Facebook Graph Search: It’s like asking friends out loud.

January 16, 2013 10:08 am Published by Bill Boorman

By now you will have read many times that Facebook has announced the launch of social search on the open graph. I have been able to grab a sneak peak of a beta version to get a few ideas of what this might really mean to users like us.

I’ve always felt that the real battle of the internet was not between LinkedIn and Facebook, but between Google and everybody else for users’ eyes. Users’ eyes are critical, because when a channel has our eyes, then it can serve content and ads to us based on our historical behavior. The more time in channel, the more opportunity to serve content, and the better the channel gets to know us.

The other big point around social search is developing a circle of trust: who do we want to make recommendations to us or provide us with information? When I want to do something or buy something, and I don’t have a solution or provider in mind, I want to ask a circle of trusted sources or friends what they recommend or what they like. I only want to step outside this circle and ask everyone if I draw a blank, and that rarely happens. It rarely happens because my network of friends are a bit like me. They’ve asked the question before, and they’ve reached their own conclusion.

When I was in Berlin recently I experienced this first hand. My friend Lars Schmidt, the talent acquisition director for NPR, had been to Berlin before me, and had left a few comments and likes on attractions via TripAdvisor. I’m quite a lot like Lars in taste; if he likes a restaurant or a museum, I probably will as well. When I was planning my days, every time I saw Lars had commented, I checked what he had to say, and I added three attractions based on his review.

Let’s consider how this works on a wider scale across Facebook. When I want to know something, find something, or do something, I can check what my friends have had to say on the topic, the place, or the service/product. I can see what content they shared on the topic. I can go to them directly if I want a bit more information, which is very powerful.

Let’s look at where the results come from: only my friends (not all of Facebook) and only the content or interactions they have chosen to show me. This is brilliant because most of my connections are in my inner circle. The ones I’m closest to show me everything, and the ones from whom I am more distant are open about less content. The closer and more open the relationship, the more I tend to trust the response. It is speculation on my part, but knowing the way that Facebook works, it wouldn’t surprise me to see a weighting of results based on relevance to the query AND relationship based on the EdgeRank ranking, giving me results according to who I have the strongest relationship with (most trust in), rather than the best match for my search terms. Interesting stuff!

What is interesting is that Facebook has gone for natural language search and applied semantics. This means that rather than typing in keywords or boolean searches to try and get the best results, I ask Facebook a question as if it were a person. It is just like asking my friends a question out loud. Instead of entering: “Coffee, San Francisco” to find the best coffee locally, I ask my friends who might have some experience of drinking coffee in down town San Fran: “Where is the best place to get a coffee in San Francisco?” or “Where is the nearest……”

In the past, I would have posted this question directly to my Timeline and then sat back hoping someone saw and answered my request. I might have been tired and thirsty by the time a reply came back. Now the same request gives me instant access to all the places my friends have checked into, liked, or commented on, and that match my question, instantly.

When there are no matches, I have the backup of results from a Bing search. I’m glad Facebook chose to go with Bing because the results are always different from Google’s. (That is why I always search both search engines when I’m looking for possible candidates.) I have other options to fall back on if none of my friends have said anything about coffee in San Fran.

I’m going to be really interested in how this works out for recruiting and sourcing. If I ask a technical question about Ruby on Rails programming (or any other discipline), the results of who is sharing or commenting on Ruby on Rails content will probably lead me to the right people already in my network. It’s going to be great for crowdsourcing when I want to view the sentiment on just about anything from my network. Ask a related question, see what they are saying, liking, etc. It might end up being as simple as asking Facebook, “Who works at Dell as a Ruby on Rails developer?” and seeing what comes back.

It seems like we’re not just stumbling onto graph search’s potential for recruiting. I found the transcript of the press conference and Zuck said: “’Let’s say we’re trying to find engineers at Google who are friends of engineers at Facebook.’ He typed in the query and found, not surprisingly, that there were lots of people who met those criteria.” That seems like a pretty sound endorsement for using Facebook to recruit!

Check out a video of graph search in action: [youtube http://www.youtube.com/watch?v=PjeppEy5e8M&version=3&hl=en_US&rel=0]

This is a great development for users, and I’m excited to see how it pans out. If I can get the answers to the questions I want to ask in Facebook where I spend most of my online time, will I spend less time asking Google and going anywhere on the web? Probably!

What do you think?

Bill

This post was written by Bill Boorman