Surveillance Capitalism
The business model that turns your everyday behaviour into free raw material, then sells predictions about what you'll do next.
Also known as Behavioural surplus capitalism · The data-extraction economy
What surveillance capitalism means
Surveillance capitalism is an economic system that treats your private experience as free raw material: it watches what you do, turns that behaviour into data, and then sells predictions about what you will do next. The term was coined by the Harvard scholar Shoshana Zuboff, who set it out in full in her 2019 book The Age of Surveillance Capitalism. Her central claim is that around the turn of the millennium, companies such as Google stumbled onto something new and immensely profitable. The traces we leave behind as we search, scroll, message and move turned out to be worth far more than the free services we were getting in return.
Under older forms of capitalism, a firm made money by selling you a product or a service. Under surveillance capitalism, you are not the customer, and you are not even the product. You are the source of the raw material. The paying customers are the businesses that want accurate guesses about your future behaviour, and the product they buy is the prediction itself.
This is why so many digital tools feel free. The search engine, the map, the social network and the email account cost you no money, because the payment is taken in a different currency: your behaviour, gathered quietly and at enormous scale.
How surveillance capitalism works
The story Zuboff tells begins with a discovery. In its early years, Google collected data about how people searched mainly to improve the search engine. After the dot-com crash of 2000, under pressure to make money, the company realised that the leftover data, the digital exhaust most firms ignored, could be used to predict which adverts a person would click. That insight became a template the rest of the industry copied.
The cycle starts with extraction. Every click, like, pause, search, location ping and voice command can be captured. Some of this is fed back into making the service better, but a great deal of it is set aside for the company’s own ends. Zuboff calls this leftover store behavioural surplus: data that goes beyond anything needed to make the product work.
From behavioural surplus to prediction products
That surplus is poured into machine learning systems that hunt for patterns. The output is a stream of forecasts about what you are likely to do, in Zuboff’s phrase, ‘now, soon and later’. Will you book a holiday this month? Are you about to switch banks? Are you anxious enough to respond to a particular kind of advert? These forecasts are bundled together into what she calls prediction products.
Behavioural futures markets
The prediction products are then sold in what Zuboff names behavioural futures markets: marketplaces where other businesses place bets on your future choices. Advertisers were the first big buyers, but the same forecasts can be sold to insurers, lenders, employers and political campaigns. Because a sharper prediction fetches a higher price, the system is permanently hungry. It always wants more data, about more people, across more of life. Zuboff calls this drive the extraction imperative.
The final stage is the one people find hardest to sit with. The surest way to predict behaviour is to give it a quiet push in a chosen direction. Zuboff names the resulting power instrumentarianism: the subtle shaping of what we do, often without our awareness, through design, defaults, notifications and rankings. Many dark patterns, the deliberately confusing interface tricks that nudge you towards a tap or a purchase, exist to keep this cycle turning.
A one-way mirror
A defining feature of surveillance capitalism is how lopsided it is. The firms come to know an extraordinary amount about us: what we read, who we love, where we go, how our mood shifts across the day. We, in turn, know almost nothing about them. We cannot see what data they hold, how their models reach their conclusions, or who is buying the predictions made about us. Zuboff calls this a new division of learning, and warns that whoever controls it holds a quiet but immense power. The knowledge flows one way, and so does the advantage. This is why she argues that the real issue is not privacy alone, but authority: the question of who gets to know, decide and shape the future, and on whose behalf.
Surveillance capitalism in everyday life
Once you know the shape of the deal, you start to see it everywhere. A free torch app that asks for your contacts and location is not being generous; it is collecting surplus. A smart speaker listens in the kitchen, a fitness band logs your sleep and heart rate, a connected car records where you go, and a supermarket loyalty card trades small discounts for a detailed map of your habits. The exchange is rarely spelled out in plain words, and the privacy policy that supposedly explains it is designed to be scrolled past.
Surveillance capitalism is closely tied to the attention economy, where your focus is the scarce resource that platforms compete to capture. The link is direct: captured attention produces behaviour, behaviour produces data, and data feeds the surplus. A recommendation feed that keeps you watching one more video is not only selling your attention to advertisers; it is also generating a fresh stream of signals about what holds you, which makes the next prediction sharper. The longer you stay, the more the system learns, and the more it learns, the better it becomes at keeping you there.
It is also a clear case of commodification, the turning of things into products that can be bought and sold. Parts of life that once sat outside the market, such as your moods, your friendships and your movement through a city, are repackaged as something tradable.
The system leans, too, on a feeling that resistance is pointless. When opting out means giving up the tools everyone else uses for work, school and staying in touch, refusal can feel unrealistic. That sense that there is simply no workable alternative is what makes capitalist realism so convenient for the firms involved. And because prediction products can be sold to a political campaign as readily as to a shoe shop, surveillance capitalism connects to older worries about how manufactured consent is produced and steered.
Surveillance capitalism, data colonialism and technofeudalism
Surveillance capitalism is one of three influential attempts to name what the tech giants are really doing, and the three are best understood together. Data colonialism describes the same extraction through the lens of historical empire, arguing that human life itself is being claimed as territory to be owned and mined. Technofeudalism goes a step further, suggesting that the platforms have stopped behaving like ordinary capitalist firms and now collect rent like medieval lords. Zuboff’s account keeps the word capitalism, but insists this is a mutant version of it, as different from the factory age as the factory was from the farm.
Which label fits best is a genuine debate, but the mechanism matters more than the name. Once you can see the cycle clearly, watch, extract, predict, nudge, sell, you begin to notice it humming away beneath the ordinary apps and devices you reach for without thinking. That noticing is the whole point. It turns a system that depends on staying invisible into something you can question, regulate and design differently.
How to spot it
Ask who pays, and what they get. If a digital service is free, the likely answer is that your behaviour is the thing being sold. Watch for permissions that reach far beyond what an app needs to work, such as a torch app wanting your contacts and location. And notice the small uncanny moments: a feed that seems to know your mood, or an advert for something you only ever said out loud. That is prediction at work.
A thought to hold onto
The old saying goes: if you are not paying for the product, you are the product. Surveillance capitalism sharpens it. You are not even the product. You are the free raw material, and the product is a prediction about what you will do next, sold to someone you will never meet.
Why it matters now
Generative AI runs on data, which hands the extraction cycle a powerful new engine and a fresh excuse to gather more of it. The same predictions that sell trainers can quietly sort people for jobs, loans, insurance and policing. Seeing the mechanism is the first step towards asking the questions that design choices and regulation still leave open.