Some Causal Processes of Video Virality

By Ema Rivas Leal

In Politics in Time, Pierson introduces a few mechanisms to think about the longevity of a causal process and its outcome. With this example of video virality, I hope to illustrate some of these mechanisms.

We will consider a video as viral if it hits around 5 million views within a few weeks. Taking video virality as the outcome, then, we should begin with the cause, the posting itself. Imagine a normal person posts a video. Hopefully, someone sees it, likes it and shares it. Resulting from this new share, someone else decides to share it, again increasing the potential viewers and sharers, and so on. In this way, we could think of a causal chain where the posting of the video (x) directly yields video virality (y) provided that x effectively initiates a sharing sequence. Since the odds are probably against the initial rounds of sharing, x and y may be temporally separated, making the causal chain difficult to establish. This implies that x initiating the sequence is not enough to guarantee outcome y.

Yet, if we consider the concept of threshold effects, the outcome could be ‘guaranteed’ sometimes. If we think that once a video has been shared a certain amount of times, the possibility of it being further shared a few million times in a short period of time becomes substantially heightened, then the process would involve a threshold. Furthermore, if it could be established that there was a critical moment when some famous person decided to share the post, immediately propelling the video close to or over the threshold, the process becomes analogous to Gladwell’s description of the tipping point. At the point where the famous person comes in, we could even identify some structural determination. Depending on this person’s level of popularity, their share could immediately make the sequence surpass the threshold, guaranteeing video virality. Finally, throughout the entire process positive feedback can be discerned, provided that sharing generates further sharing.