At the Harmony Institute (HI), our experience measuring the influence of media and entertainment has taught us that the cultural environment matters. Outside factors like public opinion and framing affect the impact of a film or social outreach campaign.

The case of the award-winning documentary Bully shows how important it is for filmmakers to adapt to external conditions and challenges. In the weeks leading up to its theatrical release, Bully faced a major hurdle when it was issued an R rating by the Motion Picture Association of America (MPAA). The rating threatened to limit the film’s availability to younger students both in theaters and in schools. Engaging this target audience was a key aspect of filmmaker Lee Hirsch’s vision of social change.

After learning about the MPAA’s decision, a 17-year-old high school student from Michigan took action. Katy Butler started a petition on the social action platform Change.org to change Bully‘s rating to PG-13. Butler’s eventual victory was decisive. The petition gathered more than 500,000 signatures, and within six weeks, the MPAA changed their decision, granting the film a PG-13 rating. The challenge is in understanding how this battle was won.

The graph below plots signatures to the Change.org petition over time. What makes this data so interesting is how the slope of the graph evolves in stages. Very early on, the line shoots upward, displaying exponential growth (the green line segment). In a single day the petition amassed more than 100,000 signatures. Over time, the rate of growth slows, following a more sustained, linear shape (the blue line segment). During this month-long period the petition reached its target of more than 500,000 signatures. The two observed stages of growth suggest two related models of influence: an early period of rapid adoption that we are calling a “viral stage” and a longer-term period of sustained growth.

While the viral stage was key in driving early visibility, the period of slower growth provided far more signatures for the petition in total—around 400,000 out of 523,475. As researchers, we were fascinated by the unusual shape of this graph and by the astounding success of the campaign. What made the petition go viral? What sustained its momentum?

Researchers Sharad Goel, Duncan Watts, and Dan Goldstein ask these types of questions at Microsoft Research. Working at the intersection of computer science and sociology, they analyze the spread of information online using huge data sets. One of their latest papers, “The Structure of Online Diffusion Networks,” helped us interpret our observations of the petition’s success.

First, Goel, Watts, and Goldstein quantify the rarity of large-scale viral events. Since the public tends to discuss only the most wildly successful viral campaigns, it is easy to ignore that the vast majority of information shared online doesn’t travel very far. They found that on Twitter, less than one percent of stories satisfy even a limited definition of viral sharing.

Given the long odds of success, it was key that Bully’s outreach team acted on this rare opportunity to support a viral event. Recognizing the early success of the petition, the Bully  team allied with Butler to promote the issue on a wide variety of outlets, from the Huffington Post, to CNN, to the Ellen DeGeneres Show. The timing of these appearances coincided with the slower, but steadier increase in signatures. For a petition widely described as “viral,” more traditional mass media coverage played a crucial role in its sustained success.

The empirical work of Goel, Watts, and Goldstein helps explain this sustained, less viral growth. Discussing some drawbacks of the viral metaphor, they explain: “One possibility…is that events such as these are not strictly viral…but rather obtain the bulk of their attention either from traditional advertising or from other coverage by the mass media.”

This important idea points to the relationship between social networks and mass media. We often think that “viral” events like Katy Butler’s petition reside exclusively on the web and that mass media like television are passive and dated. The reality is much more complex and interconnected.

One possibility … is that events such as these are not strictly viral … but rather obtain the bulk of their attention either from traditional advertising or from other coverage by the mass media.

The interaction between Bully‘s filmmakers and Butler’s grassroots petition models a powerful feedback loop with important implications for media producers interested in social change. First, the MPAA petition drew attention during its spurt of viral growth. Although the “viral stage” was relatively brief, it rallied a committed base of supporters. Picking up the viral story, more traditional mass media outlets exposed the petition and film to a much larger audience. Rather than passively receiving this information, viewers returned to the interactive domain to sign and share the petition, adding even more signatures. This positive feedback mechanism (Goel, Watts, and Goldstein call it “broadcast diffusion”) tends to spark the largest online impact. Viral sharing is only part of the larger equation.

For most media creators, producing viral media sounds like an attractive proposition. Bypassing the gatekeepers of traditional mass media, it seems possible to make a big impact with nothing more than a short video. Of course, filmmakers shouldn’t ignore these rare viral moments, but they should also look to capitalize on support from other channels, increasing the chances for larger audiences and impact. Bully‘s ability to connect with its supporters and nurture this movement ultimately turned a disappointing rating into a rallying point for supporters.

These insights are the first in a more detailed analysis of Bully’s use of social media. Check back in the coming weeks for more results from our independent study.

Image: Michael Dwyer/ The Weinstein Co.