A few days ago I wrote about the MAMMA paper from Randall Rothenberg and Joseph Frelinghuysen, which stands for Marketing, Advertising, Media, Measurement, Analytics, because apparently we needed another acronym and "MAMA" was taken by Korean pop music. The response was bigger than I expected, and the questions all landed in roughly the same place: okay smart guy, so what do we actually do about it?

Fine. Here is the follow-up. Two parts. This one is free, because I am generous and because I want you to feel the FOMO. The second part, with the operator playbook, the structural risk matrix, the regulatory timeline, and what the smart brands are quietly already doing while everyone else is still printing out the chart, sits behind the AdotaT+ paywall.

Now. Here is what you actually need to understand before you go any further.

The MAMMA paper is not really about measurement. I know, I know, it has "Measurement" right there in the name, twice if you count "Analytics," and the entire surface argument is the same lullaby we have been hearing since the Bush administration: ad measurement is fragmented, vendors are siloed, nobody trusts anybody's numbers, and the industry needs an integrated AI-driven platform layer, the now-mandatory Bloomberg-for-marketing analogy, to stitch the whole Frankenstein back together. Sure. Cool. That is the obvious read, and it is the read most trade press will give you this week, complete with a LinkedIn carousel and a webinar sponsored by a vendor that just rebranded its dashboard.

The non-obvious read, and the reason this paper is going to dominate Q2 boardroom conversations whether or not anyone has actually finished reading it, is that Rothenberg and Frelinghuysen have quietly conceded something the industry has spent thirty years pretending is not true:

measurement in advertising is political, not scientific.

It is a negotiated border between buyers and sellers, dressed up in the lab coat of methodology. That is why no "currency" ever sticks. That is why every JIC, every coalition, every "industry standard" eventually fractures into a slap fight on stage at Cannes. The numbers were never the point. The leverage was. Always has been. Anyone who tells you otherwise is selling you a panel.

Once you accept that, the rest of the paper reads completely differently, and the risks get sharper than the authors are willing to say out loud. Three things to hold in your head before you read the deep dive.

One. The "service-as-software" model is a Goodhart's Law accelerator. The paper's big idea, AI agents running planning, buying, optimization and measurement as one beautiful closed loop, is genuinely useful. It is also a doomsday device for metrics. When buyer-side AI and seller-side AI optimize against the same numbers at machine speed, with no human in the loop, the metrics collapse into meaninglessness faster than anyone can audit them. Click fraud took twenty years to become a billion-dollar industry. Agentic gaming could do it in eighteen months, and the people who built it will be on a panel explaining why it is actually a feature.

Two. The agency layoffs are not a side effect of the AI transition. They are load-bearing to it. WPP has cut more than 15,000 jobs in two years. Omnicom and IPG are stacking post-merger reductions on top like a particularly grim Jenga tower. And the roles being eliminated first are exactly the ones you need most: the senior measurement scientists, the planning leads, the twenty-year veterans who can look at an AI-generated optimization recommendation and say "that is nonsense, here is why, sit down."

Without them, MAMMA does not become Bloomberg. It becomes an industrialized self-deception engine. Garbage in, garbage out, at speed, with absolutely nobody senior enough left in the room to pull the fire alarm. The interns will be very polite about it.

Three. Look at who is missing. Google and Meta are treated like weather systems in this paper. Unavoidable. Uncontestable. Not really part of the conversation, you just dress for them. That is the tell.

MAMMA is implicitly a proposal for how the rest of the industry organizes itself around the walled gardens, not against them. The Trade Desk's Kokai, Smartly, Adobe, Mediaocean: those are the positioned winners. And the brands that do not build their own in-house "spine," the way P&G and JPMorgan already have, end up renting their measurement from operators who have every commercial incentive to grade their own homework. Spoiler: they give themselves an A.

That last part is the bit almost everyone is skipping past. So obviously that is where I spend most of Part Two.

In Part Two… I go through the operator playbook by role, what CMOs, agencies and vendors each need to do in the next two quarters, the regulatory cliff coming August 2026 with the EU AI Act, the data unification audit every brand should be running right now instead of buying another dashboard, and the specific traps buried inside the eight-category chart that vendors are absolutely going to weaponize in their Q2 pitches. The chart is the slide. The slide is the sales tool. The sales tool is the trap. I will show you how it works.

If you read one paper this year about where the ad industry is actually going, make it MAMMA. If you read one breakdown of what it actually means for the people who have to operate inside it, you already know where to click.

Part Two is for AdotaT+ readers.

Inside, I go through the operator playbook by role: what CMOs, agencies and vendors each need to do in the next two quarters. The regulatory cliff coming August 2026 with the EU AI Act. The data unification audit every brand should be running right now instead of buying another dashboard. And the specific traps buried inside the eight-category chart that vendors are absolutely going to weaponize in their Q2 pitches. The chart is the slide. The slide is the sales tool. The sales tool is the trap. I show you how it works.

If you read one paper this year about where the ad industry is actually going, make it MAMMA.

If you read one breakdown of what it actually means for the people who have to operate inside it, [click here].