In social media it is common to use or hear the expression “going viral” in the context of predicting the popularity of social content. “This video is definitely going viral.” “Can you make this post go viral?” If you work in digital, you’ve probably heard these expressions or requests many times before. But now it’s time to put a stop to it, and here’s why:

Equating the spread of social media posts (in this case, Tweets) to the phrase’s origins — the spread of biological disease — is fundamentally inaccurate when it comes to predicting content performance. We looked at predictive modeling for the spread of biological disease, and broke down a few reasons why predicting something in terms of virality in the context of digital doesn’t make sense.

First, here’s a quick overview of what basic factors are accounted for in these predictive modeling equations:

  1. Population size, age, and density. Viruses spread using different patterns based on the virus’ characteristics. Density interacts with infection rate to produce an accurate picture of how a disease might spread.
  2. Incubation rate. This is the amount of time it takes for a ‘victim’ to exhibit symptoms and become contagious.
  3. Rate of transmission and transmission type. This is the speed in which the disease spreads, which is affected by how it is spread (e.g. Airborne spreads faster than body fluids).

Now, here’s why none of that makes any sense when applied to social media:

Spreading and Sharing
While there are biological factors that impact the spread of a disease (geography, climate, etc), common viruses do not discriminate when spreading. On the other hand, the act of sharing a social post is a human decision. Factors such as incubation time and infection rate, used in measuring the spread of a biological disease, are simply useless in the analysis of social media since “spreading” a post requires concerted effort by the end-user. While the “spread” of a post can be mapped in retrospect, it’s nearly impossible to predict if any given user is going to share the post or not.

Uniformity of Population
In social, there is no uniformity of population. We’re all unique and beautiful social butterflies. The typical unit used when studying the spread of disease is the population. Within a population, it is reasonable to assume that each individual roughly has the same propensity of contracting and spreading a virus. Since social media users have broadly varying numbers of followers, the chance an individual will re-post a tweet that then goes viral is impossible to determine with accuracy.


Expected Reach
While you can estimate “expected reach” by looking at your followers, their followers, their followers’ followers, and so on, there is no way to insure that they will repost a tweet, thus hinging your model on the fickleness of humans.


So the next time someone claims their social content will go viral, or asks you to make a post go viral, you have our permission to tell them no, because it’s impossible to predict; maybe instead you tell them you will do your best to make their social content trend.