Hi friends 👋 ,

Two years ago, I wrote an article called The Final Adpocalypse about the mild concern I had that online advertising might be a bubble.

In the time since publishing that article, a lot has happened within the tech industry and the broader economy. As I've continued to read, watch and learn, I've become more convinced that the bubble is leaking and may soon pop.

So, in this post, I want to explain three things:

1. Why programmatic advertising is a bubble.
2. How Big Tech is preparing for the pop.
3. What it means for the web.

I apologize. This article is long. If it's any consolation, it used to be much longer. Let's get started...

The Subprime Attention Crisis

Programmatic advertising is an economic behemoth. Last year, marketers spent ~170 billion dollars on online ads, more than they did on print and TV combined. Why? Because tech companies like Google have convinced them that targeted ad engines like theirs cause greater conversions than traditional, mass advertisement.

But is that true?

That's what Tim Hwang, previously the global public policy lead for artificial intelligence and machine learning at Google, set out to answer with his short book - or long research paper depending on how you want to think about it - The Subprime Attention Crisis.

In it, he makes a compelling comparison between targeted advertising and the subprime mortgage bubble of 2008. Hwang outlines two major issues with the targeted ad ecosystem that seem to be inflating its value: opacity and fraud.

No one knows anything.

"The tidal wave of data that has accompanied the development of online advertising provides only an illusion of greater transparency." -Tim Hwang

Programmatic advertising is a nearly opaque system. Most of the targeted ad market is an automated auction run in a black box by computers that talk faster than even the fastest human auctioneer. This means that both advertisers and platforms like YouTube, Facebook, and Google have no idea really where ads are placed once they're sold.

This opacity in the system is what caused the adpocalypses of 2017 and 2019. Brands like Coca-Cola got wind that their logo was showing up next to content from Nazi sympathizers and pulled their advertising until platforms could "fix the issue". This problem, known as "brand safety", has become a massive technical hurdle for platforms, causing stricter content guidelines and expensive manual review processes.

And it should be noted that this is a problem unique to targeted advertising. With traditional advertising, marketers pay to be next to certain content that is controlled and created by the networks and publishers they're spending with. They know where their ads are going and so do those that run the ads. Online, no one knows. Everyone is just trusting that Google and Facebook Meta are doing what they say: showing ads to consumers. Which leads to Hwang's second point.

It's bots all the way down.

Because no one knows where these ads are showing up and who is seeing them, it begs the question: what if no one is seeing them at all? According to a study done by Hootsuite, ~50% of internet users globally use an ad blocker. That would mean that 50% of "impressions" that marketers pay Google and Meta for never actually happen*.

But there is an even greater threat to the validity of data provided by platforms: click fraud. For those that don't know, between 45%-65% of all internet traffic comes from bots, not humans. Some of these bots are known as "good bots": web crawlers and scrapers that index the internet for search engines or data aggregation sites.

But most of them are "bad bots". These are programs that visit websites with the sole purpose of propping up impressions and click through rates on ads for website publishers. This hush hush, underground practice is extremely common. Hiring a "click farm" is as easy as going to a website like BuyViewsLikes.com and paying as low as $3 for 1,000 views.

And website publishers do this way more often than you probably think. Not because they're vain, but because it can be profitable. If I ran a single ad on this website using Google AdSense - the most popular public ad exchange in the world - the average amount I can expect to be paid by advertisers through Google for those 1,000 impressions on that one ad is $3.12. That means, if I ran multiple ads on my website and just paid for traffic, I could theoretically print money.

In practice, this scheme is harder to pull off than that for a variety of reasons, but it is possible, legal, and people do it. And advertisers know this. There are entire industries built around helping advertisers protect themselves - and their ad budgets - from click fraud.

All this only scratches the surface of the problems outlined in Hwang's book. But it's all just to say there is substantial evidence that programmatic advertising is subprime, an overvalued market, a bubble.

But they're still dancing.

So, if all this is true and programmatic advertising is so grossly overvalued, why are people still spending hundreds of billions of dollars on ads? Why hasn't the bubble burst already?

Bubbles don't burst unless there is a crisis of confidence. And right now, marketers, advertisers, and tech companies exist in a state of what social psychologist Johnathan Haight calls, "structural stupidity": an institution where every actor is disincentivized from posing any question that may threaten its dominant worldview. Over time, this makes the institution "blind and stupid".

This frame of mind is eerily similar to that of bankers in 2008. There was such a deep financial incentive to continue packaging and selling bogus mortgage backed securities that they looked the other way - despite knowing that these mortgages were likely to default. The famous quote from Chuck Prince, the CEO of Citigroup at that time, says it all:

"When the music stops, in terms of liquidity, things will be complicated. But as long as the music is playing, you’ve got to get up and dance. We’re still dancing."

Well, I think the band just started playing the last song for the programmatic display advertising industry, and it's called...

The Post Pandemic Panic.

Big tech has taken a big hit this last year. Companies like Meta, Google, Snap, Netflix and Spotify have had their stock prices slashed almost universally in half. Why? Well, if we look at all of their PR statements, they all cite "worsening macroeconomic conditions" for their declining profits*.

This isn't untrue. We are in a general economic downturn caused by the pandemic, rising inflation, and the war in Ukraine*. And in a recession, people usually spend less money. If people spend less money, brands usually don't spend as much on advertising. If brands don't spend as much on advertising, publishers (like internet platforms) make less money.

This story makes sense on the surface. But when you dig deeper, it's not so simple. Tech has fallen farther and faster than other industries in this changing economic climate. Some, including myself, attributed this to a correction of valuation similar to what we experienced during the DotCom Crash of the late 90s.

However, I think there's more to this story. Analysts like Tom Forte at D.A. Davidson are predicting a "perfect storm for digital advertising" caused by changes in public sentiment, new privacy policies, and the growing realization that the Emperor of the Internet, programmatic advertising, may have no clothes.

But to fully understand this "perfect storm", we need to go back to everyone's favorite year...

Big Tech's Bashful Boom of 2020

2020 was a wild ride. A pandemic, a presidential election, and what tech journalist Shira Ovide called a, "peculiar economy that benefited some people and industries...even as it battered others"*. The largest beneficiaries by far: big tech.

As the rest of the economy was being racked by "uncertain and unprecedented times", companies like Apple, Amazon, Google, and Meta made out like bandits. Schools bought iPads, e-commerce became commerce, and with the increase in time spent on social media, advertisers bought a whole lot of ads. The combined revenue of these companies went from billions to TRILLIONS*.

In a time where most people were economically strained, these companies were flooded with cash. So much that Apple bought back 90 billion dollars in stock*, Amazon spent tens of billions on warehouse and data center infrastructure*, and Google went on an acquisition spree*.

All of this was great for their bottom line and their stock prices - which on average grew by over 50% between March of 2020 and March of 2021. However, their dramatic success in the midst of a crisis put them in the spotlight. People who hadn't payed attention to the business models behind these companies in the past were suddenly asking, "Wait, why are these companies so rich, and why does it feel like it's at my expense?"*

At the same time, Netflix released The Social Dilemma: an exposé into the tactics and techniques companies like Google and Meta use to control user attention. With interviews from top technical leadership at companies like Twitter, Instagram, and Facebook, the film brought to light the harmful externalities engagement algorithms have on society in service of programmatic advertising.

All of this reintroduced national conversations about the effects of Big Tech on politics and national security that had gone dormant since the Cambridge Analytica scandal of 2018. Advocacy for a more private, secure and democratic internet was becoming mainstream.

And with a Democratic president elect, the threat of regulation that would limit the ability of these companies to track users across the internet was on the horizon. But one member of the big tech oligarchy saw this writing on the wall and decided to take matters into their own hands - by stabbing everyone else in the back and running away with the bag.

Apple's Pivot Towards Privacy

In June of 2021, at Apple's Word Wide Developer's Conference (WWDC), Tim Cook, Apple's CEO, made a big announcement: the company was going all-in on privacy. They announced a slew of features to macOS and iOS, "designed to minimize how much of your data we [Apple] - or anyone else - can access." And one feature - a subtle little permissions toggle gone largely unnoticed by users - has caused massive upheaval to the programmatic ad industry: App Track Transparency (ATT).

ATT is a feature on iPhones that now requires developers "to get your permission before tracking your activity across other companies’ apps and websites for ads or data brokers."* Now, I want to callout something that's hidden behind this non-chalant marketing copy: the fact that, before this feature, companies were able to track you when you weren't using their app.

Using what's called the Identifier For Advertisers (IDFA), trackers were able to recognize your device and spy on your activity across nearly every app on your phone - without you ever knowing about it. The data collected from this activity log was largely how companies like Google and Meta could build detailed profiles about you and provide those to advertisers for targeted ad campaigns.

These profiles are the core value proposition of their businesses. This data is their product. And with one tiny toggle, Apple shut off this collection from every iPhone user, 55% of the mobile market in the U.S. And with this announcement being so public, the advertisers on platforms like Facebook and Google know that these companies can't pretend to provide the same value they once did.

So, why did Apple do this? Well, one answer would be to believe the marketing materials: "Privacy is a fundamental human right", "Privacy. That's Apple.", "Our apps mind their business. Not yours." But digging a little deeper we see more of the story.

ATT only applies to apps that try to track you across apps. That means they can still collect what's known as "first-party" data. For example, ATT does nothing to stop Facebook and Instagram from tracking what you do inside of the app and then cross-referencing that information between their platforms to create a profile of your behavior and interests. And for Apple, a company with a growing suite of applications and a growing advertising business, this new policy was a great move for them. It bolstered their brand and bottom line while simultaneously hurting their competitors.

And similar to how Apple saw the writing on the wall that privacy was becoming more important to the public, Facebook saw that policy changes by companies and lawmakers could decimate their targeted ads business. Which finally leads us to Marky Mark's unforgettable announcement of...

The Metaverse

In October of 2021, Facebook officially changed the name of their company to Meta. With it, they announced sweeping changes to the mission of the company and vision for the future of the internet. They were betting big on virtual and augmented reality - believing it to be "the successor to the mobile internet".

Now, when this move was announced, it seemed totally out of left field. Why was Meta making such a dramatic change to their business? Yes, they had purchased Oculus in 2014 and had been making gaming headsets like the Quest for a few years now; but people had seen this move similar to Microsoft making the Xbox - a way of diversifying revenue by selling a gaming console.

With the VR industry still fraught with technical issues and seemingly perpetually stuck in the early parts of the adoption curve for decades, no one could understand why Meta - a company most known for making a shitty website covered in ads - would bet the farm on becoming a hardware company.

Well, I hope you can see why. Meta wants to be Apple - not beholden to them. Apple has been the dominate force in the age of mobile computing. And this power comes from them controlling the hardware and operating systems that half of the U.S. uses to access the internet.

We're already seeing what Apple's ATT policy has done to Meta's bottom line - costing an estimated 12 billion dollar hit to the company's revenue in just this first year *. So, to keep his business alive, the Zuck needs to own the next era of hardware and software. And it's not going well.

The Meta Quest Pro, Meta's flagship headset, has absolutely flopped. The software is janky at best and useless at worst. But more than just not being a great product, Meta just isn't a brand people trust to bring on this future computational platform. I think people can smell the desperation of the company behind the stench of the headset itself.

It doesn't feel like Meta is doing any of this because they really want to make great VR experiences. It feels like they're doing it because if they don't, their ad business is dead.

Other companies haven't been hit as hard by ATT and an end to third-party tracking. As such, they've taken the obvious path to make up for lost revenue: doing what a normal business does and asking you for money in exchange for their services. In fact, in the months after Apple announced ATT, we experienced what I call The Summer of Subscriptions.

In the 5 month period after Apple's launch of ATT, nearly every platform that relies on programmatic advertising launched some sort of paid subscription or change to their existing subscription products.

Twitter launched Twitter Blue and Snapchat launched Snapchat+; both products give users access to additional features within the apps. Google also raised their pricing for Workspace products like Drive, Photos, and Gmail - forcing users to upgrade or risk losing their data*.

All of these changes show that these platforms are becoming increasingly aware of their reliance on programmatic advertising and the shaky ground the market rests on. It also shows how deeply intertwined the experience of the internet is with this industry. Which begs the question: without the "free and open web" being paid for by programmatic advertising, can it even exist?

The Future After The Fallout

This is the part of the post where I'm supposed to tell you what's going to happen next. Well, I'm sorry to disappoint you, but I don't find predicting the future to be particularly useful.

There is no denying that the fall of programmatic advertising represents an inflection point in the internet. And this shift in the economic engine that drives the most revolutionary communications technology since the invention of writing leaves massive opportunity for technological, social, and political change.

However, if the past two years have taught us anything, it's that no knows what's going to happen next. So, for me to declare the facts of the future would be dishonest. It would leave you with a false sense of security or distress.

So, instead I invite you to consider what's possible rather than what's probable. Because the weakening of the status quo doesn't mean things are destined to get better - only that it's easier for us, the digital citizens of the internet, to make change.

Air is being let out of the programmatic advertising bubble. The winds are picking up. And if they continue, I believe they could turn the tides away from the financialized and capitalistic sea the internet currently sits in. But it's up to us to be ready to take the helm, weather the storm, and sail towards more just, cooperative and equitable waters.

And we need all hands on deck.

For my part, I'm committed to continue writing essays like this one. Pieces that help me (and hopefully you) unravel the web so we can weave it back better. To that end, I'd love to know what this article made you think of. If you want to continue this conversation, you can send me an email at contact@drewis.cool.

Until next time.