Americans who use the internet—85.5 percent of us—have made a tacit bargain with Facebook, Google, MasterCard, Verizon, and most other sites and products we use regularly. We get access to these companies’ services, and they get to scoop up, analyze, and sell our personal information. Few people question this setup, perhaps because most of us assume that our data isn’t worth much.
But that assumption is wrong.
Earlier this year, my colleague Siddhartha Aneja and I published a deep-dive study into the value of the personal information that every major website sells access to. It’s a complicated problem. Much of the value comes from advertising revenue, disclosed in annual reports and SEC filings by public companies. But we also had to determine how much of that ad revenue is derived specifically from the micro-targeting that user data makes possible, as well as how much the companies spent to gather, analyze, and market user profiles. In the end, we calculated that internet companies earned an average of $202 per American internet user in 2018 from personal data. We believe that’s a conservative estimate.
The value reflects the extraordinarily varied and detailed data that companies collect. Google collects not only the personal information you reveal when you use its search engine, but also the data that comes from whatever you do when you visit or use any of its dozens of properties—YouTube, Gmail, Google Maps, the Chrome browser, Google Pay—or apps accessed by logging in through Google. Similarly, Facebook gathers all the data crumbs you leave whenever you visit the site itself or use its Messenger service, plus whatever you do on subsidiaries like Instagram and on apps accessed by logging in through Facebook.
Amazon is newer to this business, and unlike Facebook and Google, its basic business model doesn’t rely on data-derived profits. But Amazon’s public records show that its earnings from user data likely more than doubled between 2016 and 2018. Beyond the major platforms, hundreds of other companies take part in the burgeoning personal data business. Our study also explored the revenues from digital advertising earned by smaller internet services, ranging from Snapchat and Spotify to internet media holding companies such as IAC, which owns Match.com, the Daily Beast, and Investopedia.
For a general sense of the value of people’s personal data to these companies, we started with digital advertising revenue. In 2018, Facebook earned an average of roughly $110 in ad revenue per American user. This calculation, however, ignores what it costs to collect, analyze, and market user data. According to the balance sheet on file with the SEC, Facebook earned $55.8 billion worldwide in 2018, virtually all of it from targeted advertising. Facebook also reported that relevant costs came to $20.6 billion. That implies that the value of its users’ personal information was equivalent to $35.2 billion, or 63 percent of Facebook’s earnings.Imagine the profits General Motors could achieve if it didn’t have to pay for the steel it uses to build cars. Like GM, online companies should pay their users for the valuable raw materials they provide.
In other cases, determining a company’s data-related costs is less straightforward. Google’s cost data, for example, is embedded in the balance sheets for Alphabet, the holding company that also includes Google Fiber, GV (former Google Ventures), CapitalG, Waymo, and Google Deep Mind, as well as the suite of Google-branded internet services.
Given these issues, we developed a proxy to distinguish the value of personal information to the online platforms and the costs to gather and monetize those data. We relied on a careful study by three business school economists who analyzed certain aspects of the Ad-Choices program created by the online advertising industry. In 2010, AdChoices announced that anyone could opt out of all targeted online ads by registering at a special website. While relatively few internet users opted out, the researchers found that those non-targeted ads produced 52 percent less revenue than comparable targeted ones. This suggests that 52 percent of digital advertising revenue is derived from the personal information used to target those ads to particular individuals.
Since targeted online advertising generated $108.6 billion in overall revenue in 2018, on this basis we estimate that the profits derived from American internet users’ personal data totaled $56.5 billion. Some 279.7 million Americans used the internet last year, so the value of their personal data to these online operations, after costs, came to an average of $202 per American internet user.
Our personal information is worth so much because the profiles created from it are remarkably probing and detailed. Algorithms track and save data on what we search for, what we write in emails and messages, what we buy, and everything else we do online, whether on our phones or laptops. Not only do the algorithms then build up a basic profile based on gender, age, ethnicity, and so on; they also determine our individual interests, likes and dislikes, family background, political leanings, sexual orientation, and much more. Everything we reveal online is fair game.
As the software for data mining and targeting has advanced, the revenue from digital ads and the consequent value of the data used to target them have risen rapidly. Our study found that from 2016 to 2018, the value of the information Google mined from Americans grew 40 percent; for Facebook, 85 percent; and for the latecomer Amazon, 312 percent. Overall, the value internet companies derive from Americans’ personal data increased almost 54 percent. At that rate, the number will reach $127.9 billion in four years. Adjusting for the estimated increase in the number of Americans using the internet, people’s personal data will be worth an average of $434 per American user in 2022.
Given how profitable these data operations are, the practice is sure to spread, and the commercial, social, and other uses of our personal data—and the related side effects—will proliferate. In response, we can try to regulate it, but we shouldn’t be surprised when tech companies figure out how to deflect or defeat such regulation.
Instead of regulation, then, we should consider a more direct economic response. Imagine the profits General Motors could achieve if it didn’t have to pay for the steel it uses to build cars. Like GM, online companies should pay their users for the valuable raw materials they provide.
Behind this proposal is the principle that people have a property interest in their personal information, or at least in the wealth generated by that information. This theory does not depend on any particular position on privacy. Rather, it comes directly from John Locke’s analysis of the origins of property, which are embodied in the U.S. Constitution. Locke said that a person can create property by mixing his or her labor with materials from nature, because each of us has personal ownership of our own bodies and minds and, therefore, of whatever we produce by using our minds to direct our bodies. Every piece of personal information online companies gather, analyze, and sell is the result of that very same process, so it meets Locke’s test for personal property. Someone can persuade you to give it away—but if they mislead you about its value to them, either you’re a sucker or they’re a fraud.
Congress should recognize people’s property rights over their personal information and direct that internet companies do so as well. It’s true that the structure of the internet complicates the analysis, because the value of one person’s personal information to an online platform is bound up in the platform’s access to the personal information from many millions of other people. Economists call this a “network effect.” Just as Facebook’s value to any individual increases as more people use it, so the value of any individual’s personal data to Facebook and its advertisers increases as Facebook gathers and analyzes data from millions of other users.
In other words, internet companies and their users both contribute to the commercial value of the personal profiles that drive digital advertising. A straightforward solution is thus to require the companies to share the profits from those operations with users on a fifty-fifty basis. Of course, asking internet companies to write a check to every individual user would be impossibly inefficient. Instead, each company could write a single check to the government, and the government could distribute the proceeds to every household based on the number of its internet users. So, in 2022, a family of four internet users would receive $868 in payment for their personal data. The era of free riding for online companies would be over. Corporations have gotten rich by exploiting our data. It’s time for them to share the wealth.