Facebook Graph Search is a social search feature the company announced on Tuesday, Jan. 15. The feature is currently in private beta with a waitlist for individuals and businesses.
You can join the waitlist here.
Facebook’s announced plan is to roll it out gradually to hundreds of thousands of individuals first (English only), then more broadly for PC-based users, then for non-English languages, and then on mobile. It is not clear how quickly this expansion will occur, but several Facebook product people are on record saying they still have work to do to figure out how to scale the computationally intensive searches across millions of concurrent users. (Think of crawling a user’s social and open graph connections across hundreds of thousands or potentially millions of nodes for every search.)
Non-trivial engineering challenges stand in the way of mass availability of this feature set.
What Does It Do?
It is a very, very cool feature. When I type in a query, such as “friends who have been to Rome, Italy,” Graph Search traverses all of my relationships and those of my friends to find people who have visited Rome. It then pulls back these people and displays them alongside relevant content.
This is a simple example that illustrates the difference between the kinds of results Graph Search returns and how search results from Google (or Bing) would appear.
Another key aspect of this feature is how it appears to include implicit affinities and experiences, in addition to explicit likes and shares people have done through Facebook. When you think about the significance of that, it’s pretty impressive. Based on the content I’ve shared, as well as the check-ins, posts and comments I’ve made – plus the images I’ve tagged, etc. – Graph Search can infer what I like, where I’ve travelled to, and so forth.
The inclusion of implicit affinities is only possible due to Facebook’s massive scale and could ultimately be the component of Graph Search that makes the results valuable enough to get people to use the feature.
What’s it good for?
Will Consumers Use It?
How Will Facebook Monetize It?
Facebook hasn’t announced how they will monetize the feature. The obvious opportunity is to charge for sponsored listings much like AdWords. There are a few other options as well, including:
What Does It Mean for Brands and Publishers?
It is going to take some time (several months at the earliest) for the feature to achieve critical mass. I do not anticipate Graph Search will be something brands or publishers will be investing in directly in the first half of this year if for no other reason than monetization of the feature is one of the things still being figuring out.
What Should Digital Marketers and Publishers Do About It Now?
The potential for the feature is HUGE. Brands and publishers can and should be doing a number of things right now to benefit from it as it reaches critical mass. A simple rule of thumb is that the more content that gets shared, liked or commented on through Facebook, the greater the chances of discovery of that content through Graph Search.
These are three things brands and publishers can do right now, which are in accordance with best practices:
More Background and Reviews
All of the reviews I’ve read are lukewarm to effusive. It really does seem like Facebook is on to something BIG with Graph Search:
Facebook seems like it is ready to launch a fledgling version of Facebook Live in the near future, Osofsky said. "We have an early beta programm for a…
San Francisco-based Twilio counts Uber, Open Table, and Nordstrom among its customers. In 2015, the company's revenues were around $167 million with a…
It is hard to imagine a vote on something- even for those of us in the U.S. consumed and amazed by the daily barrage of presidential election year pol…
It is hard to see Apple in trouble. It has massive reserves and remains one of the most powerful brands in the market, yet every week seems to contain…
It seems that SoundCloud sees much of the benefit in this partnership with Twitter in the rollout of their new subscription service that is meant to s…