3 ms·
It seems like their model is too simple to draw any firm conclusions? It essentially treats all types of content the same. That is, what about this causality fl
by skybrian 8y ago
It seems like their model is too simple to draw any firm conclusions? It essentially treats all types of content the same. That is, what about this causality flow:
group -> type of content generated -> audience reach
People in different groups will prefer to author different kinds of content at different rates, and this could entirely account for audience reach.
For example, suppose we classify content by the language it's written in? One would expect people who write content in English or maybe Chinese to reach a larger audience than people generating content in a less well-known language.
But there are plenty of other dimensions that will affect audience reach.
- doomlaser 8y agoEven with a perfectly time-ordered feed with sharing/retweet functionality, a pareto distribution will emerge among content that bubbles up. That's how networked economies seem to work. http://www.labs.hp.com/research/idl/papers/ranking/adamicglottometrics.pdf http://www.labs.hp.com/research/idl/papers/ranking/adamicglo...