The Myth About Viral Marketing

The idea that messages frequently go “viral” and diffuse through social networks is now a given in corporate marketing and the culture. But recent research suggests that the term “viral” marketing does not accurately describe what most often happens online.

Truly viral diffusion is extremely rare, according to Sharad Goel, a senior researcher at Microsoft Research in New York.

In a paper called “The Structure of Online Diffusion Networks” that was presented at the 13th ACM Conference on Electronic Commerce, Goel, Duncan J. Watts and Daniel G. Goldstein (all of Yahoo! Research when they wrote the paper, now at Microsoft Research) describe how they studied seven online scenarios to see how a variety of applications and content spread. The researchers studied:

1. Yahoo! Voice, an online phone service started in 2004;
2. Zync, a Yahoo! Instant Messenger video-sharing application;
3. Friend Sense, a Facebook app introduced in 2009;
4. “The Secretary Game,” an online hiring game;
5. Yahoo! Kindness, a charitable website launched in 2010;
6. news stories sent via Twitter in November 2011; and
7. YouTube links diffused through Twitter in November 2011.

Goel said that he and his coauthors wanted to see whether adoption spread virally, “like the common cold” or some other sort of biological contagion. In such a model, “one person gets infected and then their friend gets infected and then a friend of their friend gets infected, and so on.” That process is called multistep diffusion. And the researchers’ data suggest it’s not how most messages and applications spread.

“What we see is something qualitatively different,” Goel said. Most content, he explained, is adopted and dies out within one generation. In fact, less than 4% of adoption “cascades” the researchers studied extended more than one person away from the initial adopter.

For marketers, this research suggests that it may be time to abandon the idea that viral marketing via social media will frequently lead to, say, tenfold organic growth. What, then, accounts for those rare pieces of content — such as YouTube videos viewed by millions — that become extremely popular on the Web in a relatively short time frame? According to Goel, he and three other researchers, in yet-to-be-published research, examine more than a billion news stories, videos, images and petitions on Twitter and find that even many popular pieces of content are not truly viral in the sense suggested by infectious-disease models. About one in a million times, however, they find that content on Twitter does in fact “go viral.” Understanding what sets these rare viral events apart is their latest goal.

— Michael Fitzgerald