BuzzFeed’s Pound: The future for data
We have all experienced our Facebook newsfeeds being taken over by BuzzFeed posts. We often see the same post shared over and over again between different members of our networks. BuzzFeed posts are so successful due to social sharing. Instead of telling the audience what they should be reading, they allow the audience to tell one another what they should be reading and sharing. But the success of many BuzzFeed posts is largely attributed to experimentation, which is only valuable if you can analyze data to see where the experiments were successful and where they failed.
BuzzFeed’s traditional web analytics was only able to provide limited insights about their published posts. Their previous technology revealed how many shares each post had on each social media platform. However, that leaves out a huge portion of the story; exactly what path the post has traveled. BuzzFeed’s new technology, Pound, allows data scientists to look at how posts travel throughout the web, instead of just how many times they have been shared. Pound stands for the Process for Optimizing and Understanding Network Diffusion. Every post has thousands of initial organic sharers that come directly from BuzzFeed’s website, but the majority of shares come later from users on social network sites. This process forms a tree-like structure, but Pound discovered that looking at multiple “trees” at once reveals valuable details about the path at each post. Now Pound looks at forests instead of singular trees. The information BuzzFeed has received from Pound will allow them to create more shareable content and predict the paths of the posts.
In class we have looked at networks and seen how strength of ties affects sharing of information. In the future Pound might be able to incorporate tie strength into its analysis. This would allow BuzzFeed to understand what types of posts are shared between strong versus weak ties.