Measuring viral spread of online memes

Fascinating article in Slate on the evolution of the "25 Random Things about you" meme:

Late last fall, a chain letter titled "16 Random Things About Me" began to chew its way through Facebook. The author of one of these notes would itemize her personality into "16 random things, facts, habits, or goals," then tag 16 friends who would be prompted to write their own lists. And so on and so on. Similar navel-gazing letters had popped up over the years through e-mail and on blogs, MySpace, Friendster, and the venerable blogging site LiveJournal. The Facebook strain had a good run, but by the end of 2008 it appeared to have stagnated.

Then something curious happened: It mutated. Since everyone who participates is supposed to paste the original instructions into her own note, it's easy to tinker with the rules. Soon enough, 16 things (and 16 tagged friends) morphed into 15—and 17 and 22 and 35 and even 100. As the structure crumbled, more users toyed with the boundaries. Like any disease, "Random Things" was mutating in hopes of finding a strain that uniquely suited its host. In this case, the right number was vital to its survival: The more people who are tagged, the more likely the note is to spread. The longer the list, though, the more daunting it is to compose and the fewer participants will be roped in.

By mid-to-late January, "25 Random Things About Me" had warded off its competitors. Once the letter settled on 25 things (a perfect square, just like 16) the phenomenon exploded.

The graph above shows the takeoff - and it is recognisable as a power law spread, as with any viral infection. It has a 1.27 factor (ie each person "infects" 1.27 more people on average until the overall population is infected). They also graph the time taken to write the follow on meme in the article, and note that:

The highest percentage of respondents—17 percent of those who wrote a note—composed their missive the same day they were first tagged. The numbers decay from there, and the median value is three days.

As Slate notes though, for people wishing to market virally:

....viral marketers might take note of the patterns that "25 Random Things About Me" obeyed. The best hope for someone looking to start a grass-roots craze is to introduce a wide variety of schemes into the wild and pray like hell that one of them evolves into a virulent meme. If evolution is any guide, however, there's no predicting what succeeds and what doesn't.

Nonetheless, it will be interesting to measure other viruses and see if there are some empirical laws that can be derived.