The Hype Machine

How did I stumble into this line of work? In the fall of 2001, while Mark Zuckerberg was still in high school at Phillips Exeter Academy, three years before he founded Facebook at Harvard, I was a PhD student down the street at MIT, sitting in the reading room at Dewey Library studying for two very different classes: Econometrics I, taught by the world-renowned statistician Jerry Hausman, and The Sociology of Strategy, taught by the then-rising-star sociologist Ezra Zuckerman, who is now the dean of faculty at MIT’s Sloan School of Management. Ezra’s class was heavily focused on social networks, while Jerry’s class introduced us to “BLUE” estimators—the theory of what generates the best linear unbiased statistical models.

I had my statistics textbook in one hand and a stack of papers on networks in the other. As I read the statistics text, I saw that it repeated one main assumption of classical statistics over and over again—the assumption that all the observations in the data we were analyzing (the people, firms, or countries) were “independent and identically distributed (or IID).” In other words, our assumption was that none of the people in our data should be connected to each other in any systematic way. As I read the networks papers, however, I kept seeing diagrams of the complex interconnections between people. On the one hand, we were assuming everything was independent. On the other hand, the reality we live in is one of staggering interdependence.

I realized that a great deal of what we thought was unexplainable (the variance in the “independent” models) might be explained by how we are connected to one another, how information and knowledge ebb and flow between us, and how our peers’ behaviors and opinions affect our own. In 2001 there were no large digital social networking sites, but we did have lots of digital network connections through email, instant messaging, and texting. In that moment, sitting in Dewey Library, I had an epiphany: Digital social networking was going to turbocharge how information, behavior, economic opportunity, and political ideology flowed between people. It was going to transform society as we knew it and affect everything from business to politics to public health.

I remember running to the nearest Pine terminal (a computer program for sending email) and sending an email to my PhD adviser, Erik Brynjolfsson, to request a meeting. I met Erik the next day and explained that I wanted to focus my PhD dissertation on digital social networks. I told him that I thought these networks were going to be the next big thing in personal computing and that they were going to transform society. Now, Erik didn’t study social networks, and he hadn’t ever formally thought about graph theory. He was busy trailblazing research on the impact of information technology on firm productivity and economic growth. Social networks were not on his radar. To his credit, though, he humored me. He said, “I don’t really know much about networks, but you seem pretty excited about it, so we’ll figure it out together.” I don’t know if he was thinking to himself This is a phase, it will pass. PhD students typically have hundreds of ideas that go nowhere before they land on one that works. But he was supportive nonetheless, and I wrote my PhD thesis on how information flows through digital social networks. As it turned out, social networking wasn’t a phase, and it didn’t pass. Friendster was founded in 2002, MySpace in 2003, Facebook in 2004, Twitter in 2006, WhatsApp in 2009, Instagram in 2010, WeChat in 2011, and TikTok in 2012. The New Social Age was born, and I’ve been studying it ever since.

My scientific work is firmly rooted both in my deep admiration for technology and in a healthy skepticism about how it is put to use. I’m convinced we are witnessing a new era of human evolution, one in which mass automated, digitized socialization will change the way we interact, communicate, perceive our world, decide, and act. Online social networks (e.g., Facebook), microblogging (e.g., Twitter), instant messaging (e.g., WhatsApp), and collaborative knowledge production and news aggregation technologies (e.g., Wikipedia and Reddit) have fundamentally altered the way information is produced, shared, consumed, utilized, and valued. Such changes have profound implications for many of our social, political, and economic organizations, from the productivity of knowledge workers to consumer demand patterns, and from election campaigns to public health programs and mass protests.

New technologies and new modes of communication not only change the production and dissemination of information but also simultaneously record information about human interaction with incredible precision and detail. In an article published in Science in 2009, my colleagues and I argued that these new technologies and modes of communication not only change the creation and dissemination of information but make possible the development of a new field of “computational social science,” which aims to improve our understanding of the macrolevel consequences of microlevel human interaction—a long-standing “holy grail” in sociology, economics, and other disciplines. These changes are enabling new scientific studies of human behavior at population scale, and revealing new interventions that could dramatically improve the way we deal with conflict, commerce, and health.

In addition to my scientific work, I have also been an active entrepreneur and chief scientist of multiple companies, with one foot in academia and the other at the forefront of the entrepreneurial development of these new technologies. I was the chief scientist at SocialAmp, one of the first social commerce analytics companies (until its sale to Merkle in 2012) and at Humin, a social platform that The Wall Street Journal called the first “Social Operating System” (until its sale to Tinder in 2016). I have worked directly with senior executives at Facebook, Yahoo!, Twitter, LinkedIn, Snapchat, WeChat, Spotify, Airbnb, SAP, Microsoft, Walmart, and The New York Times.

Along with my longtime friend Paul Falzone, I’m a founding partner at Manifest Capital, an investment firm that helps young companies grow into the Hype Machine. From this perch, I evaluate hundreds of companies a year and get to look around the corner at what’s next. These experiences have forced me to think deeply about the business models, technologies, and machine intelligence that drive the social economy. As a scientist, entrepreneur, and investor, I have peered into the Hype Machine up close, studied its inner workings, and participated in its development. These three perspectives are always with me, which I suspect will be obvious to you as you read this book.

As a scientist, I am obsessed with rigor. I attempt at every turn to avoid making claims I cannot prove. As a result, there will be moments in this book when I present compelling evidence but stop short of making a bold claim. I will make arguments with a healthy dose of caveats along with them. The unfortunate truth is that we don’t have all the answers yet, and the answers we do have are not simple. That is part of the challenge. While science has advanced tremendously around social media and how it is affecting us, it remains nascent and at times is constrained by the platforms’ stranglehold on data. We don’t know enough about the spread of fake news, election manipulation, filter bubbles, and digital political polarization, because the research has not yet been done. But it needs to be done. So advocating for that research will be a theme of this book.

As an entrepreneur, I appreciate the difficulty of actually doing. Innovators face impossible dilemmas. Building a successful business is difficult. Building a global platform like the companies I will be discussing is next to impossible. I respect what it took to build Facebook, Twitter, LinkedIn, and the rest. I realize that some eventualities could not have been foreseen in the decisions that were made in early days. But I also know that when faced with certain truths, we have a moral duty to act. More needs to be done about the negative consequences of the Hype Machine today. I believe the true leaders of the New Social Age will be those who make the hard decisions to put social welfare above shareholder value—or perhaps those who realize that, in the long run, these goals are aligned.

As an investor, I try to distinguish the forest from the trees. When you are building a business, you are singularly committed to the survival and growth of that business. But as an investor, you also see the marketplace as a landscape of renewal. As Steve Jobs said at the Stanford University commencement in 2005, “Death is very likely the single best invention of Life. It is Life’s change agent. It clears out the old to make way for the new. Right now the new is you, but someday not too long from now, you will gradually become the old and be cleared away.” Such is also the turmoil of the marketplace. Friendster gave way to MySpace, which gave way to Facebook. Today WeChat does in one app what Facebook, WhatsApp, Messenger, Venmo, Grubhub, Amazon, Uber, Apple Pay, and many others do individually. The ongoing success of none is set in stone. The future of the New Social Age will be forged by the choices we make as entrepreneurs, investors, regulators, consumers, and citizens. I would argue that the most consequential decisions of the New Social Age are yet to come.

My Objective

My goals in this book are to describe the science of how the Hype Machine works and to explore how it affects our politics, our businesses, and our relationships; to explore the consequences of the Hype Machine for our society, both positive and negative; and to discuss how we can—through company policy, social norms, government regulation, and more advanced software code—achieve its promise while avoiding its peril.

I’ll begin by considering fake news and the weaponization of misinformation through the Hype Machine, tracking how the design of platforms like Facebook and Twitter incentivize and enable the spread of misinformation (Chapter 2). Did Russian election interference change the results of the 2016 U.S. presidential election? What should we do to stop the scourge of fake news in the 2020 election and beyond? Stay tuned.

Along the way, I’ll examine why the Hype Machine’s rise was so meteoric and why we were so susceptible to it, as individuals and as a society. I will describe the anatomy of the Hype Machine—the underlying trifecta of social technologies at the center of this inflection point in human history—and consider, in depth, the four levers through which we can shape our technological future: money, code, norms, and laws (Chapter 3). I will describe the neurological (Chapter 4) and economic (Chapter 5) forces that have “wired us into” the Hype Machine. Understanding these neurological and economic hooks will help us answer important questions about the New Social Age from a business perspective, such as why Facebook beat MySpace in the market for social networks. They will also shed light on more fundamental questions, like how the Hype Machine will impact human evolution.

I will then discuss three key societal transformations driven by the Hype Machine that are disrupting business, democracy, and public health: the personalization of mass persuasion (Chapter 6), the hypersocialization of society (Chapters 7 and 8), and the advent of the attention economy (Chapter 9). In so doing, I’ll look under the hood at the mechanics of the Hype Machine, dig into the science of online peer effects, and explore how our new radical interdependence is changing the products we buy, the people we vote for, and even who we meet and fall in love with.

After looking under the hood, I’ll zoom out to consider the societal implications of the Hype Machine and the three trends it perpetuates—for example, its implications for what’s known as the “wisdom of crowds.” Our ability to harness the wisdom of crowds and collective intelligence rests on three basic pillars: independence, diversity, and equality. The problem is that the Hype Machine erodes all three of these pillars and turns wisdom into madness. I’ll discuss how we can recapture the wisdom of crowds in Chapter 10. Next, I’ll remind us why we invented the Hype Machine in the first place by describing its positive potential in creating an incredible tsunami of productivity, innovation, social welfare, democratization, equality, caring, positivity, unity, and social progress. At the same time, I’ll discuss why the source of social media’s positive potential is also the source of its peril and how this complicates how we must adapt to it (Chapter 11).

Finally, I’ll explore how we must adapt—how business policy, government regulation, social norms, and technology design can steer our economy and our society toward a more productive future (Chapter 12). Should we break up Facebook? How should we craft privacy legislation? Is the Hype Machine a publisher or a user-generated platform that should not be responsible for the content users post? What does that mean for free speech and hate speech? You’ll be surprised by some of the answers.

At a Crossroads

The last three years have seen front-page stories about Facebook, Twitter, YouTube, and the rest of social media’s lack of transparency; their contribution to political polarization; their promotion of hate speech, racism, and the degradation of discourse; their role in the spread of fake news; and their potentially corrosive impact on our democracies and our elections.

Lawmakers have advocated regulation. Multiple U.S. congressional committees are investigating the role of Facebook and the rest of the Hype Machine in Russian election interference and the spread of misinformation online. The Cambridge Analytica controversy, in which a political consultancy used stolen Facebook data on 87 million Americans to target political ads, forced Mark Zuckerberg to testify in front of the U.S. Congress and the European Parliament as lawmakers debate what to do about the Hype Machine’s power of mass persuasion, use of personal data, and lack of control over misinformation. Sen. John Kennedy began his questioning of Zuckerberg in the Senate with an ominous opening statement: “I don’t want to have to regulate Facebook,” he said. “But, by God, I will.”

Advertisers have also pressured the platforms to clean up their act. In 2017, Marc Pritchard, Procter & Gamble’s chief brand officer, went on a public tirade about the lack of transparency in digital advertising on platforms like Google and Facebook and advertising appearing next to fake or offensive content. Then he put his money where his mouth was and cut P&G’s digital advertising budget by $200 million. In 2018, Unilever followed suit, cutting its digital advertising by nearly 30 percent in an effort to clean up the Hype Machine’s advertising ecosystem. And these weren’t just hapless public protests. In fact, P&G reported a 7.5 percent increase in organic sales growth in 2019, while cutting its online marketing budget by 6 percent. Unilever posted a 3.8 percent gain in organic sales in the same period. Understanding how they did it requires an understanding of the Hype Machine.

In the Hype Machine, everyone is a digital marketer, whether we’re fighting for ideas or for consumer dollars. A candidate in the presidential election, trying to persuade voters to his or her side; BMW, trying to persuade people to buy the new 3 Series; the small-business owner trying to grow sales; Vladimir Putin’s Internet Research Agency, trying to sow discord through misdirection—in the Hype Machine, they are all digital marketers. They are all trying to optimize the same persuasion strategies to achieve their goals. That’s why it’s so important for us to wear different hats on this journey—the regulator, the marketer, and the concerned citizen. Frequently, I will ask you to put on your digital marketer’s hat to understand the toolbox from their perspective. To make sense of what we see in our feeds every day, we’ll need to understand what they’re doing and why.

Every nation on earth has become attentive to the role of social media in society today. Regulators around the world are debating what to do about the impact of the Hype Machine on elections, business trends, competition, privacy, and fake news. Business leaders are trying to figure out how to self-regulate, through platform policies, algorithm design, software code, and alternative business models. And all of us, as parents and individuals, family and friends, are thinking about how the Hype Machine affects our lives and the lives of our children, from the impact it has on our friendships and businesses, to how we socialize and behave, to the rise of loneliness in our society. The decisions we make today in how we design, deploy, use, and regulate social media will have far-reaching consequences for years to come.

The science suggests that while social media can help foster a transparent, democratic, egalitarian society, it can also be used to erect a polarized, authoritarian police state. Today we are at a crossroads, caught between the promise and the peril, even as the system’s design is being debated worldwide.

Tectonic shifts are a regular occurrence for social media. Keeping up with the daily changes is next to impossible. Instead, I hope to offer a lasting framework to guide our thinking about the social economy. What I have learned, over twenty years of research, has taught me some general principles about how the Hype Machine works, how information and behaviors spread, how interventions in social media change behavior, and how managers, policy makers, and individuals can interface with the Hype Machine more effectively. A rigorous investigation of these fundamentals entails a harrowing intellectual journey, with plenty of plot twists and unexpected turns along the way. It’s a journey I intend to take you on in this book. And there is no better place to start than with how social media has brought us to the precipice of what some have called the “end of reality.”

CHAPTER 2

The End of Reality

Under normal circumstances the liar is defeated by reality, for which there is no substitute; no matter how large the tissue of falsehood that an experienced liar has to offer, it will never be large enough, even if he enlists the help of computers, to cover the immensity of factuality.

—HANNAH ARENDT

The market opened quietly on Wall Street on April 23, 2013. As traders sipped lattes on an unseasonably chilly morning, stocks made modest gains from the opening bell to the lunch hour. But as everyone broke for lunch, the Associated Press broke a story on Twitter that changed the market’s mood. Phones chirped at eateries in New York, Washington, and the world over as the news was retweeted again and again, creating an information cascade that swept through the Hype Machine in seconds. The tweet, which appeared at 1:07 P.M. eastern time in the United States, simply read “Breaking: Two Explosions in the White House and Barack Obama is injured.” It was retweeted over four thousand times in five minutes, which would have informed hundreds of thousands if not millions of people of the attack on the White House.

You could almost hear the iced teas and Arnold Palmers being snorted back into the glasses of those monitoring the social network. The news was shocking. Putting aside fence jumpers, who are usually tackled on sight, there have only ever been four breaches of White House security that made it to the building. So two explosions injuring the president inside the White House was big news.

The market stuttered. Then it skipped a beat. If only individual retail investors had been influenced by the news, the financial impact might have been contained. But the Hype Machine doesn’t exist in isolation. It’s coupled to systems that sense, mine, analyze, and trade on sentiment expressed on social media in real time. Dataminr, RavenPack, and other companies are constantly sifting through social media data to find the signal in the noise. When they find that signal, they seize on it and relay instructions to their institutional clients to buy or sell ahead of market trends. On this particular afternoon, the sentiment wasn’t good, and the data miners issued sell recommendations that triggered automated-trading algorithms to unload their stocks. When they did, the Dow fell instantly by nearly 200 points, wiping out $139 billion in equity value in seconds.

But the news wasn’t true. The White House was calm, and the president was fine. The tweet was fake news propagated by Syrian hackers who had infiltrated the AP’s Twitter handle. There was a terrorist attack that day—just not at 1600 Pennsylvania Avenue. The attack happened on Twitter, and the casualties were felt on Wall Street. The market rebounded, but real people lost real money as their buy and sell orders were honored. Those who were late to the fire sale lost their shirts. The “Hack Crash” of 2013 highlights the fragility of the socio-technical systems we’ve wired into the Hype Machine. When news cascades through the network, it’s hard to stop and harder to verify with enough time to prevent panic. When the news is false, it can wreak havoc on financial systems, health systems, and democratic institutions, creating real consequences from virtual falsity.

Here’s another example. When Hurricane Harvey hit southern Texas in the summer of 2017, the flooding displaced thousands and halted production at several oil refineries in the southern United States. News of gas shortages spread quickly on Twitter and Facebook as drivers posted pictures of long lines at gas stations, with makeshift signs saying they were out of gas. A panic ensued, and drivers in the region rushed to stockpile gas as if the world were ending, creating a run on fuel in Austin, Dallas, Houston, and San Antonio.

But as the authorities later revealed, there was no gas shortage. It was false news spread through social media, then picked up and reported by broadcast media. As we later learned, there was plenty of gas to go around. The refinery and highway closures only slowed deliveries. Had everyone stuck to their normal consumption, the distribution system could have handled the disruption, and there would have been no shortage. The panic and the subsequent run on gas, however, ensured there was a shortage, wrought by manic stockpiling, driven by social media.

The signature of a fake news crisis was repeating itself: false information was spreading faster than the truth, misdirecting real behaviors with real impact. Such fake news can have dramatic consequences for businesses, democracies, and public health. And although it has been around for centuries, the speed and scale with which it spreads through the Hype Machine creates a fake news crisis on steroids.

These vignettes highlight a systematic pattern in the spread of fake news (one that bears out in large-scale studies of the phenomenon that I’ll discuss shortly). When fake news isn’t completely fabricated, it typically distorts real-world information by tweaking or contorting it, mixing it with true information, and highlighting its most sensational and emotional elements. It then scales rapidly on social media and spreads faster than our ability to verify or debunk it. Once it spreads, it’s hard to put back in the bottle and even harder to clean up, even with a healthy dose of the truth.

The Syrian hack that crashed the stock market in 2013 is a case study in the economic consequences of fake news. You’ve probably heard similar stories. The fact-checking site Snopes keeps a list of “hot 50” rumors that gets updated with alarming regularity. There was a 2008 rumor that United Airlines was filing for bankruptcy, a 2017 report that Starbucks would give out free Frappuccinos to undocumented workers, and President Trump’s tweets, in March 2018, falsely claiming that Amazon was evading taxes, which sent shares of the company plummeting to their worst monthly performance in two years. But is there a systematic effect of fake news on businesses? On stock prices? Before we can understand the full implications of fake news, we need to detour through the story of a D-list actress named Kamilla Bjorlin.