Inside the Interest Graph (Part 2): Defining The Interest Graph

Inside the Interest Graph (Part 2): Defining The Interest Graph image blog4Inside the Interest Graph (Part 2): Defining The Interest Graph

Now that we have introduced the concept of the interest graph and why it’s the ultimate recommendation engine for brands and technologists, we’ll explain the basics of what the interest graph is and how it’s mapped and analyzed to offer relevance to businesses.

What Is the Interest Graph?

The interest graph is made of Likes, Follows, and other social relationships between people and things, products, or brands. As Naval Ravikant, founder of AngelList, described in TechCrunch, the interest graph is asymmetrical, organized around interests, public by default, and aspirational.

Relationships in the interest graph are asymmetrical, meaning they’re one-way follow relationships (not two-way friendships). Users can follow or like @Rihanna without Rihanna being required to follow or like the users in return. This means that they’re organized around interests, not friendships. In fact, friendships are described by the social graph, a picture of who knows whom. The social graph has been explored for various purposes, including advertising, but its best use tends to be limited to making recommendations for more friendships. (You may have seen the “People You May Know” sections on LinkedIn that encourage you to connect with more people.) It’s important to make the distinction that knowing each other in real life isn’t required by the interest graph. You can like Tom’s Shoes, Converse, and Puma without having ever visited their retail stores.

And because mutual following is not required, people can follow based on what they want versus who they already know. This means following is aspirational. You can follow @BillGates on Twitter because you admire his philanthropic work and want to learn more about it, without Bill Gates needing to decide you’re worth following in return. Asymmetrical following emphasizes interests because it helps people reach for their wishes and hopes.

Due to the nature of social platforms, interests (and the interest graph) are public by default. Information about who and what people follow is standard public information in user profiles, which means that it’s not only revealing, it’s noninvasive too. (However, users can make their profiles private, and some ad platforms use this as an automatic opt-out feature.)

The Blended Interest Graph (BIG)

Every platform has its own interest graph. Facebook data helps us create a picture of user interests via the Like button. You can do something similar with Twitter data by analyzing interests expressed via the Follow button. Every social platform has its own version of the one-way follow relationship (on most platforms, it’s called “Follow”).

Because all social platforms also have implemented APIs and/or data streams, each proprietary interest graph is available to be analyzed in aggregate. For the purposes of this article, we’re talking about all of these secondary interest graphs taken as one — the Interest Graph. Think of it as:

  • The Facebook Individual Interest Graph
  • and Twitter’s Individual Interest Graph
  • and Google+’s Individual Interest Graph
  • and Foursquare’s Individual Interest Graph
  • and Pinterest’s Individual Interest Graph
  • and Instagram’s Individual Interest Graph

…and on and on. The sum of all individual interest graphs is the Blended Interest Graph (BIG) for online social platforms.

Inside the Interest Graph (Part 2): Defining The Interest Graph image Blended Interest GraphInside the Interest Graph (Part 2): Defining The Interest Graph

BIG = Big Data

The interest graph is made up of basic social building blocks like Facebook Likes and Twitter Follows. 140 Proof CEO Jons Elvekrog likes Domino’s Pizza on Facebook. That’s one data point for the interest graph. For example, 140 Proof CTO John Manoogian III (@jm3) follows creative agency @Mekanism and Forbes journalist @a_greenberg on Twitter. That’s two data points for the interest graph.

These individual data points add up to a huge set of information. And the interest graph is growing at an accelerated pace. Every relationship in social is represented by the interest graph.

As new users join and follow influencers, the data set grows by 2 billion Likes every day — six Likes for every person in the United States, every day. Put another way: the current world population is estimated at 7 billion, and the interest graph is about 40 times larger — currently sitting at around about 266 billion Likes and Follows.

The interest graph is a picture of the present moment. To analyze the interest graph is to understand what’s happening right now. Unfollows are discarded and new follows are included and folded into the analysis.

Inside the Interest Graph (Part 2): Defining The Interest Graph image InterestGraphDataPointsInside the Interest Graph (Part 2): Defining The Interest Graph

How Do Social Companies Analyze the Interest Graph?

Understanding and mapping the interest graph requires a dedicated team steeped in information theory, big data architecture, and lightning fast calculation. The biggest challenge in harnessing the power of the interest graph is making millions of decisions in real time. With over 2 billion new interest signals every day, any delay in processing means relevance could be compromised. At 140 Proof, an elastic architecture composed of hundreds of cloud servers grinds public social data, collects interest signals from social platforms, analyzes the data, and makes rapid decisions about people and personas. Dedicated data scientists, engineers, and statisticians ensure that computation happens not just instantly but accurately.

Contributors: Jon Elvekrog, John Manoogian III, Vanessa Naylon & Lau Ardelean

More Tech articles from Business 2 Community:

See all articles from Business 2 Community

Friend's Activity