Let me walk you through the week.

A media company figured out how to charge money for the moment you get up to pee. A measurement giant ran podcast episodes through a transcription engine until they smelled enough like display banners for a media buyer to touch them. The IAB Tech Lab shipped a protocol so the robots can close deals without a human in the loop, and buried in the announcement is a line promising that inventory will be based on real numbers, which, sure, glad we're putting that in writing now. A joint industry analysis found that AI slop beats real publishers on the industry's own quality scores. And a nonprofit watchdog asked a DSP for a seat, signed the contract, got a "woot woot" in an email, and then got dumped.

Five stories. One story. This business has gotten world-class at manufacturing inventory out of thin air, measuring it with instruments that cannot detect emptiness, automating the handshake so nobody has to put their name on it, and then explaining, with real feeling, that it has obligations to its supply partners when someone shows up with a flashlight.

Every one of these is a transparency story. Four of them have no idea.

Scott Rossman, VP of product marketing at WBD, says that pause ads are "creating new ad space" where none existed. That is the single most honest sentence anyone in ad tech has produced this quarter, and I would like it engraved somewhere. Take a screen doing absolutely nothing, drape a soft gradient over it, and congratulations, you have a CPM. WBD launched pause ads on HBO Max in 2022, announced shoppable ones at upfronts in May, and the pitch is that if you covet the shirt on the actor, you hit pause, walk into a storefront, and check out. If the exact shirt isn't available, they'll suggest a similar shirt. Commerce abhors a vacuum. So, apparently, does the vacuum.

Now the research, which is doing Herculean work here. 67% of Gen Z and millennials prefer a pause ad to a frozen screen, per DirecTV and Magna. Read that again. That is not consumer enthusiasm, that is a hostage statement. People prefer something to nothing. Astounding. The VAB says 51% "took an action," and the actions include saving information for later, which is the digital equivalent of nodding at someone in a hallway. Rossman also cites Disqo data showing over 200% lift in site visits versus benchmarks, a lovely number produced by a vendor about a format that vendor's clients sell. Hold that thought for four hundred words. The actual tell is in the plumbing: WBD sells this direct, because programmatic can't handle it, because the industry hasn't agreed on pixel counts. The Tech Lab has a proposal. Comments closed last month. We are selling the future of television and we are stuck on file sizes.

The podcast industry has spent ten years standing on a chair yelling that listening is huge and spending is tiny, as if this were a branding problem. Comscore stopped arguing and started translating. As of Wednesday it has yanked Spotify, SiriusXM, Triton Digital, Acast, and Libsyn into Proximic, using episode-level transcription to chop audio into the same pre-bid contextual segments buyers already use for web and TV. Proximic's Jessica Trainor says it brings audio "to parity" so advertisers feel more confident moving dollars over. Confident. Not better performing. The product being sold here is comfort, and I say that with zero snark, because comfort is what actually moves budgets and everyone pretending otherwise is lying to a board deck.

Spotify's Anne Bouttier says the point of a third party is that buyers stop relying on audio platforms grading their own homework, a phrase that belongs on a tote bag at every ad tech conference for the next decade. Acast's Valerie Reimer says the real prize is the long tail, that brands will fund smaller shows once they can see where the ad landed, which is the most defensible thing in the entire week's news and I'm not going to be cute about it. Good. Do that. But Reimer also says the quiet part all the way out loud: this lays the foundation for audio to be swept into agentic AI deals, so the machines can shop across channels without bothering to drill into media types. Which delivers us, with the grim inevitability of a sugya you can't get out of, to the next item.

IAB Tech Lab dropped AAMP 2.3 on Thursday, an enterprise update to its Agentic Advertising Management Protocols. It plays nice with Amazon Bedrock and Databricks, extends buying to platforms like Meta, hooks into Google Ad Manager reporting, and, per the announcement, executes programmatic transactions autonomously. CEO Anthony Katsur put it about as plainly as a standards body can: the industry has already shown what agents can do, the job now is making them reliable where advertising actually happens. Translation: the demo crushed, now let's see it survive contact with an actual supply chain.

And then you read the feature list, which is essentially a list of things that have already gone wrong. It embeds the IAB Diligence Platform and SafeGuard Privacy into the Buyer Agent. It improves pricing guardrails. It improves the integrity of transaction data. It specifies that inventory availability is based on "real, not derived" numbers. I need everyone to stop and appreciate that somebody had to write down, in a formal enterprise standards document, in 2026, that the ad should exist. We are stacking an autonomous negotiation layer on top of a market where "is this impression real" required its own bullet point. The protocol isn't the villain. The protocol may be the only grown-up in the building. But a rulebook this specific is a confession.

Grace Harmon at EMARKETER, on an analysis from TAG, the ANA, and Fiducia. AI slop is only 1.3% to 2.4% of measured open web programmatic spend, which sounds like a rounding error until you look at how it scores. Viewability: 77.2% for slop versus 74.9% for clean inventory. Invalid traffic: 0.05% for slop versus 0.32% for clean. Factor in measurability and slop gets graded premium more than 70% of the time. Oh, and 88% of it is made-for-advertising content in the classic content farm genres: recipes, personal finance, how-to.

Sit in that for a second. The machine-generated recipe blog that no human wrote and no human read is cleaner, more viewable, and more premium than the outlet paying actual journalists. That is not a flaw in the slop. That is a result about the instruments. Viewability measures whether pixels rendered. IVT measures whether a bot loaded them. Neither one measures whether a person gave a damn, and slop optimizes with total ruthlessness for exactly the things we chose to count. The algorithms are not broken. They are working perfectly, toward the target we set. EMARKETER's read is that a new measurement layer is coming and that authenticity scoring becomes a spending category, which is true and also this industry's oldest reflex: sell a vendor solution for a problem the vendor layer created. The harder question, the one nobody's budget line covers, is what we thought we were buying in the first place. Selling the pause. Selling the transcript. Selling the empty page to a robot with a rulebook. And when the woman with the clipboard asks for a seat, discovering, tragically, an obligation to our supply partners.

Meta signed an AI content deal with Newsmax, which now joins News Corp, Warner Bros. Discovery, Fox News Media, the Daily Caller, and friends in Meta's stated pursuit of balance. Newsmax paid $67 million to settle voting equipment defamation claims. Fox was ordered to pay $787 million, the largest media defamation settlement in US history. Those outlets are now training data. "Balance" is carrying an entire warehouse on its back in that sentence.

Microsoft posted revenue up 18% to $90 billion, Azure crossed $100 billion for the fiscal year, and Microsoft 365 Copilot blew past 30 million paid seats. Search advertising ex-TAC was up 10% while Windows OEM fell 7% and Xbox fell 10%, meaning the ad business is now the growth engine inside More Personal Computing. Also, quietly, Microsoft's Anthropic stake returned a $3.2 billion gain in the quarter. Everyone is an ad company and everyone is an AI investor and nobody makes anything. Snap, meanwhile, is wooing agencies with new Business Manager permissions, an Indy Agency Partner Program going into France and the UK, a Creator Agency Partner Program with nine pilot partners, and redesigned partner badges. Badges! We have reached the badge phase of the funnel. Somewhere a slide deck is being made about the badges.

Here is the pitch. You are watching something on Max. You get up to deal with the dog, the kid, the kugel. You hit pause. The screen does not go black. The screen goes to work. A full-bleed brand moment appears, tastefully lit, with a QR code sitting there like a coupon in a synagogue newsletter. You scan. You buy the shirt. Commerce, at the speed of bathroom break.

Here is the reality. Roughly 400 people out of ten million scan that code.

That is not my number. That is Brightline's own first-party data, which puts QR scan rates on connected TV at 0.004%. Streaming Media, measuring independently, gets QR click-through around 0.03%, which is nearly an order of magnitude better and still functionally indistinguishable from zero.

Origin Media ran a 10,000-household study and found that 33.5% of viewers say a shoppable TV ad makes them less likely to buy.

MarTechSeries reports that only 10% of American adults have ever completed a purchase through shoppable CTV, ever, once, in their entire lives.

So let us talk about the other numbers. The good ones. The ones on the slides.

Pause ads are not new and nobody should pretend otherwise. Hulu was testing this in 2018. Peacock shipped with it. Warner Bros. Discovery put it on HBO Max in 2022, three years before anyone said the word "shoppable" out loud at an upfront.

What is new is the standardization push and the shoppable layer, plus a fresh convoy of research engineered to make the format feel inevitable rather than merely convenient. Every ad-supported streamer runs some version now: Hulu, Max, Peacock, Roku, and increasingly the sports-adjacent stuff like WBD's "Unbreakable" suite across TNT Sports and Bleacher Report.

The player list matters, because it tells you this is not a WBD eccentricity. It is the industry's collective answer to a very boring, very real problem: there are only so many ad minutes on an ad-supported tier, and the pause screen is the last unmonetized real estate in the house. Somebody was always going to build on it. Dead air is beachfront property.

Read that right column again. Every single meaningful figure in the pause ad story traces back to the seller, a trade body the seller belongs to, or a research vendor whose invoices are paid by the platforms being graded.

I want to be fair here, because fairness is the whole job: vendor-funded research is completely standard in adtech. It is not fraud. It is not even unusual.

But it means the 67%, the 51%, and the glorious 200%+ are directional marketing data, and they should be read the way you read a real estate listing. "Cozy" means small. "Up to a 50% premium" means somebody paid it once.

Here is my favorite statistic in the entire category, and I say this with love.

67% of Gen Z and millennials prefer a pause ad to a frozen screen.

Of course they do. The alternative is a photograph of Pedro Pascal's left nostril, held for four minutes, while you find the peanut butter. That is not a market signal. That is a preference between an advertisement and nothing at all, and the nothing was losing.

You could run this study on almost any surface. Would you prefer a nice picture on the elevator wall or a blank elevator wall? Congratulations, elevators are now a 67%-preferred commerce channel.

I am not here to torch the format, because the attention data holds up better than the commerce data, and the distinction matters.

FreeWheel's Viewer Experience Lab found 58% to 61% of viewers say pause ads fit seamlessly into the content. Brightline-sourced data via eMarketer cites a 34% lift in unaided recall. Buyers are reportedly paying up to a 50% premium.

And structurally, the unit is legitimately strong. Viewability is 100% by definition, because the viewer caused it. There is no scroll. There is no skip. There is no competing creative fighting for the same eyeball. A full-screen static image on a paused television is, mechanically, one of the cleanest impressions in modern media

Attention is not the problem. Attribution is.

Every hard purchase-completion number from a non-seller source points the same direction: the shoppable pause ad, specifically, is an attention product wearing a commerce costume. Even the sympathetic voices hedge. Alan Wolk asked practitioners to show him a live, frictionless shoppable campaign, and by his account they all went quiet. AdExchanger's own July 2026 coverage of WBD's push describes the format as hard to monetize without standardization, which is a remarkable thing to find in a story built on a growth narrative.

This is the part everybody waves past, and it is the part that actually decides whether this becomes a category or a case study.

You cannot buy this inventory programmatically at scale, because the specs are still in public comment.

The IAB Tech Lab released its CTV Ad Portfolio covering Pause, Menu, Screensaver, In-Scene, Squeezebacks, and Overlay for public comment on December 11, 2025, with a window closing January 31, 2026. A subsequent version extended comment to June 5, 2026. We are talking about creative dimensions, file types, safe zones, and OpenRTB signaling. Pixel counts. The plumbing that lets a DSP know what it is even bidding on.

Until those specs are finalized and adopted, pause and shoppable inventory has to be hand-sold, insertion-order style, by humans in nice shoes. Which is precisely what WBD is doing, and precisely why it is doing it.

A year after the shoppable version was announced with a straight face at upfronts, the standards body is still processing comments on how big the box should be.

Not entirely. But it is a bet with encouraging attention data and unfinished plumbing, not a proven revenue category, and the gap between those two things is where a lot of media money goes to die.

Real: high viewability, real recall lift, defensible premium CPM, an obvious answer to finite ad load.

Aspirational: standardized programmatic trading, independent verification of purchase completion, and any credible evidence that the shoppable layer converts materially better than a static pause ad with a phone number on it.

The honest version of this pitch is that Warner Bros. Discovery monetized dead air on attention and recall grounds, and that is a genuinely good business. They should say that. It is defensible, it is true, and it does not require a single QR code.

The commerce story bolted on top is running several laps ahead of the evidence. And the people cheering loudest are, almost without exception, holding the inventory.

The full ADOTAT Scorecards. Vendor comparisons that name winners and losers, with the analysis your planning team needs before Q3 budget meetings.

The takes nobody else will publish. Independent means we don't answer to vendors, agencies, or the trade press establishment.

The archive. Every Scorecard, every analysis, searchable. The reference layer for how this industry actually works.

Read by senior leaders at every major holding company agency, GroupM, Publicis, Omnicom, IPG, Havas, Dentsu, and inside the buying and strategy teams at Netflix, Roku, Pinterest, Disney, NBCUniversal, Spotify, Amazon, Meta, and most of the platforms shaping where ad dollars actually go.

If your competitors are reading ADOTAT+ and you aren't, you're planning with less information than they are.

Expense it.

Founding member rate. Locks in for life.

Also, no ads. Zero ads. No Ads.