Scientists Claim To Have Developed Algorithm Which Predicts If A Photo Will Go Viral On Facebook

Scientists have apparently created an algorithm that is able to predict if a photo or an image will go ‘viral’ on Facebook.

Researchers at the University of Stanford have compiled a complex piece of algorithm that takes multiple cues about the image and predicts with a fair amount of certainty if that image will soon go viral or will simply be a small drop in a bucket. The researchers have been able to zero in on two essential parameters that decide if the photo will soon be viewed by many. They have been able to decipher the pattern observing just ‘how fast’ a photo is shared will eventually decide if the image has a high potency to last long on Facebook or be lost in the barge of other photos.

Stanford researchers revealed that clues to predicting if the image will go viral lie in ‘cascades’, reported Business Standard. ‘Cascades’ is used to describe photos or videos being shared multiple times. They agreed these patterns were highly elusive as it does not happen so often. Facebook’s in–house scientists were kind enough to offer data which was stripped of any personal identification tags. A total of 150,000 photos were put up for analysis, each of which had more than 5 shares in the least.

The data revealed that only 1 in 20 photos posted on Facebook gets shared even once. Just 1 in 4,000 photos get more than 500 shares. These are too few numbers and hence the scientists had to rely on multiple other parameters to extrapolate their hypothesis before arriving at a theory and then developing the algorithm, reported the Financial Times.

The algorithm developed for Facebook has worked surprisingly well. In a Research Paper titled ‘Can Cascades be Predicted?‘ that the researchers will soon publish, the algorithm has been reported to have worked fairly accurately regarding the viral prediction in 8 out of 10 times. In other words, the algorithm managed to successfully predict 80% of the time if the photo cascade on Facebook will double in its ‘shares’. Meaning, if a photo had 20 shares, the code rightly predicted if the same photo will get 40 shares & so on.

Fast and Speedy works: The old adage, ‘Slow and steady wins the race’ simply doesn’t apply to photo sharing on Facebook. In order to get a lot of shares and likes, the ‘cascades’ have to be very rapid. Simply analyzing how quickly a cascade unfolded predicted doublings, 78% of the time. The mystery of Social Sharing & the viral nature may be very complex, but researchers appear to be inching closer to unraveling the same.

[Image Credit | Science Daily]