In a previous post I discussed Facebook’s EdgeRank. It’s the algorithm that determines whether an update does or doesn’t appear in fans’ newsfeeds on Facebook.
A high EdgeRank score is important to have a successful Facebook Fan Page. But did you know that Facebook has another algorithm targeted at applications? This algorithm is called GraphRank.
What is GraphRank
Graph Rank decides if an update generated by an application, or app for short, appears in fans’ newsfeeds. So, for Fan Page owners, who use apps for marketing purposes on Facebook, it is important to know exactly how this algorithm works.
Facebook app owners & developers, take note
Essentially, Graph Rank is to application owners and users, what EdgeRank is to Fan Page owners. GraphRank consists of a set of rules that are part of EdgeRank but are only applied to updates coming from apps.
The rules that GraphRank applies are personalised and tailored to each end user. This means that the updates that will appear in a fan’s newsfeed depend on his or her interests and the level and type of interaction between user and app. The algorithm is designed to favour activities that attract a high level of engagement.
User’s friend’s tastes are factored into the algo
The algorithm also takes into account friends’ tastes, and it will alter its calculations if the user’s friends list changes. For example, when a user has many friends that share the same interests, GraphRank will include a multiplication that results in more weight being given to updates from apps that fit these criteria.
Proximity and shared interests are also factored in
Similarly, when a large number of people with common interests are checked into locations in the same area, another multiplication will happen that affects which app updates will appear in their newsfeeds. So, proximity and shared interests are two important factors that GraphRank looks at to determine rankings.
Under the Hood – The GraphRank Algorithm
GraphRank’s formula is very similar to that of EdgeRank.
coefficient (or affinity) + weight + interactions + time decay= GraphRank
The coefficient, also called affinity, reflects the relationship between users with shared interests or a high level of activity between them. Proximity is also considered to play a part in affinity. To that number weight is added. Weight has to do with the type of update.
Different weights for different update types
Video updates have more weight than picture updates, which in turn are more important than link updates, and these have more weight than status updates. This order is completely up to Facebook, and they can change it at any moment.
New features are given more weight
Generally, when Facebook comes out with a new feature, more weight will be given to the newcomer as a way of promotion. Next, interactions are added. This is the number of likes, shares, and comments an updates has. It is similar to edges in EdgeRank. The last factor that is added is time decay. The older an update becomes, the less likely it is to show in newsfeeds.
In a nutshell…
An update from an app is more likely to appear in a fan’s newsfeed if the fan has interacted with the app recently, if the update itself has attracted a high level of interaction, if the update is a video or an image, if the update is recent, and if the fan’s friends have also interacted with the app or update.