This article is part of 20in20, a series of 20 blog posts in 20 days to kick off the blogging year 2020. This is 20in20:01.
A note to put this in perspective: This blog post doesn’t pull any punches, and there will be legitimate exceptions to the broad strokes I’ll be painting here. But I believe that this category error is real and has disastrous negative effects. So let’s get right to it.
I’ve been thinking a lot about category errors: The error of thinking of an issue as being in a certain category of problems even though it is primarily in another. Meaning that all analysis and debate will by necessity miss the intended goals until we shift this debate into a more appropriate category.
I increasingly believe that it is a category error to think of online advertising as a means to fund the creation of content. It’s not that online advertising doesn’t fund the creation of content, but this is almost a side effect, and that function is dwarfed by the negative unintended consequences it enables.
When we discuss ads online, it’s usually in the framing of funding. Like this: Ads pay for free news. Or: We want to pay our writers, and ads is how we can do that.
To be clear, these are solid arguments. Because you want to keep news and other types of journalism as accessible to as many people as possible (and it doesn’t get more accessible than “free”). And you do want to pay writers, editors and all the others involved in creating media products.
Those sentences make sense if we consider them in their original context (newspaper ads) as well as the original digital context (banner ads on websites).
There, the social contract was essentially this: I give my attention to a company’s advertisement, and that company gives the media outlet money. It wasn’t a terribly elegant trade in that there are still (at least) three parties involved, but it was pretty straightforward.
Fast forward a few years, and tracking of success of individual ads gets introduced, which happens in steps: First, how many times is this ad displayed to readers? Then, how many readers click it? Then, how many readers click it and then follow through with a purchase? There’s a little more to it, but that’s the basic outline of what was going on.
Fast forward a few years again, and now we have a very different picture. Now, ads place cookies on readers’ devices, and track not just how often they’re been displayed or clicked or how often they convert to a purchase.
Contemporary tracking ads and their various cookies also do things like these: Track if a reader moves the cursor over the ad. Tracks from which website a reader comes to the current website and where they head after. Track the whole journey of any user through the web. Track the software and hardware of any reader’s devices and create unique fingerprints to match with other profiles. Match movement through the web with social media profiles and activities and likes and shares and faves. Track movement of the reader in the physical world. Build profiles of readers combined from all of this and then sell those in aggregate or individually to third parties. Allow extremely granular targeting of advertisements to any reader, down to the level of an ad personalized for just one person. And so on.
This is nothing like the social contract laid out above, even though the language we use is still the same. Here, the implied social contract is more like this: I get to look a little at a website without paying money, and that website and its owner and everyone the owner chooses to make deals with gets to take an in-depth look at all my online behaviors and a significant chunk of my offline behaviors, too, all without a way for me to see what’s going on, or of opting out of it.
And that’s not just a mouthful, it’s also of course not the social contract anyone has signed up for. It’s completely opaque, and there’s no real alternative when you move through the web, unless you really know about operational security in a way that no normal person should ever have to, just in order to read the news.
This micro-targeting is also at the core of what might (possibly, even though we won’t have reliable data on it until it would be too late, if confirmed) undermines and seriously threatens our political discourse. It allows anti-democratic actors of all stripes to spread disinformation (aka “to lie”) without oversight and it’s been shown to be a driver of radicalization by matching supply with demand at a whole new scale.
Even if you don’t want to go all the way to this doomsday scenario, the behavior tracking and nudging that is supposed to streamline readers into just buying more stuff all the time (you probably know how it feels to be chased around the web by an ad for a thing you looked up online once) without a reasonable chance to stop it is, at best, nasty. At worst, illegal under at least GDPR, as a recent study demonstrated. It also creates a surprising — and entirely avoidable! — carbon footprint.
So, to sum up, negative externalities of tracking ads include:
- micro-targeting used to undermine the democratic process through disinformation and other means;
- breaches of privacy and data rights;
- manipulation of user behavior, and negatively impacting user agency;
- and an insane carbon footprint.
Of course, the major tracking players benefit a great deal financially: Facebook, who make a point of not fact-checking any political ads, i.e. they willingly embrace those anti-democratic activities. Google, who are one of the biggest provider of online ad tracking and also own the biggest browser and the biggest mobile phone operating system, i.e. they collect data in more places than anyone else. And all the data brokers, the shadiest of all shadow industries.
Let me be clear, none of this is acceptable, and it is beyond me that at least parts of it are even legal.
So, where does that lead us?
I argue we need to stop talking about tracking ads as if they were part of our social contract for access to journalism. Instead, we need to name and frame this in the right category:
Tracking ads are not a funding method for online content. Tracking ads are the infrastructure for surveillance & manipulation, and a massive attack vector for undermining society and its institutions.
Funding of online content is a small side-effect. And I’d argue that while we need to fund that content, we can’t keep doing it this way no matter what. Give me dumb (i.e. privacy friendly, non-tracking) ads any day to pay for content. Or, if it helps keep the content free (both financially as well as free of tracking) for others then I think we should also consider paying for more news if we’re in any financial position to do so.
(What we shouldn’t do is just pay for our own privacy and let the rest still be spied on, that’s not really a desirable option. But even that doesn’t currently exist: If you pay for a subscription you’ll still be tracked just like everyone else, only with some other values in your profile, like “has demonstrated willingness to spend money on online content”.)
So, let’s recognize this category error for what it is. We should never repeat these statement again that ads pay for content; they do not. (Digital ad spend goes up and up but over the last 15 years or so newspaper revenue through digital ads have stayed pretty flat, and print collapsed.) Ads online today are almost completely tracking ads, and those are just surveillance infrastructure, period.
It’s surveillance with a lot of negative impact and some positive side effects. That’s not good enough. So let’s start from there, and build on that, and figure out better models.