Anatomy of Viral Social Media Events

Abstract

Discussion topics go sometimes viral in social media without a seemingly coherent pattern. Existing literature shows these discussions can reach a very high level, but, notably, they prevail to varying degrees. This paper investigates the anatomy of viral social media events using a dataset of 960 viral social media discussion topics that have been identified by an algorithm from a variety of social media sources over two years’ time. A negative binomial regression shows that the average daily amount and the relative change in the daily amount of social media platforms at which the event has been discussed has a positive effect on the duration of the event. Average or relative amount of posts or authors has no or very little effect on event duration. The results suggest that viral social media events last longer when people using different social media platforms get exposed to them. This finding contributes to the literature on social media events, virality, and information diffusion.