How Memes Spread On Facebook

It takes humor, good grammar, and straight-up pleading to make content go viral.

Consider the lowly chain letter, the original viral content. More than a century before the rise of doge, a Methodist women’s academy in Chicago sent out a letter asking for donations, along with a request for each recipient to make three copies and continue the chain. A fascinating new study posted by Facebook’s data science team suggests those women may have been harnessing a behavioral pattern as old as life itself.

In a look at hundreds of “Please post this” status updates, researchers found that good grammar, a call to action, and users’ political affiliation affected whether a meme would go viral or die a quick death. Each mutation that survived created its own ripple of reposts, and inspired additional variations down the road.

To show what this looks like, the Facebook team visualized the life of one 2009 status update supporting health care. The result is an evolutionary tree map, the same kind that biologists use to show genealogy, where every bloom represents a unique phrasing of the original status. “It’s long been suspected that ideas or memes may be evolving like genes,” Lada Adamic, a data scientist who worked on the paper, says in a phone interview with Fast Company. “But there wasn’t data to show that this was the case.”

When Richard Dawkins coined the term “meme” in his 1976 book The Selfish Gene, he had biological analogies in mind. Now, thanks to the team's research, Facebook may have proven he was on to something.

In 2009, the world’s largest social network removed its 160-character limit on status updates and ushered in a brave new world of communication. One particular species of status poetry thrived, sneaking into newsfeeds, complete with replication instructions. You know the type: “Please copy and paste this to your status,” they began, “if you love firefighters/hate cancer/want to help find a missing three-year-old.”

Take this status, copied verbatim by more than 470,000 Facebook users in September of 2009: “No one should die because they cannot afford health care, and no one should go broke because they get sick. If you agree, post this as your status for the rest of the day.”

Other users modified or messed up the original message as they manually copied and pasted (Facebook had yet to introduce a share button). Like the world’s biggest game of telephone, “the rest of the day,” eventually became “the next 24 hours,” or “the rest of the week.” At some point, a user inserted, “We are only as strong as the weakest among us,” after the first sentence.

“What you tend to see with a lot of these memes is that they’ll grow exponentially, which is something you see a lot in biology,” says Lada Adamic, a data scientist at Facebook who worked on the paper.

The paper revealed some other intriguing details. Facebook users were more than twice as likely to repost iterations with direct calls to action, such as “please post this” and “copy and paste,” than they were those that didn’t. Grammar hounds will be pleased to know that memes with typos, on the other hand, tended to die out.

While Facebook made it clear that the data was kept anonymous, a study like this could lead produce valuable insights into the qualities of shareable stories and ad content. “Understanding how memes diffuse can influence a story’s ranking in the News Feed,” Adamic says. Knowing why users share stories is crucial to keeping a network’s circulation moving, and Facebook could work that information into a variety of its products.

Of course, not everyone took the bait. Facebook’s data team found a number of parody memes, each with its own popularity among certain political demographics. Conservatives were more likely to post variations about beer and taxes, liberals preferred spoofs involving Jabba the Hutt or zombies. “Even if you’re somehow not susceptible to passing the original meme on," Adamis says, "there’s usually some variant that still gets you.”

[Image: Secrets via Shutterstock]

Add New Comment

3 Comments