Someone recently asked me on Quora how eCairn scores influence. It was a great question, and one that I was more than happy to answer, but I think it was a good prompt to address this here. It’s also a good follow-up to our post on the two pillars of influencer analysis. First we need to quickly define influence.
We define an influencer as a person who, through writing and being read, affects the opinions of peers within a community around a certain topic. For example, Chris Brogan is an influencer of the Social Media Marketing community. He is influential because he writes and shares content that is widely disseminated throughout the community, and because that content largely affects or reflects the feelings of that community.
Some of our competitors have a different definition. They look at the popularity of certain sources across all industries, then try and correlate that to their ability to drive “action”. But influence is relative. You can’t take the expert social media advice of Chris Brogan and say that his word will affect or reflect the car enthusiast community just because he happens to mention cars a lot.
Influence in eCairn Conversation™ is ranked by analyzing the connections and the level of networking between blogs over time. The primary variable which affects influence rank is the number of times other blogs within the community link back to posts from the target blog. eCairn Conversation’s algorithm then ranks the blogs and determines whether the blog’s connectivity is statistically significant enough to be considered a high, medium or low influencer or if they have no influence at all. There are a few other factors which help to determine influence level, but they don’t carry as much weight. Below you’ll see how the sources in a community tend to appear in an influence graph:
You’ll notice that the influencers found using eCairn’s method are the true nodes of communication between bloggers. These people act as the links in the community who both build and transmit a majority of the content that affects the community and drives action. It’s most apparent when you look at a map of the influencers and highlight the high influencers:
Above I’ve colored the high influencers blue in the video gaming blogger community. Notice how they act as the centralized nodes of conversations.
There are several benefits to this method:
- It is less easy to “game” this system.
- You’re going to be sure that the targets are both influential AND relevant.
- There is little room for noise from irrelevant sources and bots.
Blogs provide their users the ability to broadcast original, in-depth and insightful content for their readers. When compared to the paltry character count of alternative social networks, we can see how blogging provides a better avenue for developing insight in the community. (Also, let’s face it, most of the best tweets link to a blog, right?)
But this doesn’t mean we can completely discount these other social media channels. In fact, it’s definitely a good idea to leverage these other channels to maintain your relationships in different circles. This is integral in any social media campaign, and it’s the reason why our application also supports aggregating the feeds from, and managing engagement with, these sources as well. Today, blogs are the core of the system. In the future, we’ll expand our algorithm to take inputs that also make sense of twitter, Google+ and Facebook.
Why use eCairn’s method?
By not relying on factors which the blogger has control over, we make it more difficult for the blogger to game the system. Other existing methods of influence ranking use factors like follows/subscriptions which can be artificially increased through the use of bots, fake accounts and collusion with other users. Post count is another easy variable to skew as well. It’s easy to create a content bot which can constantly publish or re-publish content that can be relevant and interesting, but in no way original or worth following.
Building your influence rank by starting with the community ensures that you find influencers who are relevant. Say you’re trying to find the experts on movie production. You don’t want their interactions with the music recording community to affect your influence measurement of the movie developer community. You’re trying to find the most influential people on a specific topic, not the most influential people who happen to have an interest in that topic. Since you build the community, you control the level of noise that can come from irrelevant sources and can manage the objectivity of the ranking.
Posted on eCairn's blog: November 10, 2011