How Online Content is Received by Users in Social Media: A Case Study on Facebook.com Posts

In social media platform such as Facebook.com and Twitter, there are many settings in which users can publicly post content. A number of these sites offer mechanisms for other users to make responses to these content: a canonical example is from Facebook.com, where posts come with annotations like“1,492 people like this”or“view all 307 comments”. Usually these user-generated content have the effectiveness of gaining influence for their publishers in social media by getting positive comments or “like”. The influence gained from users varies widely across different posts, and reasoning about the effectiveness of a post for gaining influence is an important task in social media analysis. In this paper we develop a framework for analyzing and modeling how the online content get influence in the social media platform, using a large-scale collection of Facebook.com posts as the dataset. We find that the effectiveness of a post to gain influence depends not just on its content but also on when it is published and in a subtle way on how the post relates to their users’ interests. As part of our approach, we also propose a simple and natural method to model and predict the effectiveness of a post in gaining social influence. The experimental results on the real-world dataset is consistent with our findings. Our work provides a new method to analyze the social media from data mining view, which contrasts with a number of theories from marketing and sociology.