If Then

Long before Facebook and Cambridge Analytica, a small New York company called Simulmatics promised to forecast how people would vote, shop and fight. Founded in 1959, it built what it called a People Machine, a computer simulation of voters, worked for the Kennedy campaign and the Defense Department, went public in 1961 and closed in 1970.

Lepore, a Harvard history professor, New Yorker staff writer and author of These Truths, found the company's papers in MIT's archives. She tells the story through its people and the politics of the 1960s, and argues that Simulmatics planted the habits of modern Silicon Valley: confidence that behavior can be modeled, a hunger for data and indifference to the harm of getting it wrong.

The book was longlisted for the 2020 National Book Award for nonfiction and shortlisted for the Financial Times and McKinsey Business Book of the Year. Publishers Weekly found it colorful but disjointed, with a confusing chronology, and said Lepore does not always separate the company's self-promotion from what it achieved.

This is a history book, not a finance book. Finance readers will find a useful early case of a data company selling prediction it could not deliver, and an IPO built on hype, worth keeping in mind when judging claims about modeling human behavior.

1. Prediction has always been oversold. Simulmatics promised to forecast human behavior with early computers, and some clients found its results overpriced, sloppy or wrong.
2. Data businesses are old. Voter segmentation, consumer targeting and behavioral modeling date to the 1950s and 1960s, which puts claims of novelty in today's data firms in perspective.
3. Hype can fund a company for a while. Simulmatics went public and won government contracts on the strength of its pitch, and fell quickly once results failed to match it.
4. Models carry their makers' assumptions. The company's work in Vietnam and on riots shows how simulations encode the views of the people building them, with real consequences.
5. Ethics lag technology. Lepore argues the founders rarely asked whether their tools should be used, a question that returns with every wave of data-driven business.

Lepore tells the story chronologically, from the company's founding to its collapse, with an epilogue tying it to the present.

1. Founding. Greenfield and Pool assemble social scientists in 1959 to sell computer simulation of behavior, first to advertisers and then to political campaigns.
2. The Kennedy campaign. In 1960 the People Machine models voter blocs and argues that a firm stance on civil rights would help Kennedy, and the firm's role becomes public after the election, setting off alarm about computers in politics.
3. Going public and expanding. Simulmatics sells stock in 1961, builds election-night models for The New York Times, and becomes the thinly disguised subject of Eugene Burdick's 1964 novel The 480.
4. War and riots. Through Pool's government ties, the company wins Pentagon contracts to study Vietnamese villagers and support pacification, runs a Saigon office, and studies media coverage of the 1967 riots for the Kerner Commission.
5. Collapse and legacy. Contracts end amid criticism, debts and personal turmoil, the company closes in 1970, and Lepore traces its ideas forward to today's data-driven platforms and targeted campaigns.

Longlisted • National Book Award (Nonfiction) Best Books of 2020 • Financial Times Best Books of Fall 2020: O, The Oprah Magazine, The Observer, Boston.com Most Anticipated Books of Fall 2020: TIME A revelatory account of the Cold War origins of the data-mad, algorithmic twenty-first century, from the author of the acclaimed international bestseller These Truths. The Simulmatics Corporation, launched during the Cold War, mined data, targeted voters, manipulated consumers, destabilized politics, and disordered knowledge—decades before Facebook, Google, and Cambridge Analytica. Jill Lepore, best-selling author of These Truths, came across the company’s papers in MIT’s archives and set out to tell this forgotten history, the long-lost backstory to the methods, and the arrogance, of Silicon Valley. Founded in 1959 by some of the nation’s leading social scientists—the best and the brightest, fatally brilliant, Icaruses with wings of feathers and wax, flying to the sun”—Simulmatics proposed to predict and manipulate the future by way of the computer simulation of human behavior. In summers, with their wives and children in tow, the company’s scientists met on the beach in Long Island under a geodesic, honeycombed dome, where they built a People Machine” that aimed to model everything from buying a dishwasher to counterinsurgency to casting a vote. Deploying their People Machine” from New York, Washington, Cambridge, and even Saigon, Simulmatics’ clients included the John F. Kennedy presidential campaign, the New York Times, the Department of Defense, and dozens of major manufacturers: Simulmatics had a hand in everything from political races to the Vietnam War to the Johnson administration’s ill-fated attempt to predict race riots. The company’s collapse was almost as rapid as its ascent, a collapse that involved failed marriages, a suspicious death, and bankruptcy. Exposed for false claims, and even accused of war crimes, it closed its doors in 1970 and all but vanished. Until Lepore came across the records of its remains. The scientists of Simulmatics believed they had invented the A-bomb of the social sciences.” They did not predict that it would take decades to detonate, like a long-buried grenade. But, in the early years of the twenty-first century, that bomb did detonate, creating a world in which corporations collect data and model behavior and target messages about the most ordinary of decisions, leaving people all over the world, long before the global pandemic, crushed by feelings of helplessness. This history has a past; If Then is its cautionary tale.