Let me tell you about the most expensive lie in marketing.

It doesn't come from a shady vendor. It doesn't arrive in a cold email at midnight promising 10x ROAS if you just sign right here. It comes from your own dashboard. It's sitting right there, refreshing in real time, looking authoritative and precise and completely, catastrophically wrong.

It's called last-touch attribution. And if your marketing program runs on it, which statistically speaking it almost certainly does, you might as well be making billion-dollar business decisions based on which intern happened to be standing closest to the printer when the sale came in.

I sat down with Skye Frontier of Incremental, who has spent the better part of his career doing something genuinely rare in this industry: telling marketers things they absolutely do not want to hear. He is, depending on your disposition, either the most useful person in any room he enters or the most annoying. Possibly both. I found him completely, bracingly refreshing.

Skye described last-touch attribution to me as "the high fructose corn syrup of measurement. We all know it's bad for us, but it's everywhere. It's ubiquitous and it's much of a hassle to avoid. So we live with it and we have to get off of it."

That is not a throwaway line. That is a precise and damning diagnosis of an entire industry's measurement infrastructure. The U.S. digital media market is somewhere around $300 billion. And a significant chunk of that, Skye estimates roughly half, is being optimized toward metrics that have no meaningful connection to actual business outcomes.

None. Not a little connection. Not an imperfect-but-directionally-useful connection. No connection whatsoever.

Here is the thing about last-touch attribution that should keep every CMO, every CFO, and frankly every board member up at night. It works on a simple, seductive, completely fictional premise. Someone saw your ad. Someone clicked your ad. Someone bought your product. Therefore: your ad caused the purchase.

Correlation. Dressed up as causality. Given a dashboard. Given a budget. Given a promotion.

As Skye put it with the weary clarity of a man who has explained this at too many conferences to count, "there's no connection whatsoever between last-touch attribution and incremental sales or incremental impact." That is not his opinion. That is the conclusion of approximately twenty years of independent research into click-based attribution. Study after study. Same conclusion. Over and over again.

And the industry looked at all of that research and said: yes, but have you seen how easy the dashboard is to read?

Let's pause on that number. Twenty years.

In twenty years, we landed rovers on Mars. We sequenced the human genome. We put a supercomputer in everyone's pocket. We invented, hyped, crashed, and partially resurrected the entire crypto industry twice.

And in that same twenty years, the digital marketing industry looked at mountains of evidence that its foundational measurement approach was broken, shrugged collectively, and kept buying on last-click because it was easy.

Skye, who came up through the measurement side of this business, is not without sympathy for how we got here. "I was told really early on in my advertising career," he told me, "that media dollars follow easy." And he's right. Last-touch attribution is easy. It's available on every platform. It produces numbers that go up. It tells a story, however fictional, that everyone in the room can follow without a PhD.

And crucially: it almost always makes your advertising look fantastic.

Think about what that means structurally. The metric that's easiest to use also happens to be the metric most likely to flatter your results. That is not a coincidence. That is a system that has been optimized, whether intentionally or not, to make the people selling advertising look good, the people buying advertising feel good, and the people asking hard questions feel like they're being unreasonable.

Here is the part that genuinely fascinates me, and the question I put directly to Skye: why do obviously intelligent people, people with advanced degrees, real experience, and the demonstrated ability to build complex things, cling to objectively broken metrics with the devotion usually reserved for a toxic romantic relationship?

The answer is not stupidity. It is incentives.

"Last-touch attribution and ROAS often paint a pretty rosy picture of advertising's contribution to the business," Skye told me plainly. "And there's job security in that."

There it is. The thing nobody says out loud at the conference panel.

If your KPIs are built on last-touch ROAS, and last-touch ROAS always makes your work look good, then fixing the measurement system is not just an intellectual exercise. It is an existential threat to the narrative that justifies your budget, your team, your promotion, and possibly your entire department's reason for existing.

You are not clinging to a bad metric because you're dumb. You are clinging to it because the alternative, actually measuring whether your advertising is causing anything, might reveal that a meaningful portion of what you're doing is not working. And nobody gets promoted for discovering that.

I asked Skye if he'd ever staged an actual intervention. Pulled someone aside. Said: we both know this is not real, right?

"I've done that," he said. "I have certainly had a few too-honest conversations after a martini or two on the topic."

He has not, for the record, been invited to Google after hours. Make of that what you will.

So if last-touch attribution is the problem, what's the answer?

Skye's entire professional existence is built around incrementality, which sounds like jargon until he explains it, at which point it becomes almost insultingly obvious. I gave him fifteen seconds at the end of our conversation. No hedging. No jargon. His answer:

"It's the sales caused by advertising that you wouldn't get without advertising."

That's it. That's the whole thing. The question is not "did someone click my ad and then buy something?" The question is: would they have bought it anyway?

That distinction, between correlation and causation, between credit-claiming and actual impact, is worth, depending on your budget, tens or hundreds of millions of dollars a year. The entire field of causal inference, foundational in economics, biology, social science, basically every serious discipline except marketing, exists to answer exactly this question.

The marketing industry built a different tool instead. A worse tool. An easier tool. And then automated the whole thing so the algorithms could make the same mistakes faster and at greater scale.

"You're now automating and ingraining really bad decision making into algorithms and rules," Skye told me, with the specific grimness of someone who watches this happen daily. "That for me is the scary part."

It should be the scary part for all of us.

Then on Wednesday night, ADOTAT goes dark.

For those who don't know: I'm Pesach Lattin, yes, Pesach, as in Passover, as in my parents had very strong opinions, and Wednesday night begins the holiday of Pesach, the Jewish festival of freedom, during which I will be doing considerably more eating and considerably less writing about attribution models.

We'll be back Saturday night. Recharged. Fed. Still skeptical.

Chag Sameach to everyone celebrating. And to everyone else: use the time to finally look at your incrementality numbers.

Parts 2 and 3 of this series are available exclusively to ADOTAT+ subscribers. Here is what you are missing if you're not one:

Part 2: The $150 Billion Automated Mistake. The platforms themselves have hardwired last-touch thinking directly into the algorithms buying your media right now, in real time, while you read this sentence. This isn't a human making a bad choice anymore. This is bad decision-making running at machine speed, at machine scale, completely unsupervised. Skye walks through exactly how deep this goes, what it's actually costing you, and the one number that literally cannot lie to you. If you run any kind of paid media program, this part will make you want to call an emergency meeting.

Part 3: You're Renting Your Own Customers Back. Retail media is now a $50 billion industry. Every grocery chain, every big-box retailer, every company with a website has decided it's also a media company. And brands are writing the checks with a smile while knowing something is deeply wrong. Skye explains why retail media is structurally closer to a platypus than a media channel, who is actually winning this game, and why the transparency problem is even worse than you think. This one is going to make some very important people very uncomfortable.

Both drop this week. Both are for ADOTAT+ only.

[BECOME AN ADOTAT+ SUBSCRIBER]