Once upon a time, some people developed a new technology for freely communicating with people around the world. While initially the purview of techies and hobbyists, it didn’t take long for commercial interests to notice the insanely popular new medium and rapidly move to displace the amateur stuff with professional content. But these companies had a problem. For years, people had gotten used to the idea that if you paid for the equipment to access the content, you could receive the content for free. No one wanted to pay for this new, high quality (and expensive to make) content. How could private enterprise possibly make money (other than selling equipment) in a market where people insisted on getting new content every day — heck, every minute! — for free?
Finally, a young techie turned entrepreneur came up with a crazy idea. Advertising! This fellow realized that if he could attract a big enough audience, he could get people to pay him so much for advertising it would more than cover the cost of creating the content. Heck, he even seeded the business by paying people to take his content, just so he could sell more advertising. Everyone thought he was crazy. What? Give away content for free? How the heck can you make money giving it away for free? From advertising? Ha! Crazy kids with their whacky technology. But over the course of a decade, this young genius built one of the most lucrative and influential industries in the history of the world.
I am talking, of course, about William Paley, who invented the CBS broadcast network and figured out how to make radio broadcasting an extremely profitable business. Not only did Paley prove that you could make a very nice living giving away content supported by advertising, he also demonstrated that you didn’t need to know anything about your audience beyond the most basic raw numbers and aggregate information to do it. For the first 80 or so years of its existence, broadcast advertising depended on extrapolated guesses about total aggregate viewing audience and only the most general information about the demographics of viewership. Until the recent development of real-time information collection via set-top boxes, broadcast advertising (and cable advertising) depended on survey sampling and such broad categories as “18-25 year old males” to sell targeted advertising — and made a fortune while doing it.
We should remember this history when evaluating claims by Facebook and others that any changes to enhance user privacy will bring the digital world crashing down on us and force everyone to start paying for content. Setting aside that some people might actually like the option of paying for services in exchange for enhanced privacy protection (I will deal with why this doesn’t happen on its own in a separate blog post), history tells us that advertising can support free content just fine without needing to know every detail of our lives to serve us unique ads tailored to an algorithms best guess about our likes and dislikes based on multi-year, detailed surveillance of our every eye-muscle twitch. Despite the unfortunate tendency of social media to drive toward the most extreme arguments even at the best of times, “privacy regulation” is hardly an all or nothing proposition. We have a lot of room to address the truly awful problems with data collection and storage of personal information before we start significantly eating into the potential revenue of Facebook and other advertising supported media.
Mind you, I’m not promising that solid and effective privacy regulation would have no impact on the future revenue earning power of advertising. Sometimes, and again I recognize this will sound like heresy to a bunch of folks, we find that the overall public interest actually requires that we impose limits on profit making activities to protect people. But again, and as I find myself explaining every time we debate possible regulation in any context, we don’t face some Manichean choice between libertarian utopia and a blasted regulatory Hellscape where no business may offer a service without filling out 20 forms in triplicate. We have a lot of ways we can strike a reasonable balance that provides users with real, honest-to-God enforceable personal privacy, while keeping the advertising-supported digital economy profitable enough to thrive. My Public Knowledge colleague Allie Bohm has some concrete suggestions in this blog post here. I explore some broader possible theoretical dimensions of this balance below . . . .
First off, I’ll take this opportunity to flog by white paper on the basic principles for effective privacy legislation. Since principle #1 is about putting users back in control of their personal information, it won’t surprise anyone that I start with this as the basic bedrock for any effective privacy protection scheme. I own my personal information. I have to provide access to it to some degree to function in society. But we have long recognized that I should have better choices than either to waive my rights entirely or to live in a cave and somehow figure out how to collect food and water without being spotted by drones.
To return to the instant point, we need to understand the economics of advertising and why collecting personal data has made such a huge difference in advertising economics.
Why Personal Data Fuels Modern Advertising.
Most advertising does not work. This is inevitable. People have only so much money and so much time. Even the most successful advertising campaign yields relatively modest results as compared to the number of people who actually go out and buy the product, use the service or vote for the candidate. To make matters worse (from the advertiser’s perspective), people and tastes change. Something that works today falls flat tomorrow, with no warning as to why yesterday’s super effective pitch is tomorrow’s total dud. Furthermore, lots of people generally dislike the concept of advertising (yes, there are exceptions) and like to think of themselves as independent thinkers unsusceptible to media hype.
So an advertiser has a serious challenge. How to persuade someone who doesn’t want to even see your message, let alone be persuaded by it, to buy something they hadn’t been thinking of buying, or do something they hadn’t been thinking of doing, or stop doing something they are doing. And you have to do this in an environment saturated by lots of other people doing the same thing.
In this environment, several things matter. For one thing, size. The more eyeballs you can deliver, the more valuable your real estate. That’s why the Superbowl commands such a premium. But it’s not just about number of eyeballs. Anything that gives the advertiser even a slight improvement in absolute terms in the effectiveness of ads potentially translates into millions of dollars in increased sales. So, like olympic athletes spending whatever it takes to boost their performance by even a fraction of a second, advertisers will pay premiums for anything that boosts the effectiveness of their ads.
Another problem for advertisers is that it can be damn hard to determine if your ad is actually effective. Sometimes it can be super obvious. Release a new advertisement, everyone starts talking about it, and sales shoot up astronomically. But most times it is a lot harder. Even when the product is outrageously successful, it is hard to determine what, exactly was effective (and with whom) v. what was entirely wasted effort. Black Panther has now topped over $1 billion in worldwide gross revenue. Lots of things contributed to that. As someone trying to figure out how to replicate that success, I’d love to know what worked and what didn’t to make that result happen.
Mind you, it’s not all bleak for advertisers. Consumers actually want information about products they are interested in buying. If advertisements are sufficiently entertaining, people actually seek them out. Going back to the Superbowl, consider how an entire cottage industry has emerged around the advertising independent of the game. (Some) People actively look forward to seeing advertisers strut their stuff in front of the largest audience of the year. In an ideal world, therefore, an advertiser would carefully craft an advertisement to appeal to you specifically and only show it to you when you were most likely to respond favorably to the ad. Additionally, it would be nice to have some sort of mechanism that lets me directly measure the effectiveness of the advertisement — such as an ability to monitor whether the person viewing the advertisement actually engaged with the advertisement in some way or, even better, directly linked the advertisement to the sale of the product.
Welcome to the Internet and the age of digital advertising.
How Facebook and Other Free Online Internet Services Put All This Together.
Probably the easiest explanation is this clip from Adam Ruins Everything. To put in words, modern digital advertising offers advertisers the above described Nirvana of advertising. Using information collected from individuals, digital advertisers (particularly Google and Facebook, which jointly account for about 50% of global online advertising revenue), construct profiles based on web browsing (“behavioral advertising“) to determine what sort of ads are most likely to be effective. Additionally, through the magic of predictive analytics, advertisers use literally billions of discrete bits of information to make predictions about your behavior, select the time of greatest receptivity to show you ads most likely to alter your behavior.
The more data I have, the more correlations I can establish and the more accurate I can make my predictions. This is why the race is on to sluck up ever more data from users. Even things that on their own seem pointless or trivial can become useful when crunched as part of a monstrous data set developing correlations that are even marginally better at predicting your future actions and reactions. The more data I can collect, both in terms of the number of people I surveil and the number of data points for each individual, the more accurate my predictions about what advertisements will work specifically on you. Additionally, the more I monitor you, the easier it is for me to confirm that my advertisement is actually altering your behavior — or tell me when you are adapting and my previously effective ads are less effective. This, in turn, allows me to charge advertisers a higher premium to deliver their ads.
Both Google and Facebook have all the elements needed to charge the maximum premium and dominate the market. Like the broadcast networks of the 20th Century, they aggregate gigantic audiences on a scale hitherto impossible to imagine. They do this by providing people, at no charge, access to services and content that users could not otherwise obtain. This alone would justify a huge premium similar to the premium networks can pull based on high ratings. But these companies have also pioneered super-sophisticated data collection, data analysis and ad placement technology. All these factors combined give them an enormous advantage in the market (which is why they dominate) and why they can charge such enormous premiums for their advertisements. (“Enormous” is relatively speaking. The individual placement/impression is a microtransaction, but the aggregate comes out to significantly more per person than a 30 second Superbowl ad.)
There is nothing nefarious about this. Heck, in many ways it is a lot less deceptive than things like product placement in television and movies. But just because something makes good business does not make it good public policy or devoid of negative consequences. Although people seem to have forgotten this, it is actually the job of Congress (and the states) to determine where the public interest lies and then set appropriate rules of the road for doing business rather than the “responsibility” of business to “self-regulate.” The problem is not the failure of Facebook to “behave responsibly,” but the failure of Congress to DO ITS FREAKING JOB! That’s what the whole bloody “rule of law” and “government of laws, not men” thing was supposed to be about – remember? While human beings have inherent rights, “to secure these rights, Governments are instituted among men.” These are wise words, enterprising men should quote ‘em.
Not To Interrupt This Fascinating Digression, But to Get Back To The Point of the Blog – Doesn’t All That Mean Sheryl Sandberg Is Right? Regulating Facebook Data Use Means We Lose Free Facebook?
We do not need to kill free advertiser driven service to create better privacy regulation that genuinely and meaningfully protects user information. That’s why I went through the little history lesson about broadcasting. Regulation might, depending on the nature of the regulation, diminish the premium companies like Facebook and Google can charge. But even here, it’s not entirely clear that’s the case. Advertisers pay for whatever advantage they can get. If no one could offer this level of targeted advertising based on this level of information, advertisers would still need to pay for advertising, and would pay premiums for the best audiences available. And reality is often funny. When cable TV came along and fragmented television audiences, people predicted that the advertising value of network programming would plummet. But the opposite happened. The fragmentation of audiences actually made broadcast network programming (which still generally attracts much larger audiences than cable programming) actually made broadcast advertising more valuable, because the comparative “bang for the buck” for broadcast programming was actually much higher than for any of the alternative specialized venues. So even though broadcasters had a much smaller share of the overall television audience, their comparative value to the alternatives allowed them to extract an even larger premium. Rather than bankrupt programmers as predicted, cable’s increasing aggregate audience share actually made them more profitable.
But even if there is some cost, the fact that Facebook or other companies might make slightly less money does not mean they would go from profitable to unable to sustain their current business model. This kind of simplistic argument (“regulation raises cost, therefore it will kill service/eliminate innovation/lead to worldwide famine and plague”) gets raised constantly, and – if regulators actually have the guts to follow through – generally turns out to be utterly false. Profit is a thing of margins. If a regulation imposes modest cost, the provider may have to eat that cost and take a lower margin. If the business is a low margin business, it may indeed drive some folks out of business. But Facebook, Google and other major advertising driven businesses are not marginally profitable. They are hugely profitable. Regulation may impose additional costs, but we have an awful lot of headroom before we reach the point where it would threaten free Facebook.
[I’ll skip the whole Coase thing about social cost and how what we are doing now is basically shifting cost from private companies to the public generally, etc. Y’all can read that separately if it interests you.]
Additionally, as David Dayen points out in this piece proposing that the U.S. simply ban targeted advertising, we don’t really have a lot of strong evidence about whether this targeted advertising stuff actually provides such a huge advantage over traditional advertising (and, in fact, we have some evidence to the contrary). Furthermore, as we keep learning, targeted advertising itself (not merely the act of data collection) has some potentially significant social costs. Targeted advertising, as Dayen points out, allows people to engage in racial discrimination in advertising. Scammers use it to locate vulnerable marks. Potential stalkers, for a relatively modest fee, can use targeted mobile ads to track their victims. Even if we might see the end of free service (or a switch to “fermium,” where some applications are free and others require payment), it may very well be worth it. Health codes increase the cost of restaurants. Fire codes increase the cost of rental housing. Maintaining a functional, national 911 system raises the cost of offering phone service. But we have decided as a society these things are worth the cost. We may, ultimately, make the same decision about free advertiser-driven services. But we have a long, long way to go before we have to worry about it.
It is far too early to suggest that creating real privacy protections, or even banning targeted advertising, would drive out free Facebook (or other free, advertiser-supported services). We still have free over the air broadcasting, and it remains a very profitable business despite the continuing predictions of its demise. I will add from over 20 years of experience in media and telecom law that I have learned to treat the claims of broadcasters that any new regulation (or even continuing existing regulation) will mean the end of free over-the-air broadcasting.
Stay tuned . . .