graph.  And two, Graph Search is an advancement for natural language processing (NLP).  While Google relies on short, distinct searches for the best results, Graph Search actually works best when searching with more natural phrases.  A search of “restaurants in New York that are near my house and serve sushi” will work better than “New York sushi.”

Graph Search Justifies Facebook

One major take away I see in Graph Search is a partial justification for many of Facebook’s more controversial actions.  We love to hate Facebook, and for good reason.  Facebook’s privacy policies are shady at best; their IPO has been called a scam.  To date Zuckerburg’s use of our data has been tainted. 

Facebook is held high on a pedestal of wrong doing by many Internet activists, a symbolism of how in a free service we the users are the product.  But, Graph Search finally makes many of Facebook’s infamous overreaches worthwhile.  We too benefit from our own information.  Users can now interact with and benefit from their own social data, rather than merely facing targeting advertisements from Facebook’s business partners.

The Critics

Expectedly though, Graph Search has its fair share of critics.  Many people have and will find fault with my assertion that Graph Search may indeed be worth the privacy sacrifice.  Over at Forbes, Anthony Kosner argues that due to the new take on search that Graph proposes, Facebook could face many unintended consequences.  Most notably, he highlights that Graph Search provides a product that few users currently desire.

Multiple sources also reiterate a specific criticism: the OpenGraph API’s reliance on “likes” and user content.  Steve Cheney points out a fact that most Facebook users can resonate with: that our “likes” online do not necessarily correspond with what we truly like offline.  And over at TechCrunch, Semil Shah argues that many users do not send all their social data to Facebook.  As an example, I rarely use Facebook to “check-in,” foursquare is my default location app.  Because Facebook’s Graph Search relies upon the social data we provide, any distortions can lead to poor results.

The Future, Regardless

In my opinion, these criticisms are partially an extension of the popularity of hating on Zuck and Co.  Yet, it should go without saying that the most radical inventions deserve the most scrutiny.  Therefore, the critics should have their day.

And, in many ways, critics of Graph Search may turn out to be correct in their assertions.  But, Graph Search is the future, regardless, and here’s why:

  1. Graph Search is the next iteration of search in general.  Rather than layer social over traditional information searches (re Google), we can now search social itself.
  2. OpenGraph is a refreshing attempt at understanding the vast web of social connections we produce.  It is an advancement of data science and natural language processing.

 Whether Graph Search will be a revolution itself, or lead to products built upon its foundations remains to be seen.   But OpenGraph’s forward thinking take on search should not be shrugged off.  And just because users may not be “clamoring for this” does not mean the product will fail.  As the saying goes, we do not miss what we have never had.  Here’s to the future.

Billy Ceskavich

Billy is an Engagement Fellow and Masters candidate in the iSchool's Information Management program. He's an avid fan of consumer technology and the startup space, having worked in the industry both in Syracuse and Silicon Valley. He's active on Twitter (@ceskavich) or can be contacted directly via email (bceskavich@gmail.com).

More Posts - Website - Twitter - Facebook - LinkedIn


  • @MichaelFDownes

    Only time will tell to see if people love or hate Graph Search. I, too, think that this is huge for the future, and as you said above, “Graph Search justifies Facebook.” As someone who is getting pretty sick of the social network, now has me rethinking its value. I have used the feature already, and it definitely needs work. But in theory, this is huge. And once it hits the mobile application, it’ll be even bigger. I do have to say, however, that this will undoubtedly spark an uproar or two due to its privacy implications. I realize Graph Search doesn’t make changes to privacy permissions, but it does make it damn easy to find out certain information about people.

  • Pingback: Facebook just unlocked the reputation economy | LEWIS PR()