New algorithm can detect if photos will go viral

New algorithm can detect if photos will go viral

Scientists have developed an algorithm which can predict if your photo will go viral on Facebook based on how quickly it is shared.


When you share a picture on Facebook, you always want it to spread and get huge amount of likes and comments but even though you would want it go viral, fact is, you do not really expect much. However, now researchers have found a way to know which picture may or may not get famous and go viral.

Stanford University researchers are working with Facebook data scientists and will present their findings at the upcoming 23rd International World Wide Web Conference which is to be held in Korea. The team of researchers which includes Jure Leskovec, assistant professor of computer science at Stanford, Jon Kleinberg who is a computer scientist at Cornell University and  Facebook researchers Lada Adamic and P. Alex Dow. The researchers will describe in detail how they can accurately predict 8 out of 10 times if a photo shared multiple times would double in shares and how they arrived at this conclusion.

According to the data provided by Facebook and University researchers, only 1 in 20 photos uploaded on the social network gets more than 1 share and 1 out of 4000 pictures gets more than 500 users to share it which is not really a lot when you talk about going Viral.

“It is very hard to quantify what going viral means,” said Leskovec. “Anyone would say ‘Gangnam Style’ went viral, but that’s a singular event,” he said, referring to the YouTube video that has been viewed almost 2 billion times.

The researchers began their study by analyzing 150,000 photos uploaded on Facebook which got more than five shares. Initially it was found that a picture shared more than once had a 50-50 chance of getting double the amount of shares.

When various criteria’s were applied to the test they were able to predict with 80% percent accuracy if the photo would get double number of shares and photos which were shared over a 100 times had a accuracy of almost 88% percent.

Among the various influencing factors the speed of sharing and the wide range of social networks over which the photo was shared was the best predictor of a photo going viral.

“Even if you have the best cat picture ever, it could work for your network, but not for my boring academic friends,” Leskovec said, adding, “You have to understand your network.” and there was no one single trick of doing this.

Eric Brown
Connect on



  1. John D

    This is an interesting study and will benefit marketers in understanding what works for them and what doesn’t.

Leave your comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.