Welcome to the finale of my three-part summer field trip into the world of online advertising. Part one introduced our main characters. Part two got dark, with publishers and middlemen struggling as the Internet decayed. Today we wrap it up as the stakes get bigger with the arrival of enormous consumer platforms. Finally, the people will have their say. Is a happy ending even possible for our intrepid heroes? Read on…
Platforms are even bigger than you think they are
It’s hard to understand how enormous modern Internet platforms are if you haven’t worked inside of one. It’s effectively impossible to learn about giant platforms from observation. You need to use statistical analysis to look at things like prevalence, variation, and so forth. Unfortunately, human brains are not good at dealing with huge numbers or probabilities. Maybe some real-world references will help.
When I left Google Ads at the end of 2018, Google Ad Manager was delivering more online ads per day than the total miles driven daily by every motor vehicle in the world (40 billion). Facebook has as many active users as the estimated total number of eggs eaten worldwide per day (3 billion). One bad car crash or a dozen rotten eggs is not evidence that automobiles and hens are inherently unsafe.
The same holds for company “decisions.” The media doesn’t help when they personalize platforms like they’re kaiju destroying a major city. “Apple announced,” “Google launched,” “Meta apologized,” etc. In reality, company announcements are often just the chaotic side effect of hundreds of thousands of people trying to get promoted, survive internal politics, and avoid getting laid off. Amazon employs as many people as the population of Trinidad and Tobago (1.5 million). If some jerk from Trinidad steals your girlfriend, it’s not proof that the Trinidadian government has an organized plan to ruin your love life.
Ad networks are a side business
The ad networks that platforms like Google, Facebook, and Amazon operate are enormous compared to any one publisher, even the biggest. Disney reported $7.4 billion in ad revenue in 2024, but Google’s network business had that much revenue just in Q2 2025 alone. The catch is that being big is relative, and at platform scale, even a giant ad network is more of a nice-to-have. To understand why, look at net revenue.
That $7.4 billion Google earned on the network? As a middleman, they only keep around 32% of that, and pay the rest out to their publisher partners. The resulting $2.4 billion is still huge, but is dwarfed by the “Search and other” category, which took in over $54 billion in the same quarter. If you’re in charge of Google’s ad business, the math implies the network should only get about 5% of your attention and resources. [Meta doesn’t break out revenue for the Facebook Audience Network, but similar calculations would apply.]
Net revenue also explains why Amazon is bucking this trend and putting a lot of investment into its ad business. Their core business, online retail, has very low net margins. Their North American profit margin in Q2 2025 was 7.5% (and even lower, 4.1%, internationally). Even a 20% or 30% take rate as an advertising middleman looks pretty good in comparison. Amazon’s ads business is growing fast and was over $15 billion in Q2 2025. [Note this is probably a mix of network and first-party advertising.]
It’s not you, it’s me: differential growth is destiny
A common sensation of working with a big platform over many years is loss. One day you are great partners, the next you can’t get a phone call returned. The platform didn’t pull a fast one on you, and they aren’t just fickle. Often the reason is differential growth. When one part of a business grows faster than others, incentives turn upside-down. Here are three examples.
One: customers. Early in my DoubleClick career, the Washington Post was a top ten customer, and my (small) team and I spent a lot of time with them. A few years later, the Post staff started complaining about feeling ignored, even though their business had grown. They weren’t wrong. Other customers had grown much faster during the same time, and now the Post wasn’t even in the top twenty. We still valued the relationship, but they would never have the same influence again.
Two: businesses. In 2008, Google’s display ad network was small, so adding DoubleClick’s $300 million in platform revenue was a big deal, and the platforms got a lot of executive attention and investment. The display ads business grew quickly, hitting a $2.5 billion annual run rate in October 2010 and $5 billion in January 2012. The vast majority that growth was from the network, so it was logical for attention and investment to go there, which would only make it loom even larger over the smaller, slower-growing platform business as the years went by.
Three: paradigms. As mobile apps took over the world and desktop web usage stopped growing, our fastest-growing customers started to see us differently. Snapchat’s ad business started out on the Google platform, but within a couple of years they’d completely outgrown us and had to build their own stack. I’ll never forget sitting in Snap’s Venice Beach office while their ads team patiently explained all the ways that we just couldn’t handle their scale. It still stings. [Washington Post team, this is your schadenfreude moment.]
People will pay for value
With all of the problems with the online ads experience I covered last week, it’s no surprise that people have voted with their feet by installing ad blockers, opting out of tracking, and favoring mobile apps over ad-saturated web pages. A decade or so ago, a newspaper executive famously said that their competition wasn’t other newpapers, it was Angry Birds. Well, Angry Birds won. [Cue superheroes turning to dust and drifting away in a gentle wind. Somber music. Credits.]
Fortunately, that isn’t the end of the story. Smart publishers have found a viable path forward by going back to the basics and building products that earn user attention, instead of exploiting it.
I gave the New York Times a hard time last week, but since 2015, Meredith Kopit Levien has led a remarkable turnaround built on successful subscription products. NYT Games and Cooking, plus the Athletic for sports, have thrived, and almost 80% of their online revenue now comes from subscriptions.
Readers will also pay for useful, unique reporting, especially if they aren’t being overcharged to cover the salaries of highly-paid executives. Publications like Talking Points Memo (politics), 404 Media (security and abuse), and The Information (tech business) have all done well enough to support small teams of journalists. Even individual reporters can thrive with newsletter subscriptions.
When I started Platformer, one of the thoughts I had was like, “All I need is a thousand people to pay me a hundred bucks a year, and I have a pretty good job; 2,000 people is an amazing job; 3,000 people, that’s more than anyone will ever pay me.” And the internet is still big enough, and there are enough amazing readers out there that they will create that job for someone.”
Online advertising still has a role to play in this new world, but as what it probably always should have been — an add-on. A publication with an authentic, trusted relationship with their audience can make some money by respectfully asking them to share a bit of attention with a partner. Think of it as a break for these important messages, but not the whole show.
Post-credits scene
There is still one big problem that the new world of subscription media doesn’t solve. Even if all the best stuff survives by going behind paywalls, the Internet will still be flooded with free junk. The path of least resistance (and the only affordable path for many) will be consuming really low-quality information. Even worse, this could end up being biased information that is only free because someone with an agenda is paying for it.
For what it’s worth, this is not a new problem. It also foreshadows questions of online identity that I’ll be writing about this fall. Stay tuned, true believers.
Besides, as the vilest Writer has his Readers, so the greatest Liar has his Believers; and it often happens, that if a Lie be believ’d only for an Hour, it has done its Work, and there is no farther occasion for it. Falsehood flies, and the Truth comes limping after it; so that when Men come to be undeceiv’d, it is too late; the Jest is over, and the Tale has had its Effect.
Ideas? Feedback? Criticism? I want to hear it, because I am sure that I am going to get a lot of things wrong along the way. I will share what I learn with the community as we go. Reach out any time at [email protected].
