Hello! How was your lunch? Mine was most delicious. I’m just kidding. Livebloggers don’t eat lunch, silly goose! But enough about my starving belly – up next we have Ronald Yager and Rick Berry. I’m a little nervous about this one, mostly because the speakers look like college professors. I think we’re all about to get our minds blown.
We’re here to talk about social network analysis…interestingly, neither one of the panelists have a Twitter account. I almost want to go talk to them and find out what that’s like. And if they’re at the right conference. And what they do with all that free time?
Li Evans is moderating this session and says that before you do anything in social media, you have to do research. The guys on this panel have done a lot of research. Hold on to your hats.
Up first is Dr. Yager. He goes to the podium to introduce himself because he feels more comfortable talking there. Heh. Okay then. Dr. Yager looks a bit like a mad scientist. Think Doc from Back To The Future. This should be good. Or maybe scary.Dr. Yager says there’s an exponential increase in universal interconnectivity. We need technologies to help us better take advantage of this environment. We’re going to talk about intelligent social network modeling. We’re talking about personalization based on a Person’s Web Persona (P3). We’re modeling how a person is getting information and how they read it.
Querying social networks – Huge amounts of information exist in social networks. Just like we query a database, we want to query social networking using marketing relevant concepts. But there’s a problem here. It’s the ability to express our queries in terms of network relevant concepts and still express ideas meaningful to human comprehension and marketing focus. He knows we all understand statistics so he’s not going to harp on that too much.
First you have to locate the clique. Then look inside of it. Ask yourself:
- Who are the trendsetters?
- Who are the leaders?
- Who are good listeners?
- What is the strength of the relationship?
Clique’s Macro Concepts
Categorize them – Is the clique artsy, techie, fashionable, into sports? Are they cost conscious? Are they baby boomers? [If they’re online, jury says no ;)]
- Provides a more human-friendly summarization of data than means and averages
- Substantially more informative
- Easy for managers to make decisions
Intelligent social network analysis requires communication, cooperation and coordination between man and machine. We want the human to express their information and then represent it in machine language. The major difficulty is that humans communicate in language and machines communicate in mathematical formulas. We have to find a way to bridge this gap.
Bridging the Gap
- Should be human focused
- Communal Vocabulary
- Enable machines to comprehend and manipulate linguistic concepts
- Linguistic concepts are fuzzy and imprecise.
He’s talking about fuzzy computing. Is that like fuzzy math? I remember fuzzy math. He says it extends the capabilities for analyzing social networks by enabling the use of the human like concepts with fuzzy logic technologies.
Fuzzy Set Theory
- Allows representation of human focus linguistic concepts in a formal way
- Has operations that allows machine manipulation to linguistic concepts in a manner similar to human reasoning. It provides the bridge between the man and the machine. [Are we teaching machines to speak like humans? Don’t we know how this ends? I see a Will Smith movie in the making.]
He says we’re going to fall behind in this world if we’re not good at mathematics. Um, yeah, then I’m screwed. I need someone to come hold me. People are confused. This guy is way smarter than everyone in the room.
A relational social network is a set type object. Things are connected. He says something about nodes. Oh, oh, he says if you don’t understand what he’s talking about, don’t worry about it. It won’t matter in a minute. Heh. Thank god.
Inspiration and Motivation
Formal representation of social network is a set of theoretical object.
Framework for Intelligent Social Network Analysis
Take human focused concepts, represent them in fuzzy set representation, then take a social network and represent it in terms of a relational structure. Now we have a structure and we can connect it. It’s a bridge. Or something. We need to have a communal vocabulary – a collection of terms between the man and the computer.
- Inter-species communication uses vocabulary
- Man uses linguistic term
- Machine uses fuzzy set representation
- Man determines the content of the vocabulary
Things we want to model
- Strength of Relationship: This is a network-based concept
We want to search social networks in a way that you search databases. Everyone here can deal with a database (um, not this dummy). We want people to be able to query social networks in the same way. The network view – you have the nodes in the network and there’s a connection between these people. On top of that, everyone has attributes (age, friendships, etc). One has to be able to deal with this type of environment. One can query using Rich Concepts – a way of representing human concepts. We start off with basic ideas we can measure and we build out sophisticated concepts. Love is a sophisticated concept but maybe we can break it down to simpler things like buying someone presents is a sign of love (oh, that’s a good path to go down). That is easier to identify.
He’s done and everyone in the room is wondering what the hell just happened. We went from search conference to doctoral thesis in 10 minutes.
Nick is up next. I hope he speaks in English, not Math.
Facebook, just how big is it? There are 6.8 billion people on the planet. 30 percent of the connected world has a Facebook account. If Facebook was a country it’d be the third largest country in the world. What’s the most popular windows app of all time? It’s not Word, it’s not Notepad, it’s Solitaire. People like games.
Activity on traditional gaming sites is falling 20 percent a year. Not because people are playing fewer games but because they’re now playing them on Facebook, not on traditional gaming portals. He says you want to build your brand inside Facebook because it’s the Internet inside the Internet. Facebook isn’t the only social network in the world but it’s pretty large. He shows a lot of smaller ones. No one really cares.
Facebook gaming is big business. Example of top games
- Farmville: 84mm monthly user. 32mm people use the application once a day. 40 percent of customers visit every day.
- 25 percent of Facebook’s population comes from the US.
There’s a huge number of college-aged people using Facebook. Most of the mobile people are college-age. There’s also a huge number of users 65+. He shows the different gender make ups for all the different countries. It’s interesting.
Games and Interests
Bejeweled Blitz – Lots of stay at home spouses (he means women) playing BeJeweled.
Call of Duty – 92 percent male
He shows the difference in audiences for Star Wars and Pilates; Jay vs Conan; Twilight; Star Wars vs Star Trek. It helps creates corrolations between “people who like X” also like “Z”.
Glee: What are fans of Glee interested in on Facebook? The Sims, Monopoloy, Tetris, Wii, etc. They like Jersey Shore, MTV, America’s Next Top Model. They like Taylor Swift, Lady Gaga, etc. They like Pepsi over Coke or Redbull. Shows how you can use this information for targeting.
Micro Transactions: People used to sell packaged items. In Facebook, you don’t sell thing. You provide things for free and get money in different ways. Are you targeting your game to the paying audience?
- Bling: You can dress your character up in hats and swords. They don’t affect the game, but people will pay money for them for vanity. You don’t have to get everyone to pay.
- Accelerators: Kill 10 more ORcs or Pay 10 gold coins so you don’t have to pay. If you’re busy or you’re competing, you get to use your time better.
- Power ups: Pay to get 10 gold coins to get 10 percent attack bonus for 12 hours.
The economic beauty of micro-transactions
Once upon a time, games were sold for $19.95. Then they said, “lets sell them cheaper”. In doing that, they sold more games. With micro transactions there’s no fixed price, instead there’s the long tail. There’s also the Tall Tail – people who spend $100,000 a month on their games. I…who ARE these people?
Big brands are starting to get it – larger brands are getting involved in Facebook. When playing games, people are at 95-99 percent focused attention. This blows away all over media. With just 15 seconds to a brand, there is >80 percent unaided recall. There’s a honeymoon period for brand owners with fans. 43 percent of fans visit Facebook several times a day. 50 pecent of fans come to FB for entertainment. 92 percent say that being a fan has a positive impact on reccomending to friends. 84 percent of fans consume the brand in real life. There is no disconnect between brands followed on Facebook and regular consumers. The average person fans 9 brands.
We’re now on the second generation of social games. You need to get the mechanics reasonably polished before general release.
That’s it. This was quite possibly the most confusing panel ever. Lots of really interesting stats, but we flew through them all. I need a nap.