by Greg Satell
FIGURE 2.5 The Beta Model
* * *
This is exactly what Watts discovered mathematically. The more he looked, the more it became clear that it takes very few random links in the network to shorten path lengths considerably. In fact, adding only five random links, he found, would cut social distance in half, regardless of the size of the network. Not only was it possible for tight clusters to combine with short social distance; networks actually have a strong tendency to become tight clusters combined with short social distance. In effect, Watts’s mathematical model predicted that small-world networks are not a special case, but a natural state.
These random links played exactly the same role that Granovetter’s “weak ties” did in his job-hunting studies. In one fell swoop, Watts had found an explanation that agreed both 66with Erdős, Rényi, and Rapaport’s mathematical investigations and with the real-world studies of Milgram and Granovetter. Networks really are “small worlds,” clustered into tight communities, but also connected through links over long distances. Small groups, loosely connected, have a strange power to synchronize.
When Watts and Strogatz applied their model to actual data—the network of movie actors in the Internet Movie Database (IMDB), the neural network of the worm C. Elegans, and the power grid of the western United States—their theory fit precisely with the behavior of real-world networks. The paper Watts wrote with Strogatz, Collective Dynamics of ‘Small-World’ Networks, would prove to be a landmark in the study of how things connect and remains one of the most highly cited papers in the field today, 20 years later.22 High clustering going hand-in-hand with massive reach, it seemed to be a universal law. Loosely connected small groups can drive massive synchronized behavior, as long as some element of randomness is introduced in the system (and in the real world, randomness almost always comes into play).
THE MAKING OF A CASCADE
* * *
The unlikely success of his network studies led Watts to begin studying another network-related phenomenon known as cascades. A particularly vivid example of a cascade occurred during the summer of 1996.23
On August 10 of that year, an exceedingly hot day, air conditioners across the West Coast were going at full blast when a transmission line near Portland, Oregon, failed. This was not an unusual occurrence, and redundancies were built into the system to cover for just such a failure. The load of electricity was supposed to be rerouted by a relay to another substation. However, with the entire system operating at its limit, the surge caused another line to fail. The excess power from that backup line was rerouted to yet another overly taxed component of the system, causing yet another failure. Soon, power surged throughout the system, causing cascading failures wherever it went. Within 70 seconds, the entire power grid between California and Oregon was out. Then the surge moved east, creating outages in six other states.
The power company’s report on the blackout listed many possible causes: maintenance problems, failure to address warning signs, and bad luck. Yet a general comment slipped in at the end of the report caught Watts’s attention—the power company admitted that it simply didn’t fully understand the interdependencies of the system.
Although it may not be immediately obvious, the disruptions we’ve discussed up to this point are very much like the West Coast blackout of 1996. They are also cascades. A disturbance in one part of the system eventually ripples through every other part of the system, finding vulnerable clusters as it travels. Eventually, the entire order is disrupted, much like the cascading failures in the electrical grid.
The electrical grid was also a network, and Watts, fresh off identifying the structure of small-world networks, moved quickly to study it. What he found was that while cascades were known to occur in a variety of environments, most of the examination focused on the nodes, not the networks. When a flu epidemic spreads through a school, or financial contagion spreads from a crisis in one country to entire continents, we tend to look for a cause in the initial failure rather than in the system as a whole. Social fads work in a similar fashion. Whether the craze is for Harry Potter or the iPhone, we tend to think there’s something magical about the idea itself and often fail to notice the network the idea travels through.
Yet in our rush to ascribe significance to a particular node, we often miss the dynamics of the network. Certainly, there was nothing particularly special about the transmission line that started the blackout of 1996. So why do we tend to assume that there’s something special about the idea of Harry Potter or the iPhone? Sure, they are powerful ideas, but so are many others that die on the vine. What’s the difference between those and many fairly ordinary ideas, like LOLCats and other Internet memes, which somehow get uplifted by extraordinary cascades?
Just as important, Watts argued, is the structure of the network, which can influence a cascade’s probability.24 If you think about it, that’s undeniably true. Would the financial crises of 1997 and 1998 have happened if the emerging markets of Asia and Eastern Europe were stable to start with? Was the amazing success of Harry Potter due only to the talents of J.K. Rowling, completely removed from context of time and place? Would the iPhone have been as successful if it had not been created by Apple, with its dense network of brand loyalists?
That’s a crucial point, because the structure of networks is something that we can do quite a bit about and that can have a dramatic effect on whether a movement cascades or not. We can choose to depend on a few all-stars, or we can disperse responsibility around the organization. We can insulate ourselves, like Milošević in Serbia, Yanukovych in Ukraine, and the various dictatorships in the Arab Spring, or we can choose to build open, transparent management structures, like General Stanley McChrystal did in Iraq. We can lock up our organizations in airtight, siloed fortresses, or we can encourage communication across the enterprise.
We tend to see success and failure through the prism of strategy and tactics, and that is, to a certain extent, true. But we also act according to how we see the world. If we see the world as a chessboard, with various powers vying for dominance, we will act one way. However, if we see the world as a series of networks and ecosystems, we are likely to act differently. What we see determines how we will act.
SEEING THINGS FROM A NETWORK VIEW
* * *
In the Introduction to this book, we looked at three pairs of examples: Occupy and Otpor, Blockbuster and McChrystal’s forces in Iraq, and Route 128 outside Boston and Silicon Valley. Look at these cases one by one, and you will see differences in strategy and tactics. While Antioco changed his business strategy, McChrystal transformed his culture. The Route 128 companies sought proprietary competitive advantage, while Silicon Valley companies created an open industrial ecosystem. Occupy chose angry rhetoric, while Otpor chose street pranks.
Yet when taken collectively, a different picture emerges. Antioco, the Route 128 firms, and Occupy saw hierarchies and framed the challenges they faced in terms of traditional resources, such as strategic capacity, competitive pressures, and market dynamics. McChrystal, the Silicon Valley firms, and Otpor, on the other hand, saw networks to integrate, and that made all the difference. They looked to build connections to other nodes in the system and recognized that true power resided in those linkages.
Otpor set out to identify pillars supporting the existing power structure—not to knock them out, but to draw them in. So for them, every arrest was a chance to win converts at the police station, and those converts, in the end, were what proved to be decisive. Occupy, on the other hand, eschewed all connections to the existing order and quickly devolved. McChrystal saw that, “to defeat a network, you must become a network” and transformed his organization from one that relied solely on a top-down command structure to one that worked to build horizontal connections that empowered a “team of teams.” Antioco and his executive team at Blockbuster merely saw a strategy to be executed and neglected internal networks. Silicon Valley saw an ecosystem that needed to be constantly nourish
ed and renewed, while the Route 128 firms believed in vertical integration.
In The End of Power, Moisés Naím pointed out that “power is easier to get but harder to use or keep.” That is undeniably true, but I also think it misses the point somewhat. The greater truth is that in a world connected by digital technology, power no longer lies at the top of hierarchies, but at the center of networks. We move to the center by making new connections and drawing others in.
Networks have important advantages over hierarchies. Hierarchies are expensive and difficult to maintain. Leadership must provide standards, governance, oversight, and resources. The Tea Party, to take just one example, was able to take the country by storm, seemingly overnight, because it didn’t have to provide any of those things. It only needed to empower the groups that formed of their own accord, let them connect with each other, share ideas and practices, and, most of all, help them to understand that they were not alone, but there were many others who shared their ethos.
This kind of distributed authority can be amazingly effective. Consider the Orpheus Chamber Orchestra, one of the most accomplished ensembles in the world. It operates without a conductor. By giving everybody a voice in which arrangements will be performed, how they will be interpreted, and what role each instrument will play in a particular piece of music, it has unleashed the creative energy of its members and won multiple Grammy awards in the process.25 In a similar vein, Morningstar, a $700 million dollar food processing business, operates without any management hierarchy at all.26
Networks, unlike hierarchies, push authority down to the lowest level of an organization. They recognize that people on the ground have the most current information and are most able to act on it quickly and effectively. However, there is a catch. Just as well-functioning networks can cascade toward a common purpose, poorly governed ones can spin wildly out of control. One or two rock throwers can quickly transform a peaceful demonstration into a horrifying riot. Factions can form in an organization and go off in competing directions. Much like Tolstoy wrote about families, all well-functioning networks are alike, but each can go wrong in a multitude of ways.
Every movement for change has a trigger. In Serbia and in Ukraine’s Orange Revolution, an election triggered the mass protests that brought down the respective regimes. Later, during Ukraine’s Euromaidan protests, it was the pullout from an EU agreement that set things off. With Occupy, it was the financial crisis, and in Silicon Valley it was the shift from microcomputers to PCs that set the stage for Bay Area dominance and sealed the fate of Route 128. More recently, videos of police violence have led to massive protests and the Black Lives Matter movement, while the mass shooting at Marjory Stoneman Douglas High School in Parkland, Florida, ignited the March for Our Lives movement.
Yet notice that not all of these movements have been successful. A cascade makes change possible, but it doesn’t make it inevitable. As impressive as the mass protests that Occupy was able to mobilize were, they never amounted to much. They made a point, but they never made a difference.
And that’s what’s crucial to understand about cascades. They are a means to an end, not an end in themselves. Mobilizing thousands of people to walk the streets or sleep in parks or post messages on social media do not in themselves make change happen. In fact, such actions can—and often do—backfire if they motivate those who oppose change to mobilize themselves. Every revolution sets the stage for a counterrevolution. Cascades are only useful if they result in influence, and to do that, they must travel far beyond where they start.
That is what we turn to in the next chapter, in which we will see how small groups, loosely connected, but united by a common purpose, lead to transformational change and how harnessing these forces often determines whether a movement succeeds or fails.
CHAPTER 3
How Cascades Create Transformational Change
What all information cascades have in common, however, is that once one commences, it becomes largely self-perpetuating; that is, it picks up new adherents largely on the strength of having attracted previous ones. Hence, an initial shock can propagate throughout a very large system, even when the shock itself is small.1
—DUNCAN WATTS
My best friend growing up was a boy named Robbie, and he had a special gift. He was the kind of guy everyone loved to be around. Although he was not a great student, teachers had a soft spot for him. He was a good athlete, but not a star. Coaches loved him anyway. He was not what you’d call suave or debonair. Nevertheless, girls liked him. Robbie had something. He was funny, had a zest for life, and loved being around other people. He had a certain way of saying and doing things. He would tell a joke in the locker room, and the next thing you knew, everybody was repeating it.
I think everyone knew someone in high school like Robbie. So, when Malcolm Gladwell published The Tipping Point, which offered an explanation why people like Robbie can be so infectious, it became an instant bestseller, in part because he claimed people like Robbie can lead to much bigger things. Many great movements have charismatic leaders. We see Mahatma Gandhi, Martin Luther King Jr., and other legendary figures leading the charge, inspiring large masses of people to act in the service of a cause. In business, we see larger-than-life figures like Steve Jobs and Elon Musk at the head of their own revolutions.
So the idea that the personal qualities of rare individuals determine whether a change effort succeeds or not makes intuitive sense, although as we will see, things don’t really work that way. It is, in many ways, an urban myth, and if we try to create change based on that assumption we are unlikely to succeed. So it’s important to understand why it doesn’t hold water. It is not “influential” people who create change, but as Duncan Watts put it to me “easily influenced people influencing other easily influenced people”2 that creates a cascade.
Gladwell’s explanation for how personal qualities drive change is based on what he calls the “Law of the Few.”3 “The success of any kind of social epidemic is heavily dependent on the involvement of people with a particular and rare set of social gifts,” he writes and then goes on to classify these gifts according to three classes of highly influential people: Connectors, Mavens, and Salesmen. Connectors know a lot of people, but they are more than just social butterflies. They actively seek out folks to introduce and, by doing so, attract even more people to them. The Mavens influence others through their deep knowledge and expertise. They are subject matter experts that others actively seek out for advice. Salesmen seem to have supernatural powers of persuasion.
We’ve all met people like this and we remember them. Connectors, Mavens, and Salesmen have played important roles in our lives because they have, as Gladwell notes, skills that are rare and valuable. So the idea that people like these play an outsized role in creating change just feels intuitively right. It is, in many ways, a more modern and trendy version of the “Great Man” theory of history. If true, the question of how to create transformational change would be relatively easy. You simply have to find a transformational changemaker (an “influential” in popular modern parlance), and this person could then convince everyone else to join in.
Unfortunately, the Law of the Few, while seemingly plausible, is dangerously misleading. As we will soon see, it is not “special people” who create change, although some with great talent can help to inspire it, but small groups, loosely connected, and united by a common purpose. Each of those three elements is crucial, because small groups engender strong bonds, loose connections provide greater numbers, and a common purpose gives direction. Every change effort, if it is to succeed and provide lasting results, needs each of those elements because it requires long chains of sustained influence to make a real difference.
However, in order to understand how small groups, loosely connected, but united by a common purpose, give rise to cascades and lead to transformational change, we first must understand why the Law of the Few and the idea of influentials isn’t a viable explanation for what makes a mo
vement succeed. To do that, we first need to understand where Gladwell’s idea came from.
OPINION LEADERS AND THE TWO STEP FLOW COMMUNICATION MODEL
* * *
Long before Gladwell published The Tipping Point, Paul Lazarsfeld and Elihu Katz, two eminent sociologists, developed what they called the Two Step Flow Communication Model. It was based on Lazarsfeld’s earlier research into the 1944 presidential campaign,4 which had inspired subsequent studies throughout the 1940s and 1950s.5
The basic argument was that mass media doesn’t actually affect the masses, but works through “opinion leaders” who amplify messages. The existence of these special people was predicated on three findings from their research of presidential politics. First, they found that people who had changed their minds or selected their candidate late in the campaign were likely to cite the personal influence of those they knew as a factor in their choice. Second, they found that the “opinion leaders” who influenced the choice of others tended to be of the same social class as the people whom they influenced. So, and this is a crucial point, the opinion leaders weren’t influential by virtue of celebrity or position, but through a personal relationship. Third, the opinion leaders were exposed to mass media more than others.
