Facebook studies the evolution of memes using gene technology


By S.H. Andras

table of variants

Ever notice how memes take on a life of their own? Memes are similar to the game of “Telephone,” as the original phrase often changed considerably as it is whispered from person-to-person. Sometimes a person in the chain changes the phrase for comic effect, sometimes not. The same can be said for Internet memes.

Facebook decided to model memes as genes to study their evolution as they travel through the social networking site. Until now, it had been impossible to trace the evolution of memes because of lack of data. Facebook was able to study more than 460 million individual meme instances. They found that as memes were changed slightly, depending on need or niche, they sometimes became more popular. For example, adding the phrase “please post this” to the meme made it two times more likely to be shared.

In September 2009, 470,000 Facebook users posted the following as their status update: “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.”

According to Facebook’s blog post:

At some point someone created a variant by prepending “thinks that” (which would follow the individual’s name, e.g., “Sam thinks that no one…”), which was copied 60,000 times. The third most popular variant inserted “We are only as strong as the weakest among us” in the middle. “The rest of the day” at one point (probably in the late evening hours) became “the next 24 hours”. Others abbreviated it to “24 hrs”, or extended it to “the rest of the week.”

For this particular post, 121,605 different variants of this meme appeared in 1.14 million status updates. Facebook decided to to connect the variants similar to the way geneticists trace the lineage of a particular genetic sequence using a gene tree and then used a simpler process called Yule.

Tracing genes is much more difficult because, to study them, one must go back generations to examine mutations. With a meme, Facebook easily started at the origin of a post and then traced when copies and mutations occurred, which are the two basic ingredients in the evolutionary process of a gene. Using the Yule process, they showed that every instance of a meme variant has a constant probability over time:

  • How popularity will be distributed among variants;
  • Characterized popularity distribution: If it lands close to 0, all variants are equally popular, close to 1 and a few variants are extremely popular; and
  • If most of the meme copies are made among friends, lineage can be constructed very easily.


The parallels between the evolution of memes and genes will continue to be traced as posts propagate and evolve through Facebook and beyond. While this study only traced memes on Facebook, it revealed how ideas transmit themselves through social ties in general.