John F Kennedy on the US presidential campaign trail in 1960 © The LIFE Picture Collection via Getty

In 1960, media reports of dark forces behind John F Kennedy’s winning presidential campaign caused what Jill Lepore calls a “national hullabaloo”. America’s new leader, it was widely reported, had clinched the victory with the help of a “secret weapon”: a super computer that crunched troves of data to profile voters, allowing Kennedy to better target his political messaging before the polls opened.

Such seemingly awesome and ominous technology was the creation of the Simulmatics Corporation, a long-forgotten tech company that pioneered the use of data science in politics and, according to Lepore, in the process “invented the future”.

In her book If Then, the Harvard professor and New Yorker writer meticulously chronicles how Simulmatics laid some of the earliest foundations for the field of predictive analytics, today wielded by internet platforms, advertisers and political strategists to help sell consumer products or election candidates. The chaotic band of scientists, psychologists, and slick Madison Avenue advertising prophets have been — until now — the unknown grandfathers of Facebook and Google, and all of their whizzy algorithms.

“Commentators accused the Trump campaign of using a ‘weaponised AI propaganda machine’, describing a new and ‘nearly impenetrable voter manipulation machine’,” writes Lepore. “New? Hardly. Simulmatics invented that machine in 1959.” And so history repeats itself. The book could not be more prescient as Facebook continues its soul-searching in the wake of the Cambridge Analytica scandal and ahead of a highly polarised November US presidential election.

Launched in 1959, Simulmatics injected big data and computing into politics against a testy backdrop of Eisenhower-Kennedy-Nixon cold war paranoia, civil rights protests and later, student antiwar movements. Headed by the magnetic Ed Greenfield — a man who “sold nothing so well as himself” — it espoused lofty, liberal aims: Greenfield dreamt of using predictions to defeat communism, protect democracy and win the Vietnam war.

Through conducting surveys and computer analysis, he planned to be able tell a candidate “the consequence of taking a position on any issue, anywhere, state by state, county by county, voter by voter, issue by issue”. To do so, it would split up the US population into 480 voter types, cross-checking these against previous voting returns, in order to forecast future behaviour.

Simulmatics Corporation at the New York Times in 1962

Such high-flown ambitions never quite survived contact with reality, and Simulmatics collapsed into bankruptcy in 1970. And yet its mission would become a blueprint for millions of companies in the profit-seeking private sector, from advertisers to banks, according to Lepore: “Collect data. Write code. Detect patterns. Target ads. Predict behaviour. Direct action. Encourage consumption. Influence elections.”

Despite its visionary acclaim, the work the company carried out was consistently shambolic. One contract in Saigon was aimed at helping the US Department of Defense work out how to win over Vietnamese “hearts and minds”. Simulmatics ran up against translation issues, unco-operative locals, then “allegations of negligence, malfeasance, and even fraud”. Some of their methods were so ill-thought-through as to be verging on embarrassing. (In one instance, the company hired a Freudian psychoanalyst to analyse a handful of Vietnamese citizens. His conclusion? That they all had Oedipal issues.) The contract was pulled.

Another domestic effort, this time designed to help local authorities in Rochester, New York, predict when race-related riots were likely to happen, failed to produce any concrete results. Even the Kennedy campaign argued they did not follow the company’s recommendations in 1960, despite commissioning research from the group.

In Lepore’s telling, it is not just the company that loses its way. She outlines the unravelling of many Simulmatics staffers and their long-suffering, intellectually frustrated wives. Alcoholism and mental illness are commonplace. Their moral compasses waiver as the world around them grows more polarised and bleak, with the drawn-out war in Vietnam and assassinations of JFK, Martin Luther King and Bobby Kennedy weighing especially heavy.

In an arresting image of the company’s elitism, Lepore describes Simulmatics’ Saigon headquarters, headed by the lightly qualified Charles Ramond, as a lavish venue for expat parties. “Inside the villa, the Ramonds were serving cocktails. Outside, people were dying in the streets, blasted, dismembered, splattered.”

So, why the failure to deliver results? Perhaps Simulmatics was “hobbled by its time”, Lepore offers — by “the technological limitations of its day” such as scarce data and weak models.

Still, she does not explicitly explore the value of predictive analytics today, which is still up for debate, leaving unresolved the question of whether it is indeed a science. (To some, Cambridge Analytica sold snake oil, others believe it was a democracy destroyer.)

Without any clarity on how impactful their work was, and a fleshed out comparison to today’s multibillion-dollar data science industry, it is hard to determine how dangerous Simulmatics’ legacy is in reality. From the outset, Greenfield’s sci-fi vision — “if they collect enough data about people and feed it into a machine, everything, one day, might be predicable” — has its naysayers, even from within. Eugene Burdick, a political scientist-turned-novelist who worked with Greenfield at one point, warned: “This may or may not result in evil . . . Certainly it will result in the end of politics as Americans have known it.”

Tellingly, Lepore gently advocates for regulation of the field, including around privacy, noting that while the US government in the late 1960s focused on how to manage the potential explosion of its own agencies wielding big data, they did not consider that private enterprises might go on to do just the same.

But her conclusions sometimes seem contradictory. In the epilogue, she casts Simulmatics as building “a very early version” of “a machine that applies the science of psychological warfare to everyday life . . . manipulates attention . . . divides voters, atomises communities . . . and undermines democracy.” With the same breath, both the work of Simulmatics and today’s equivalents are dubbed “flimflam”.

For all of Simulmatics’ efforts at automating prediction, it is company executive Ithiel de Sola Pool, an MIT academic with a focus on social networks, who in Lepore’s telling proves to be the most accurate prediction machine — foreseeing the “data-mad and near-totalitarian twenty-first century” that he was instrumental in helping to create.

“In the coming atomised society, the information the citizen gets will arise from his own specific concerns,” he wrote in 1968, predicting a communications revolution, “customised news feeds” and the dismantling of party politics for a “politics of self, every citizen a party of one”.

Hannah Murphy is the FT’s technology correspondent

Get alerts on Non-Fiction when a new story is published

Copyright The Financial Times Limited 2021. All rights reserved.
Reuse this content (opens in new window)

Follow the topics in this article