It's 4 AM. I have a migraine. I am not doing that thing where a writer says "I couldn't sleep thinking about this" as a rhetorical device.

I am sitting in the dark, my skull is splitting open, and I am furious about programmatic advertising.

And then at 5:30 AM someone I trust sent me an email that made the migraine worse. But we'll get to that.

First, the question that's going to frame this entire series. The question my brain keeps circling back to every time I look at the data:

Is this illegal?

Not "is this inefficient." Not "does this industry need to do better." Is what is happening in programmatic advertising, at scale, across the supply chain, actually criminal?

I'm trained in computer forensics and intelligence analysis. That's my background. When I pull apart supply chains and look at data flows in programmatic, I don't see "market friction." I see patterns. And the patterns look like organized fraud. I think about RICO. I think about what a federal prosecutor would do with subpoena power and the log-level data I've seen.

That's the lens for this series. Not "how do we optimize." Is this legal.

And because of what I've been seeing, what I've been sent, and what I was told this morning... I am going to start treating parts of this ecosystem the way my training tells me I should. Like a possible criminal enterprise. I will be taking notes. I will be investigating. I will be reporting.

For those who read this newsletter, you should know who else reads it. The FTC. The SEC. The NYC Department of Consumer Affairs. California Consumer Affairs. The California Attorney General's office. The New York State Attorney General's office. That last one sends me press releases now. Hi, everyone.

This isn't a threat. It's a promise to continue my career and my ethics the way they've always been. Reporting the truth.

Let's start.

Here's something that sounds like a contradiction but isn't: advertising works and advertising is broken. Both are true simultaneously.

The research is clear. Paid media delivers positive profit ROI. CTV works. Search works. Online video works. When a well-run campaign reaches real humans with a relevant message in a quality environment, advertising does what it's supposed to do.

But that describes a minority of what's actually happening in programmatic. For the vast majority of buyers, "programmatic" means spraying money into an opaque supply chain and hoping. Display CTRs have collapsed to under 0.1 to 0.3 percent. A shrinking minority of users account for the majority of conversions. And a meaningful chunk of every buy is going to bots, MFA sites, and fee layers that never deliver a single real impression.

When someone says "digital advertising is more effective than ever," what they mean is: the best campaigns are more effective than ever. Everyone else is lighting money on fire and watching a dashboard tell them it's working.

2009 U.S. online ad revenue: $22.7 billion. The whole market.

2024: approximately $259 billion. Over 10x in fifteen years. Globally, from the mid-$50 billion range to roughly $440 to $450 billion, heading toward $700 to $800 billion.

Programmatic now accounts for roughly 90% of all digital display dollars worldwide. Over $100 billion annually in the U.S. alone. In 2009 it was single-digit percentages.

And fraud? Fraud scaled faster than the legitimate market.

In 2009, click fraud audits flagged 13 to 20 percent of PPC clicks as fraudulent. But the pie was small. A few billion in waste.

Now? Global digital ad fraud is estimated at $100 to $120 billion per year. Some methodologies push it toward $250 billion. Projections show it hitting $172 billion by 2028, roughly 20 to 23 percent of all global digital ad spend.

One in five dollars. Gone. Not "wasted on low-performing inventory." Gone. Stolen. Siphoned into an extraction layer of bots, MFA sites, and supply chain grift that has become an industry within the industry, roughly the GDP of a medium-sized country, feeding on top of legitimate advertising.

The pricing tells the story if you know how to read it.

In 2009, open-web banner CPMs were $1 to $3. Today, cheap banners are about the same in real terms. But nobody's growth story is about banners.

Programmatic CTV: $40 to $75 CPMs. Programmatic video: $15 to $35. That's where the revenue growth is, and that's where the fraud incentives are, because faking a $60 CPM impression is a lot more interesting than faking a $2 one.

The industry points to rising CPMs and says "the market values this." But rising CPMs in a market where 20 percent of impressions are fraudulent doesn't mean value went up. It means the price of nothing went up. You're buying air at premium prices and the receipt looks fantastic.

So. That email.

Someone I trust, someone who has been in this industry a long time and has seen the pipes from the inside, emailed me this morning. Yeah, he's up early too. Maybe this industry gives everyone migraines.

What he told me is something I've suspected but hadn't been able to confirm from a source I consider reliable: some agencies are participating in this laundering. And benefiting from it.

Let me be very clear about what I'm saying and what I'm not saying.

Not all agencies. Not everyone. There are good people at agencies doing good work for their clients. I know some of them. Some of them read this newsletter. They're as frustrated as I am.

But some agencies are benefiting from the opacity of the programmatic supply chain in ways that range from ethically questionable to potentially illegal. And the spectrum matters here, because this is the question at the heart of this series.

Here's how it can work. And I'm being careful with my words because, again, my newsletter subscribers include people with regulatory authority.

Arbitrage. An agency buys inventory at one price and bills the client at a higher price, pocketing the spread. Sometimes this is disclosed and contractually allowed. Sometimes it isn't. When it isn't disclosed, when a client is paying for "transparent" media buying and the agency is running undisclosed margin on top? Where exactly is the line between "business practice" and "fraud"?

Kickbacks and rebates. The ANA's own landmark studies have documented undisclosed rebates flowing from media sellers to agencies. Money that should be reducing client costs or being disclosed as agency revenue, instead quietly flowing into agency P&Ls. Some of this has been restructured since the ANA reports. Some of it has just gotten more creative about hiding. At what dollar amount and level of concealment does an undisclosed rebate become something a prosecutor would be interested in?

Looking the other way on supply quality. This is the subtle one. An agency knows, or should know, that a significant portion of the programmatic inventory it's buying for clients is low-quality, MFA, or outright fraudulent. But the volume looks good on reports. The CPMs are cheap, which makes the agency look efficient. The client sees big reach numbers and green dashboards. Nobody asks hard questions because the answers would mean less spend flowing through the agency's systems, which means less revenue. Is willful ignorance a defense? How willful does the ignorance have to be before it becomes complicity?

I don't have all the answers yet. But I know the questions, and I know they're the right questions. And I know that "this is just how the industry works" is not a legal defense. It has never been a legal defense. "Everyone does it" didn't work for Enron. It didn't work for the mortgage brokers. It won't work here.

Let me map the layers of complicity, because complicity has tiers.

Tier One: They know. MFA publishers building pages for arbitrage, not humans. SSPs seeing the log-level data, the insane refresh rates, the impossible CTR patterns, and keeping those sites live because they generate fees. Fraud-as-a-service operators running bot farms and click injection networks. These are overt choices.

Tier Two: They choose not to know. Mid-tier networks and platforms seeing metrics that no honest inventory could deliver at those prices, but accepting it because it hits goals. Agencies (some, not all) hiding behind "we pass MRC filters" even though those filters miss a huge share of sophisticated fraud. Strategic ignorance is still a choice.

Tier Three: They're in the dark. Most advertisers don't have log-level transparency or the staff to analyze it. They see dashboards and attribution outputs and think the machine is working. It's not working. It's performing.

The gap between Tier One and Tier Three is where the money disappears. And somebody is pocketing it at every layer.

I want to be fair. CTV actually works. When it's real inventory, reaching real humans on real screens, it delivers. The brand lift data is strong.

But people like Dr. Augustine Fou claim CTV fraud is far worse than the industry admits. I take that seriously enough that I'm going to dig into it separately. The question is whether the analytical methodologies are sound. Are these analysts catching real fraud, or are we making claims based on limited data? That column is coming. Not today.

Here's what I know needs to happen.

Stop buying open exchange garbage without curation. If you don't control the domains and apps you buy on, you are the mark.

Demand log-level data. If your partners won't show you the raw auction data, the answer to "why not" is never good.

Stop rewarding reach. The obsession with cheap CPMs and massive impression counts created the MFA ecosystem. You get what you incentivize. We incentivized garbage at scale.

Audit your agency. Ask hard questions about undisclosed margins, rebates, and supply quality. If the answers are vague, they're vague for a reason.

Somebody with subpoena power needs to look at this. $100 billion a year, at minimum. Multiple actors, coordinated activity, sustained fraud across state and international lines. This deserves prosecutorial scrutiny it has not received.

We built a $450 billion global industry on the promise that technology would make advertising smarter and more accountable. For a small number of sophisticated players, it did.

For everyone else, we built the most sophisticated grift in the history of commerce and called it "innovation." A hundred billion dollars a year, at minimum, disappearing into an ecosystem of fraud, arbitrage, and willful ignorance, with participants at every level of the supply chain taking their cut.

I care. That's why I'm up at 4 AM with a migraine writing this. That's why my brain won't stop mapping supply chains like evidence boards. That's why I keep thinking RICO when I should be thinking sleep.

Starting today, I'm treating this the way my training says I should. I will be investigating. I will be documenting. I will be reporting. The FTC, the SEC, the state AGs, the consumer affairs offices... they know where to find me. I know where to find them.

This isn't a threat. It's a promise to keep doing what I've always done.

Reporting the truth.

Next in this series: How deep do the agency conflicts go? And: What would a prosecutor actually need to bring a case?