Sramana Mitra: When you’re working within an industry sector where the data is all self-sufficient like a video game industry, that’s obvious. What is the state of the union in other sectors? What percentage of the e-commerce clients that you have are actually using social sign-in and trying to marry social graph data with their transaction data?
Dmitri Williams: Certainly, none of them are trying to do this. This is not on anyone’s radar. I’m educating and evangelizing as I go. For the most part, when we walk in, the answer to your question is do they have social sign in, and what percentage of their user base is using it. You can’t sample from the social graph. It’s not like when you look at the US population and you want every nth person. You do need every person on every node because if you’re missing people, it’s fundamentally flawed. You need a good chunk of people to be on it. Luckily, that doesn’t mean you need every person to have signed in because they will actually show up on their friends list.
For example, you may sign in with Google Plus and I may not but I’m in your Google Plus network and I’m on your list. My name shows up with you and also shows up in the spending or activity database on logs. Therefore, I’m captured that way. You can bootstrap into this. We think what you need is 10% to 15% of people to have signed in to catch everybody. There’s this six degrees of separation which is helpful.
Sramana Mitra: There are some other things that are going on. I’ll just use our example. I run an online business. Our business is a virtual incubation business that’s completely online and we have a variety of different ways that we offer our service to our audience. We have free online roundtables that happen every week. These are mentoring sessions that entrepreneurs from all over the world attend every week. That’s on a WebEx platform and to register for that, you have to register through Eventbrite. Eventbrite has social sign in available. That’s one place where there is a social connection for us.
We also have a blog as part of this system. The blog commenting is all on social. We also have sharing capabilities available through that as well. Do you have the ability for a company that has that kind of configuration to tie all this stuff in and draw those inferences?
Dmitri Williams: Yes, we actually are pretty agnostic to the edge data sourcing. In social graph, you call the people as nodes and the connections between them as edges. To some extent, we don’t really care where the edge is coming from. It just happens that the existing well-known technologies like Twitter, Google, Pinterest, and Facebook are the most adopted and the most likely to be used. We’ve spent time making sure that we understand them and can build graphs and connect them that way. Any connection between people is good enough for us.
This segment is part 5 in the series : Thought Leaders in Big Data: Dmitri Williams, Professor USC and CEO of Ninja Metrics
1 2 3 4 5 6 7