Yesterday, Facebook announced the release of Graph Search, which will allow users to search Facebook using personalized, social data. If you’re looking to find a local restaurant that is recommended by your friends and people they trust, you might search for “denver restaurants liked by friends and friends of friends.” Or if you’re planning a trip to Italy and want to find the most picturesque spots, you might search for “photos of friends taken in Rome, Naples and Venice.” Or if you’re the jealous, stalker type, you might search for “ex-girlfriends of my boyfriend that lives in my city.” Whatever your situation, you will be able to do it soon.
According to Mark Zuckerberg, Graph Search will be very different than searching on Google. “Web search is designed to take any open-ended query and give you links that might have answers. Graph Search is designed to take a precise query and give you an answer, rather than links that might provide the answer.” As you might expect, Facebook is not working with Google on fine tune their search functionality.
Following are a few sample screenshots from Facebook:
Facebook Graph Search launched immediately in beta, and did not indicate when the full rollout would take place. During the beta, Graph Search will focus on four areas: people, photos, places and interests, with more features added in the future. Graph Search pulls from profile data of its 1 billion users to populate search results, and delivers items that will be unique to each user’s search parameters since each user’s friends and affinities are unique.
Graph Search functions within the current search bar. All results delivered will be from content already shared with users, therefore, Facebook states there are no privacy concerns. After the personalization layer, comments and engagement will be the next ranking signal (e.g., photo results will be sorted by engagement, with the ones with the most like and comments appearing toward the top).
If Graph Search cannot deliver a result, results will be pulled from Bing web search, but still displayed within Facebook. Bing explains the integration in a blog post:
“Now when you do a web search on Facebook, the new search results page features a two-column layout with Bing-powered web results appearing on the left-hand side overlaid with social information from Facebook including how many people like a given result. On the right hand side, you will see content from Facebook Pages and apps that are related to your search.”
Bing Web Search Results in Facebook Graph Search
“Imagine searching for Jay-Z concerts on Facebook, and not only finding Facebook content, but also web results from Bing including concert tickets, news about the tour and other web results—annotated with Facebook Likes and Shares. We think this is a powerful combination.”
What does this mean for brands? It’s making Facebook even more important as a means for sharing content and engaging with users, and may make Bing a larger focus for SEO and overall search marketing.
Say you are a women’s clothing apparel chain, you’ll definitely want to ensure that you have a Facebook Place page with accurate and up-to-date information. You’ll want to encourage check-ins in order to 1. increase reach and potential for showing up in personalized results if someone searches for “women’s clothing store in San Diego visited by my friends” and 2. increase total volume of engagements on your Place page, which is another layer of ranking signals. Sharing your products, videos, events and other custom content creates more opportunities for appearing in Graph Search results and broadening reach. And who knows what paid advertising opportunities Facebook has up their sleeve. But integrating search queries into Facebook Marketplace, similar to Google AdWords, seems like a natural step and high potential revenue generator for Facebook.
This is particularly relevant for SMBs, franchise companies and multi-unit businesses with a local presence. Local search is growing overall, especially on mobile (BIA/Kelsey estimates that mobile will generate 27.8 billion more queries than desktop by 2015) and with the success that Facebook has already seen with mobile ads (mobile ads generate 13x more clicks and 11x more revenue than web, according to one study).
While Facebook Graph Search is available now, there is still a waiting list to test the beta product. To sign up, learn more and get a small taste of the results, visit https://www.facebook.com/about/graphsearch.
I tried out a sample search to test Graph Search through my personal Facebook account. The test shows “Friends that live in my city.” I’m quite aware of which friends live in Denver, but it’s interesting to see of all the friends living here and which ones show up first. My guess is that other profile information is being matched to mine to create a (potentially) more relevant experience. The first result is a friend that lives in Denver, but that is from my hometown. I only have two friends that fit these criteria, and I engage with this one more than the other, so I would imagine those factors–direct relevance to my search, profile information, engagement and volume of mutual connections–are what is driving this result. The other two are colleagues of mine. I find it interesting that they show up over other colleagues, who I likely have more information in common with and engage with more often. I’ll have to play more with the beta to get a better understanding of the ranking factors, and I imagine we’ll never be 100% certain as all search algos are mysterious and EdgeRank has always been a particular puzzle.
I am very intrigued by Graph Search on both a personal level and for brands–there are numerous earned and paid opportunities for promoting brands and content, and I look forward to learning more as the product fully rolls out. Stay tuned for updates in the coming weeks.