Last week, Ashoat Tevosyan, an engineer on Facebook’s search quality and ranking team, discussed the nuts and bolts behind adding posts to Graph Search. Wednesday, during the social network’s third-quarter earnings call, Co-Founder and CEO Mark Zuckerberg spoke about the reasoning behind the effort.
Responding to a question from RBC Capital Markets Analyst Mark Mahaney, he said:
To your question about how do we make the knowledge that’s been shared in Facebook more useful, the first thing we needed to do is just index it all and build the infrastructure to start being able to use it in different ways. I mean, the first data for Graph Search, we had indexed to more than 1 trillion connections, friendship connections, group membership, like connections, and then for post search, we indexed more than 1 trillion of the posts that people that people have put into the system.
The basic insight that we think we are operating on here is that right now, a lot of the behavior and engagement on Facebook is very day-to-day, where you are sharing something and Facebook is the best place for you to share photos or events that’s going on in your life if you go to News Feed and see what’s going on with the people around you.
What has happened is that over the past almost 10 years of this behavior, this amazing base of knowledge has been built up, but there are trillions of pieces of content and information that now we are just trying to find different ways to expose and basically make that more useful to people, instead of just the stuff that’s been shared in the last day or so.
Graph Search is one way that you can see that coming to light in terms of people being able to do directed queries for different types of content. There are other kind of services that we think we can build, as well, that give people more utility from the corpus of knowledge that’s been built out, and that’s going to be a big focus for us over the next few years.
Later in the call, Zuckerberg responded to a question from Sanford C. Bernstein & Co. Analyst Carlos Karjner with:
On your question about the graph of Facebook, the graph that people share on Facebook versus the Web overall, I think they are pretty different. In terms of quantity, they are getting to a pretty comparable size, so engineers on the Graph Search teams have told me — folks who have worked on other Web search engines before tell me — that the scale of things like post search are as big or bigger than any Web search index that’s out there, but they are they are just different use cases with different kinds of knowledge, and people are going to use them for different things, so our approach with Graph Search is not to build something that is Web search.
We think that companies have done that, and they are doing a good job at that, but there are different kinds of knowledge — things that you want more opinion on from people that you trust — that I think inside Facebook, we need to do a better job of servicing that, so that’s going to be the focus on that.
I mean, over time, there are a lot of possibilities for things that we can do, and I am not really ready to talk about a lot of them today, but you can kind of look at what we have launched in the couple of the Graph Search launches that we have done so far, and we are pretty early in that journey.
Readers: Have you started using Graph Search and, if so, what are your impressions of it?