It was possible to draw all sorts of conclusions from data about “HuffPo,” both flattering and not. Yet two numbers in particular stood out for what they suggested about the nature of the enterprise that Huffington, Jonah Peretti, and the host of that first lunch, Ken Lerer, had built: 40 million and 19,000. The former is the number of unique visitors who came to the site in January. The latter is the number of names in Arianna Huffington’s contact list.

Each represents a network that together constituted something far greater than what each represented by itself—an aspiration: the potential power that comes to those who can build, nurture, and harness a network that is at once vast and loyal.

Emphasis on potential.

1. Connected

Duncan Watts, an Australian who had come to Cornell for graduate school, was 32 and possessed a disposition that could be mistaken for curmudgeonly. The rapid growth of the Web had proven a boom time for social scientists, who could suddenly perform all sorts of research on large samples in very quick time and at relatively little cost. The result was papers and conferences and books that, in Watts’ view, transformed sociologists into engineers. Time and again, people approached Watts with questions for which he could offer only the maddening answer of “it depends.” This did nothing to make sociologists popular. Nor did it stop all sorts of people from coming around, among them Ken Lerer.

Six Degrees examined the nature of what Watts called “small worlds.” The conceit—later to be adopted by fans of Kevin Bacon and playwright John Guare—had come from a 1967 experiment by the social psychologist Stanley Milgram, in which he tracked the number of connections it would take for a letter to reach a certain recipient unknown to the original sender. The answer: six. That conclusion suggested that it was possible for any one person to reach any other person in the world by establishing a network of diminishing familiarity—start with a friend, then a friend of a friend, and so on—until the connection was complete. That premise had fueled the study of what Watts called the “science of networks,” one that dated to the 18th century and which had over time drawn in a host of scientists who seemed to think all things could be measured. And if they could be measured, they might be graphed and charted in a way that revealed patterns that could, in an ideal world, be replicated.

Perhaps the dream of creating vast networks of connected strangers was possible, if only one could identify the proper links. Or, in the parlance of the network world, making weak links into strong ones.

Watts’s book was filled with images and drawings that could be confounding—how many threads are necessary for connecting a mass of buttons in a way that created a button network?—yet tantalizing, especially, it seemed, for someone trying to find a way to take on one of the nation’s most powerful lobbies.

Still, networks were eternally undermined by the inevitable force of randomness. It was one thing, say, to go to a baseball game and hear the stirrings of rhythmic clapping that then cascade around the ballpark so that quickly everyone is clapping in unison. A powerful thing to behold—so much so that an inning later, you yourself might want to start the whole stadium clapping. Maybe the person to your left joins in, and maybe five or six others do, too. Until the clapping dies. In Watts’s view, networks were a wonderful phenomenon to observe, but all but impossible to replicate. Why did everyone in the ballpark feel the desire to join in the clapping in the sixth inning but not in the seventh? What was different? Could you somehow recreate the precise conditions that made that ephemeral but resoundingly successful sixth-inning network happen?

Watts doubted it. There were simply too many variables at work. Still, you could, in theory, try something: Start to clap, see if anyone joins in, stop if they don’t, wait for a new set of conditions to arise—another player to bat, a runner reaches second base? Or third? In other words, experiment, and measure the results as they occur, all the while adjusting, tweaking—try clapping louder, say, then faster, maybe adding a chant—but do so having accepted the likelihood that animates work for all scientists: failure.

Ken Lerer listened, and he was not deterred. Networks did, in fact, occur—vast networks through which previously disconnected people suddenly found themselves joined together, perhaps to share an idea, a song, a sentiment, a cause. Why not then try to create a network that could challenge the vast and powerful and sustaining network of the NRA?

Michael Shapiro is a contributing editor to CJR and teaches at Columbia's Graduate School of Journalism. His most recent book is Bottom of the Ninth: Branch Rickey, Casey Stengel, and the Daring Scheme to Save Baseball From Itself.