The Mavrck Platform utilizes a patented algorithm to determine the Influence ranking of an individual on a given social network. This algorithm considers approximately 20 different signals about the individual and their posts to determine the level of engagement on their posts that they are driving from their friends/followers on a social network.
At a high level, the algorithm calculates a ratio between the reaction (likes, comments, shares, clicks, conversions) an individual earns from their posts versus how much reaction they should earn.
A key aspect of the algorithm is that having a large audience can cause someone to have less calculated influence than someone with a smaller audience.
For example, let's say two individuals each earn 500 reactions every month from their posts, but one has 1,000 friends and the other has 2,000 friends. The Mavrck algorithm will likely determine the individual with 1,000 friends to be more influential than the person with 2,000 friends.
The Mavrck algorithm applies weighted factors at three different levels:
- Level 1: likes, comments, and shares on a Post
- Level 2: the Post itself
- Level 3: the Individual who created the post.
Additionally, the algorithm only looks at the most recent 90 days of activity from the individual on the social network, since influence is extremely dynamic.
An example of the factors considered include:
- How often an individual posts
- How private the post
- How old the post is
- How often a friend/follower engages with the individual who created the post
- How many friends/followers the individual has
- What type of media is in the post (e.g. link, video, photo)
The Mavrck platform plots individuals on a 0 to 100 percentile curve by taking that individual's calculated influence score and comparing globally to all individuals ever scored by Mavrck.
This percentile should be used as a predictive indication of the level of engagement an individual can drive when creating content on behalf of your brand.
Algorithm Patent Number: US20130179511