Previous Next

How Does Social Content Spread?

March 31, 2014

Why is some social content shared more than others?

A team led by Justin Cheng at Stanford with others from Facebook and Cornell took on the question from a new angle. They studied 150k photos on Facebook and instead of examining how the photos spread from the onset, which other studies have done, they started by looking at an already re-shared photo and create an algorithm that tried to predict the likelihood that it would double its shares.

Their task was to predict whether the sharing cascade for a photo will double in size and to do this they looked at a number of factors; the type of image, the number of followers the original poster had, the velocity of the cascade and its shape.

Here’s what they found:

  1. Something that starts off by spreading fast is likely to continue to spread
  2. Captions matter- being newsworthy, or attached to a current meme helps
  3. The more reshares- the easier it is to predict the outcome
  4. Timing and alignment matters- the right content needs to be placed in the right network at the right time

Although the study focused only on photographs, the findings of Chen and his team seem to echo the way new sites especially Upworthy and Buzzfeed are operating. Both are focused on optimizing their headlines in order to maximize their appeal and Buzzfeed examines every article for an hour to see how fast it’s spreading and use that data to predict the likely future success of a piece.

Some lessons here for agencies and social content.

  1. If it’s not spreading fast from the get go -it’s not going to become a hit
  2. The right headline is critical- it might not turn a bad piece of content into a good piece, but it will make sure something good is more likely to be seen
  3. Connected to 2- you need some sense of “zeitgeist” you need to know/feel the overall conversation, either through data or instinct, or the combination of the two

We are still at the very early days of understanding the science of social sharing and the more these massive datasets are analyzed, the smarter we will become.

Related Articles

The social network evolves and fragments- 3 new trends
To date, social networks been defined by the...
The challenges of scale- facebook
Big is good. It shows scale, popularity and for...
Social media is the new marketing job
From the FT..At Dell:"Bob Pearson, former senior...
Enemies- a social idea for a divisive age
While the country seems increasingly polarized...
Screen Shot 2013-08-12 at 2.19.07 PM
Contagion:The Spread of a Viral
There’s something cool about seeing how a...