predicting popularity

predicting popularity

Whether it’s a rush of customers at a cafe, a change in fuel prices or the spot where lightning decides to strike, life can seem a bit random at times. As someone who writes online articles, what captures an audience’s attention can be equally perplexing.

If the Monty Python motley are correct, order may be found in the bladder of a sheep. Unhappy with the prospect of hurting a sheep, and with a scarcity of sheep nearby, an investigation to find someone with an answer and a method to predict the future was necessary.

Carlos Castillo and his colleagues uncovered key online indicators recently that provide a window into the future of an online news article and a method for predicting popularity. With these predictive variables on their side, the researchers were able to anticipate the lifespan of news articles from the first 10 minutes after the news is released.

The team monitored interactions with news articles such as page visits, uptake and affects on social media and reader referrals immediately following an article’s release. Using data from the Al Jazeera English television network, and it’s companion Al Jazeera English website, a hypothesised relationship between an article’s shelf-life and it’s early presence in social media was examined. The Al Jazeera network use social media to broadcast and promote latest publications. Previous research found that the news provider enjoys stronger audience engagement and broadcast referrals than any other global media organisation of its kind.

Previous studies observed distinct patterns of user interactions associated with different types of online news articles. Accordingly, Castillo et al distinguish breaking news from in-depth news articles in their examination of online diffusion. During the study, the team additionally identified three classes of short-term response to new involving social media they describe as decreasing, non-decreasing and rebounding patterns of interaction.

While in-depth news articles exhibited complex and highly variable patterns of short-term audience responses, response patterns to breaking news articles were more standardised. Roughly 80% of breaking news engagement consistently declined over time, around 10% gradually increased in the 12 hours following release online and 10% showed a gradual increase in audience interactions.

The findings of this study reveal that social media is not merely limited to providing an additional platform for media broadcasts, but offers a functional online environment generating meaningful feedback on the trends of core media business activities, such as a TV network’s main website operation.

Thanks to Matt Stempeck for his assistance.


 C. Castillo, M. El-Haddad, J. Pfeffer, M. Stempeck: “Characterizing the Life Cycle of Online News Stories Using Social Media Reactions”. Unpublished manuscript. 2013.


Online Advocate/Australian Higher Education/Community Health/Youth Mental Health. Follow me on Twitter @writerinsight

1 Comment

  1. Pingback: Twitter To Squawk: Social Media Political Content Migrates Offline | Student View

Leave a Reply

%d bloggers like this: