Ad Tech in the Wilderness
Ad Tech in the Wilderness
The multi-layered multi-directional opacity of the ad tech industry makes it harder to fix
I write to you from a plane returning from San Francisco*, home of our tech overlords, where I spent two days presenting a bird’s eye view of the intersection of ad tech, publishers and marketing to Mozillians (who are not our tech overlords) and others.
I was very satisfied, especially because the group leveled a bunch of questions at me that I hadn’t thought deeply about but should have and also because I think at least some of what I had to say was new to them.
The workshop I presented was a restructured and expanded version of the two 2.5 hour presentations I gave at The New School titled “Ad Tech: Broken by Design”. I split the topic into four parts:
The history of publishers ignoring the opportunity of the web advertising space (and why) as they entered digital publication and then by allowing it to metastasize into bad tech and systems. Then how this created an opportunity for ad tech middlemen to come in and start stealing money away from publishers.
Why—given the framework of an ad space on publishing sites, which fails to integrate the advertising involved in any way—the current publisher consideration of ad tech creates room for significant fraud and just a few of the many ways that opportunity for fraud is realized.
The requirements for and use of user data that creates a vicious cycle that has platforms and algorithms creating and tracking more data and agencies and advertisers seeking ever more user information to the extent that the need for user data has become completely integral to the entirety of the modern ad tech stack.
How metadata, combined with an aggressive handling of potential Javascript code exploits and browser-level tools, can be used to potentially empower publishers to create a better ad tech space.
The reason I frame the discussion in this order and around these points is because I think it is pretty much impossible to understand how ad tech works or why it is so bad without a real overview of how we got up to this point both in terms of publisher and publishers-as-platform-clients history. If you were to dive deep into the ad tech space as it stands right now it would be impossible to understand why it works in the insane way it does. Something as basic and fundamental to ad tech as 'why are there five trackers on a single ad' or 'why does VPAID exist' doesn't make sense without dissecting the distrust between publishers and advertisers, the breakdown of agency fees, which itself increases agency fraud, the lack of transparency in the system that makes such fraud possible, the circle of vulturous ad tech providers that enable fraud to succeed, exchange incentives… and here we go down the well.
Depending on what slice of this crazy ad tech space you deal with, you are only going to encounter perhaps a quarter of the overall problem. And without the whole picture, it’s almost impossible to understand why we've ended up where we are now.
Most of the time, when I talk about these topics, I'm talking to the publishing side, people who have seen the publisher side of things but don’t know what has brought them to this stage. It was great to get very different feedback from people who encounter these problems from another angle. The browser side encounters the technical consequences of ad tech malarky—the tracking, the slow mess of network requests, the shitty code—without needing to deal with what the ad tech space does or doesn't do for publishers and advertisers.
Browser user performance and concerns create an entirely different set of worries and priorities of course and that manifested itself in some ways I hadn't thought about.
One big one: We don't think much about what will become of people who have turned off tracking but not blocked ads during the browser-led transition of the ad tech space to mostly first-party targeting. At least, I didn’t. But the first thing that came up after I talked about ad value’s link to user data was that people browsing third-party-cookie-free will spend that period of transition as the lowest-class-citizens. The first adopters of privacy will end up having the worst experience of all.
As long as ad tech still uses deep third-party data, users without those cookies will be seen as data-lite and, therefore, the lowest-value target for advertising, with the result of them getting bottom-of-the-bucket ads (and potentially a higher percent of malicious ads as a result). Not great. We talked over a few solutions, maybe more on that later.
Song break! Do you know someone at Waxwork Records who you can convince to put the Chopping Mall soundtrack on Spotify? It would make me so happy.
Another question came up when we were diving down into content fraud of all types: The impact of fake followers.
We were discussing different types of fraud when, during our conversation about follower fraud, I noted a study in which the following was stated, in regard to a Fiverr user generating fake social accounts:
“In other words, one user from Moldova has earned at least $1.5 million/year, which is orders of magnitude larger than $2,070, the GNI (Gross National Income) per capita of Moldova.”
One of the participants asked: even if we could somehow extend the reach of our laws to another country, should we penalize these people who have managed to make a living exploiting the platforms? It was something I hadn't really thought about from that angle but after some consideration I realized: No.
In the period immediately following the US Revolutionary War, the fledgling United States essentially jump-started its economy by violating international intellectual property agreements, particularly those of the British crown. In fact, the Copyright Act of 1790 extended the right to intellectual property only to American citizens creating a marketplace for reprints of mostly British works (see p320-324 of the linked PDF), including numerous textbooks, without paying royalties back to the original authors.
In the case of successful fake follower creators, it feels more like they are following our revolutionary-era example than anything else. The bounty of the powerful, in this case the effectiveness of social media platforms, can be harvested by less wealthy nations. No, I don’t think the penalties should fall on these users, but on the platforms and analytics tools who promise genuine engagement to advertisers with no capacity to deliver. At a fundamental level the platforms and ad tech understand what they are doing is unsustainable. The people who take advantage of the giant wealthy systems that have imposed its way into their lives, all of our lives, shouldn’t be the target of our efforts to better the web. After all, they’re just leveraging the space between the truth and the bullshit in the ad tech pitch.
I think the platforms realize this. As protests to the current methodology of ad tech grow louder, most of the world thinks about Facebook and Google and how to punish or regulate them. But it isn't just that. If ad tech's problems were quite so centralized they would be far easier to solve. Is Facebook a monopoly? Yes. It should be broken up. But the source of its monopoly-like power has little to do with its placement in the marketplace or primacy in the app store created by absorbing other platforms.
The power of Facebook is its monopoly on user data because it is given greater amounts and more accurate user data than any other platform and can directly turn that user data into money. Facebook's break up would do little to resolve this issue. The transaction and trade of user data has become built into the mechanisms of much of ad tech and it may still be able to connect that database to the theoretical post-breakup Baby Facebooks, even if those systems are legally separate.
What these companies fear then are not the regulatory power of governments (which they have proven the power to compromise anyway) on their moderation techniques nor the anti-trust lawyers (though they should be subject to such powers) but the revolutionary independence of user data. Like the early Americas, we sit on the precipice of a system that could potentially free us from imperial chains of ownership; our data and ourselves belong to Facebook and Google. Perhaps a law, perhaps a browser, but likely a combination of both could liberate that data. But, and this is important, for it to work consent isn’t the only thing we have to consider.
In ad tech, fake users seeing ads make as much money for most of the middlemen companies (including Facebook and Google) as real users. The value isn’t even the user though, it is their behavioral data. If we are going to regulate how big platforms operate and derive value, we have to start with the control of data, which is where the money is made. At the end of the day, Facebook, Google and the rest of them are networks for communication. They don’t need to own our data for that to happen. They don’t even need to own our data to make money showing us ads. We need a middle ground where we can use these platforms, make our data visible to whom we choose, but refuse to give control of it to any centralized power.
A version of Facebook where the revolutionaries operate without the oversight of the Facebook data stores is a one where Facebook becomes safe, far more effectively than any anti-trust action could make them. Right now the sun doesn't set on Facebook. But history shows that those types of Empires don't last. King Zuck can only reign for as long as we acknowledge the platform is a necessary repository for the power we give it. Not Google, not Facebook, no ad tech company should be a repository for that power because at a fundamental level it should reside with individuals. When we move to punish platforms (instead of users) by removing their most important assets—our data—from their grasp: So sets the sun.
*Ok, I started on the plane back, but all trips back from SF (and only SF, it doesn’t happen coming back from Seattle) kill me with various types of fatigue, abnormally bad jet lag, and/or disease, so I ended up spending most of the week recovering from existing in SF and then edited and finished off the newsletter at a later point. I was happy to visit and had a lot of fun but, honestly, how do people live there?
Interested in talking more about this newsletter? Join the Keybase.io team to chat about it: https://keybase.io/team/gneist.newsletter
Readings:
Progress: