Does Twitter Have What It Takes To Predict A Viral Tweet In Real Time?

Being able to produce a “viral sensation” on social media may seem like magic, but Twitter hopes to uncover the secret sauce behind the world’s more viral tweets. The question is, can you really boil down “going viral” to an algorithm?

While it’s still in stealth mode (i.e. protected), the newest Twitter experiment appears to be aimed squarely at figuring out what makes a tweet soar above its peers in terms of popularity.

TechCrunch has been tracking the activity of the Twitter account @MagicStats for a few weeks, and noticed that the account seems to be favoriting certain tweets. Despite the fact that the account is now protected and its tweets unreadable, the TechCrunch team was able to see all of its tweets and favorites before it went dark.

The account seemed to pop a favorite onto a tweet very quickly after it was sent – but never acting first. This might indicate that it was running off an algorithm attempting to detect the tweets that were rising quickest to popularity.

@MagicStats was originally related to yet another experimental account from Twitter, @MagicRecs. This account offers up recommendations to users via DM (to get yours, just follow the account).

These and other Twitter experiments indicate a bigger trend going on on everyone’s favorite 140-character network – the need to understand what’s popular, before it becomes popular.

Capitalizing on trends has always been a drool-worthy proposition for media companies (which, one could argue, is what Twitter is), but what if, instead of reacting to trends, a company could actually predict them?

The @MagicStats account appears to be one of many experiments aimed at doing just this. It’s goal is to ferret out the top tweets on Twitter as they’re gaining in popularity rather than after they’ve reached it (which is what Twitter’s Trending Topics shows us).

Predicting virality, however, is not as easy as looking at how many retweets a tweet is getting and assuming it will “go viral”. Just think about Gangnam Style… who could have ever predicted how far and wide this infectious pop phenomenon would travel? And plenty of marketing firms have spent top dollar on YouTube videos with all the elements of “virality” – like cute cats, memes and hipster-cool spokespeople – only to have them flop.

Plus, there are different degrees of online virality. Some videos, photos and other content seem to bubble to the collective surface of the internet – again, think Gangnam. I bet even your grandma would recognize the tune. But other forms of virality exist, and might be even more important. What’s hot among my peers or colleagues would probably hold my attention longer than a generic trend, and I’d consider it more valuable to be able to tap into what’s “going viral” in my own network.

Twitter is clearly focusing on personalized experiences with the @MagicRecs account, so it’s not a stretch to assume they’ve got something in the works to identify virality on a more personal level.

If they can successfully identify upcoming trends, in real time, and go the extra mile to personalize them, they’ll be way ahead of the competition – and they might just create the value proposition that hooks new users and entices them to stay.

(Magic crystal ball image via Shutterstock)