A history of a forgotten company from what seems like a radically different time. The company, Simulmatics, essentially invented the field of data science applied to social sciences, and tried to apply these ideas to advertising, warfare, and law enforcement. Sound familiar?

It’s probably unsurprising that Simulmatics is forgotten: it was barely ever a functional company, and would perhaps never have achieved even the limited results it did without the quagmire of the Vietnam War giving license to all sorts of unusual and even desperate ideas in the pursuit of a way out for the US – any way out. It doesn’t work out. Perhaps it could never have worked at that time, given the limitations of computing power and data availability, but there’s a weird unwillingness to properly engage with the funded projects, partly because of attitudes of the time preventing senior staff from engaging in field work, or with working computers, or working as equals with the women and local interpreters they hired – all of whom could have helped make sense of the situation.

It’s impossible not to read Simulmatics as a precursor to other trends too: to over-promising and under-delivering, to radically hyping-up every small incremental improvement in technology, to generalising from tiny successes to society-changing possibilities. Lepore sees this as leading directly to modern Silicon Valley techno-utopianism, and it’s hard to disagree.

There was a feeling at the time that predictive analytics was ipso facto anti-democratic and even immoral. I suspect the reality is slightly more nuanced than this, and that the problems start when the prediction becomes a lever to enforce (or avoid) a particular outcome. It may not matter whether the people with their hands on the lever are government or companies, conservative or liberal; it may not matter that the lever has only a very limited power to move opinions. Understanding social trends slides very easily into manipulating them in pursuit of a specific goal, which may not be transparent to those meing manipulated.

The conclusion nails many of the defining features of the early 21st century that emerge directly from Simulmatics’ techniques, even when they remain unacknowledged. They also skewer the techno-optimism, emerging in part from the 60s counterculture, that led directly to some of today’s problems. By using data to personalise people’s exposure to information and comment, and by simplifying who could become a journalist and publisher, the internet and the web encouraged atomisation and the emergence of communities never exposed to contrary ideas: everyone gets their own facts alongside having their own opinions re-affirmed. It’s definitely the major social challenge of our times.

But the conclusions are marred slightly by their purely American focus. It’s true that the US government declined to regulate data collection and processing by private companies, but not true that governments declined: the EU’s General Data Protection Regulations do precisely this, however imperfectly, and if there is a be a solution to the widespread abuses of big data then those (and similar) regulations will play a large part.

5/5. Finished Sunday 30 March, 2025.

(Originally published on Goodreads.)