THE BLOG
02/01/2013 10:29 am ET Updated Apr 01, 2013

What You Need to Know About Facebook Graph Search

2013-01-30-FBSearch2.png

Facebook announced on January 15th the beta rollout of a vastly enhanced search capability they have dubbed Graph Search. As the name implies, Facebook eventually will give all their users the ability to search their friends' profiles and activities (part of what makes up Facebook's proprietary "Social Graph" of information) for a variety of results.

What Marketers Need To Know

• This elevates the importance of ongoing engagement over total likes
• Rich content will perform best
• Local businesses may have an incentive to create/up their presence on Facebook
• The natural language search that's powered by a user's social graph is fundamentally different from traditional search

How Graph Search Works

The Graph Search toolbar will appear as a header throughout Facebook. The search results act as their own page, with the query as the page title - this page can be shared as a piece of content in its own right by users, and the title can be re-edited to place the results into a new context.

Results Powered by the User's Social Graph

The search results an individual will get are powered by his or her unique social graph - in other words, relevancy and results will be determined by the content and pages a user has interacted with (Likes, Shares, Comments), a user's friends, and the content those friends have interacted with (and by extension their friends' friends). Any privacy restrictions that users have enabled will extend to search results automatically.

For now Graph Search can only surface results in four primary categories: people, photos, places, and interests. To help users get used to these parameters Graph Search also gives some suggestions of categories to start with and to build on.

Natural Language Queries

Unlike Google and Bing, where users tend to string words together as keywords, Graph Search is designed as a Natural Language Search Engine - think of how Siri works or Wolfram Alpha. In practice human nature will likely still reduce this down to shorter phrases like "friends in New York who like thai food."

Who Can Use Graph Search?

Graph Search is invite beta only, with a slow roll-out. It is not enabled on a brand basis (i.e. page admins will not be able to search fans) but is solely meant for individuals. Presently there are no public plans to expand beyond US English-speaking countries. On the Beta Invite page located at https://www.facebook.com/about/graphsearch you can see a sample search based on your own Social Graph.

What Are the Limitations of Graph Search?

Results

If Graph Search focuses on four main categories, what is being left out? Two big areas - Open Graph Actions (ie songs Listened to on Spotify or photos Pinned on Pinterest) and Posts (ie status updates) are not included. However this is likely to change as the Graph Search is upgraded.

APIs

There is currently no Graph Search API, nor is search included in the Facebook Insights API or ads API. This means that reporting on performance of brand pages or content in Graph Search using Facebook's Insight tools will not be possible for the time being. There may be some workarounds for shared content such as images or videos however.

What Does Graph Search Mean for Marketers?

Engagement Still Trumps Page Likes

At present, the same rules for optimizing content in people's newsfeeds holds true for search results as well - content that is commented on, shared, and Liked by a large number of people is more likely to rank well and pages that have more of this content will rank higher by default, even over pages with more Likes. On the other hand one of the sorting mechanisms is the number of Likes a page has received so this would appear to bring added value to the total number of Likes regardless of engagement a Facebook page has. This is welcome news to marketers who may have spent to gain new fans only to find Facebook's algorithm limited their presence in the newsfeed to a mere handful of them. Still, while users will be able to sort by number of Likes, it is unclear how many will choose to do so. Engaging, relevant, and useful content wins.

What Is a Like, Really?

In theory, a user's Social Graph should make his or her results more relevant and personal and therefore more useful. In Facebook, this is complicated by the fact that Likes are not necessarily synonymous with endorsements or testimonials. For instance when Facebook moved to a structured data platform, people's freeform descriptions of themselves in the About Me section were transitioned to Likes that were meant to be equivalent but sometimes were not. The phrase "I miss Saigon" might be translated into Liking "Miss Saigon." Some users simply accepted the new format without editing these.

A New Era for Social Search?

While the idea of social search optimization has been around for some time now, search as a tool is used in many ways by different platforms in social, and social platforms are inconsistently displayed by traditional search engines. As social has become a huge engine of content distribution, sites like Pinterest have mirrored YouTube as discovery points for users looking for specific content (versus Twitter where search tends to be clustered around finding people, as well as conversations tied to events happening in real time such as the Super Bowl). Where Google has salted its search results with Google+ data, the Google+ platform hasn't had enough wide adoption from users and businesses to make these enhanced results truly useful. Meanwhile Google's YouTube is the Diet Coke of search engines, ostensibly a niche but a strong number 2 in the market.
This has not escaped Facebook, and the fact that photos and shared content are highlighted as primary search queries shows that they feel that content navigation and discovery are one of the keys to revenue growth on their platform.

Down the road it is possible that results in Bing could be influenced by Facebook and vice versa on the Web tab. Given that Facebook queries will be radically different from the way users search in Bing this might open up a whole new aspect of SEO.

Advertising

Sponsored results continue with the existing ad formats and no update has been announced. Expect some more targeting parameters and potentially the ability to work with Facebook's Strategic Preferred Developer partners such as SocialCode and Shift once this is included in the ads API.

Can Facebook Alter Native Behavior?

Will Facebook users habitually search for anything beyond their friends' names? Presumably the publicity around Graph Search will lead users to try it out but if the results seem iffy or not useful they may not have a reason to use it twice.

This is complicated by the fact that tools like Facebook Places are used inconsistently by users. It remains to be seen if there are enough check-ins at, say, restaurants within Facebook, to make a search of "friends who have had Italian food in New York" a result that leads to action.

What Does Graph Search Mean For The Future of Facebook?

While Google approached social search from the standpoint of using their dominant search engine to jumpstart a social network with Google+, Facebook has taken the opposite approach - Using their dominance in social to push into search.

Facebook has been sitting on a mountain of rich user data, particularly the interconnection of behavior and relationships that powers the Social Graph. In almost every case the data has been leveraged for "push" discovery to consumers - seeing how many friends Liked a post, a page, or an ad for instance. For the first time, the data is being used to power consumer driven "pull" discovery - show me a restaurant, a jazz song, a picture - that is liked by or shared by friends.

In turn this adds a new layer of data, "handraising", that Google has traditionally been master of. Presumably if I search for friends who Like or Checked In at Italian restaurants in Newark, I am also thinking of going to an Italian restaurant in Newark.

This may go hand in hand with a further push into reviews and local pages, assuming there enough volume and referrals are generated from Graph Search.

Facebook could also decide to further the relationship with Bing and serve up Facebook enhanced results to logged in users on Bing's platform - arguably a huge advantage over Google in the realm of social results and another potential place to leverage ads. In addition, if Graph Search becomes even a moderately used tool on Facebook's platform, it enhances the current Bing relationship as those results will continue to show up under the Web tab in Graph Search, and might be expected to lead to more conversions as volume increases.