The attention economy: a history of the term, its economics, its value, and how it is changing politics

Work your way to the end. I explain why TikTok users value the site by $5,600 per year and why the politics of attention explains Cambridge Analytica, Hunter Biden’s laptop, and cancel culture.

A lot of new concepts from tech have entered into the wider culture in the last five years. Two of them have been especially sticky, the attention economy and surveillance capitalism.

There are good reasons that these ideas have caught on.

Surveillance capitalism analyzes platforms as suppliers of online goods and services. In contrast, the attention economy brings focus to the consumption of attention, the limits of focus. The attention economy is thus broadly concerned with demand.

Surveillance capitalism and the attention economy. Supply and demand. The two are twinned in digital spaces.

What follows is my way of working through a small piece of the attention economy. It explores the history of the idea, the fundamental economics, and some of the implications. Right now I am in the middle of a research project with a student to define and systematize the interactions between these two concepts. There is a lot of research but there isn’t much systematization.

Importantly, it provides an intellectual base for my work at CGO explaining how digital competition works.

It is also meant as an extended response to that old adage, if you aren't paying, then you're the product. I have always hated that saying. You are never not paying. Even when you are on Twitter and Facebook you are paying, you pay with your attention.

Social media platforms might be free in price, but they are never free of opportunity cost. After all, time spent on Facebook can’t be spent hiking, playing basketball, or birdwatching.

There is a lot of criticism directed at Facebook and Instagram and TikTok. But not as many pieces that seek to understand why people stay on these sites. The attention economy always entails an opportunity cost. But the reverse is true as well. It means that people value the attention economy to a large extent.

These are the foundations, I think, to understand the new attention economy.

A short history of the term

In a 1971 paper, written as part of a colloquium, Nobel Laureate Herbert A. Simon coined the term, the attention economy. The title is mundane, “Designing Organizations For An Information-Rich World,” and the paper is short. But, the content is fascinating since it walks through two questions that we are still living with today, “How long has the world been rich in information? What are the consequences of its prosperity, if that is what it is?”

Simon was quick to point out that, yes, information is abundant. But information abundance is not the most interesting thing to dwell on. The consequences of its prosperity are far more interesting.

As he famously wrote, a wealth of information causes a “dearth of something else: a scarcity of whatever it is that information consumes.” Information consumes attention, and thus, a “wealth of information creates a poverty of attention and a need to allocate that attention efficiently among the overabundance of information sources that might consume it.”

As he continued, “It is not enough to know how much it costs to produce and transmit information; we must also know how much it costs, in terms of scarce attention, to receive it.” Because time acts as a constraint and because it takes time to consume information, it is possible to measure the scarcity of the resource by the time that a recipient spends on it. We’ll come back to this later.

But Simon’s idea would never have taken off, it would never have become a commonplace term, if it weren’t for Michael Goldhaber.

Goldhaber began noticing the changes brought on by the computer in the mid-1980s. As Charlie Warzel explained in the New York Times,

[Goldhaber] was obsessed at the time with what he felt was an information glut — that there was simply more access to news, opinion and forms of entertainment than one could handle. His epiphany was this: One of the most finite resources in the world is human attention. To describe its scarcity, he latched onto what was then an obscure term, coined by a psychologist, Herbert A. Simon: “the attention economy.”

Goldhaber continued to work on the idea and lecture on the topic throughout the 1980s and into the 1990s. Then in 1997, an article penned by him titled “Attention Shoppers” appeared in Wired. The article was prescient in laying the attention economy’s dynamics and consequences. But it was just as important in giving a name to an economy that was just emerging.

Goldhaber democratized the concept.

Rereading that Wired article nearly a quarter of a century later, it is impressive how much Goldhaber got right. He was right that an “economy built on [attention] will be different than the familiar material-based one” because of the inequalities inherent in a transaction.

If I pay attention to you, he explains, then I cannot pay attention to someone else. This new way of interaction “promises nearly everyone a chance at attention from millions…or by an audience of peers whose attention we value most.”

Goldhaber was also correct that “The Net also ups the ante, increasing the relentless pressure to get some fraction of this limited resource.” The new attention economy, he further surmised, would create “huge inequality between stars and fans” and it would “keep us from reflecting, or thinking deeply (let alone enjoying leisure).” Altogether, “we will be so engrossed by efforts to capture our attention that we will shortchange those around us, especially children.”

Where Goldhaber strayed was in believing that advertising would be completely supplanted online. Instead, ads have supported the development of the attention economy. But he never strayed far from economic analysis.

Economic choices in the attention economy

Simon (1971) is still worth reading today because it offers a language to understand our information-rich world and some key questions to consider. Indeed, I find it helpful to categorize the questions into two levels of analysis. There is a personal choice component to the attention economy but there is also a systems approach as well. In other words, the analysis from Simon might be split into micro and macro. In this essay's last section, some macro insights are laid out.

Much like the microeconomics of econ 101, the microeconomics of attention economy is best characterized as choice subject to a budget constraint.

But this is not any normal budget constraint. Users of Facebook and TikTok don’t have to pay $10 a month for the service. They are not limited by money. Users are limited in their time instead. In a world of zero-priced information, attention becomes the budget line.

Simon makes this clear, “Human beings, like contemporary computers, are essentially serial devices. They can attend to only one thing at a time.” Naturally then, a key method to understanding attention markets lies in understanding which activities are displaced and the rate at which this displacement occurs.

People’s lives are filled with other demands. These are demands for education, income, taking care of the kids, cleaning up the house, exercise, and play. The attention economy has to vie for priority, and it might be chosen last for many working-age people or parents. But if there are no other good options, it might be a good first choice.

Gray Kimbrough’s recent paper on playing video games seems to confirm this notion. Through time-use data, Kimbrough has found that video gaming has largely displaced other kinds of entertainment over time like watching television, movies, and streaming video for men.

At the same time, employed men saw only modest increases in gaming. The biggest gains in video playing could be found instead in full-time students and non-employed men. Most telling, men who recently left employment didn’t “prefer to spend significantly more time gaming than currently employed men, contrary to the hypothesis that preferences for gaming could be enticing men to leave the labor force.”

The labor-leisure choice helps to capture what is going on here.

Someone can either work or they can relax. (Let’s hold off on education for a moment.) I tend to think of labor in the broadest sense. Labor isn’t just working for a wage, it can be used as a stand-in for all of those things you must do to survive. You must work, but you must also clean the house, and fix dinner at a minimum. Unpaid domestic work is still labor nonetheless. In the same way, leisure is also a broad category, which includes the choice to participate in the attention economy, ride a bike, watch a movie, play video games, and everything else.

The decision to go onto Instagram or login to Fornite lay downstream of that initial labor-leisure decision. Employed men have already chosen to work, so it is not surprising that they aren’t playing video games anymore. Video games might replace other forms of leisure, but the amount of leisure is basically fixed. On the other hand, unemployed men, even if laid off, can consume more leisure and play more video games.

While there is limited work on the U.S. labor force, Vilhelmson, Thulin and Elldérit (2016) found for people in Sweden that online activities fill the space created when “other fixed and mandatory commitments of everyday life are fulfilled.” 

As for education, the canonical labor-leisure choice is typically expanded to include school as an investment phase. So now, it is an invest-labor-leisure choice. Again, the same insights apply. If you are young and aren't working but are going to school, then it makes sense why video games would take up more of your time.

A small aside: I think the debate over teen screen time is focused on the wrong issue at hand. Teens and children should be learning about the world, they should be in an investment stage. In other words, they tend to use devices to consume media when they should be learning.

The attention economy highlights the precious resource of focus. Individuals use their limited attention by playing games, consuming media, or interacting online. Attentions can be understood as a kind of budget line, a limit on our ability to consume. The value of the attention economy is the time spent trying to obtain and consume that information. 

The value of the attention economy

Simon’s 1971 paper also contains an anecdote that I find amusing. It seems that Simon had a habit of telling his friends how much they really paid to read the New York Times or Washington Post. Apparently, “Making the calculation usually causes them some alarm, but not enough for them to cancel their subscriptions. Perhaps the benefits still outweigh the costs.”

The idea is simple and key to understanding the attention economy.

By spending an hour reading the newspaper, he would say to them, you give up the ability to work another hour and make another hour's worth of a wage. The opportunity cost is every hour you don’t work.

Let’s say that one of his friends made $100k a year and worked 37.5 hours a week. In this case, the hourly wage comes to $51.28. Reading the newspaper is time spent not working. So just over 45 minutes a day reading the newspaper comes to about $40 when the opportunity cost of wages is added.

That Sunday New York Times doesn’t just cost $5 for the hardcopy, he would tell them. It costs you $45.

While Simon was given a Nobel prize for his work in decision theory, he was also a democratizer of this more inclusive concept of cost. Economists now tend to put this method of cost accounting under the bigger umbrella of the shadow price.

I find the idea of the shadow price compelling. But as a policy nerd, I find its origin story even more fascinating.

Stick with me for this digression.

After WWII, California began to grow in population. A big number of these new people ended up in the arid areas around Los Angeles and San Diego. Strain on water resources spurred the State of California to start the California State Water Project (SWP) in the early 1950s. The massive water management project collects water from rivers in Northern California and pulls it down south via aqueducts, pumping stations, and power plants.

One of the constituent projects in the SWP was the damming of the Upper Feather River Basin, a river shed with the most precipitation in the state. The dam was built between 1962 and 1967, but planning began in 1955. Two years later, an plan to dam the river had been put together and included a section trying to estimate the value of the reservoir’s recreation.

But project managers at the time were flummoxed: “The nature of recreation benefits is such that a detailed economic sizing study could not be made for reservoirs whose principal use would be for recreation purposes.”

Meanwhile, in going through a spree of building dams through the 1950s, California leadership also learned firsthand how big the demand for recreation was. In some cases, reservoirs that were just for recreation were dangerously overcrowded.

Andrew Trice and Samuel Wood saw both of these things happening and recognized the connections. At the root of all of these problems, the artificial lakes weren’t being properly valued. Their seminal paper estimated the recreational value of the lake by adding up all of the time that people gave up to get to it.

The total cost of travel to the lake turns out to be a good proxy for its value. More importantly, the travel cost method is a flexible framework for pricing difficult-to-price goods because it focuses on time. While the original research in this space was concerned with pricing lakes, parks, and other recreational places, the logic can be extended to attention markets.

To be fair, trying to put a price on these types of environmental goods seems grisly. It is understandable why some look at economists with suspicion. I can hear the response now: How could you price a lake or a river? That’s monstrous.

I think that’s missing the point.

Shadow prices aren’t perfect proxies for the price, obviously, but they bring into focus those things that are given up. Shadow prices focus our attention on the opportunity cost of a thing.

What is my shadow price for writing this newsletter? It’s high.

In the standard workhorse model, it is assumed that people will choose to work until their take-home pay equals the marginal rate of substitution for leisure. Since it is the effective wage that matters, economists value leisure time through after-tax average wage rates.

Let’s run through the calculation real quick for TikTok:

• Users spend about 40 minutes per day on the site.

• Average hourly wage earnings were $32.08 per hour in June 2022.

• The Congressional Budget Office estimates that the economy-wide marginal tax rate on labor income at 28.3 percent.

• Roughly speaking then, the current after-tax average wage rate stands at $23 per hour. This rate offers a rough estimate of the average value of an hour on social media. 

• Forty minutes a day for an entire year comes to a total value of TikTok of $5,596.66.

• By way of reference, a couple of years back I ran through the numbers for Facebook and found that its value was $4,886.56 for the average user.

I will be the first to admit that these numbers are rough approximations. Who knows how much they’re off. But even if they are off by half, people on TikTok still value the site by a minimum of $2800. Everyone faces tradeoffs. Shadow prices are really just tools to count them up.

With all of this as background, if you aren't paying, then you're the product, makes little sense. Social media platforms might be free in price, but that doesn’t mean they are free from opportunity cost. Consumers clearly give up their time to go online and interact. That has got to be worth something, right?

The macro-level of Simon

Reading Simon (1971) in 2022 is a bit jarring because so much is contained in so few pages. While a lot has been said about his micro analysis, Simon also considered what would happen to organizations. His analysis stressed the filtering effect from new technologies.

How is attention conserved at the organizational level? Well, companies, non-profits, governments, and other organizations use information processing systems, which he abbreviated as IPS. In turn, an IPS acts as an information condenser.

You know them and use them every day. Microsoft Word, Excel, Google Search, and Spotify are all information condensers. These systems are designed with a goal in mind. They condense information into a platform or an application, which allows for easy access of a type pad, a spreadsheet, a powerful search function, and on and on.

Naturally then for Simon, “the crucial question is how much information it will allow to be withheld from the attention of other parts of the system.” The macro level analysis here describes perfectly what is happening online. How much information is being withheld from the attention of others is a concise definition of the problem of attention politics.

Echo chambers, filter bubbles, Cambridge Analytica, Hunter Biden’s laptop, deplatforming, AI bias, cancel culture. At core, these tensions are about information being withheld from the attention of other parts of the system. These are questions about filter politics. The Filter Bubble Transparency Act, Texas’s HB 20, and Florida’s SB 7072 are all bills meant to deal with filter politics as well.

The attention economy has brought about an age of filter politics. That’s a big change we will be digesting in the 21st Century.

I will have more on all of this later.