Podcast episode

WTF is the IAB's Redefining Media Types standard?

by Ken Rona ·July 29, 2026 ·The Digiday Podcast

TL;DR

The Digiday Podcast interviews Jamie Finstein, IAB VP of Media Center, about the newly released "Redefining Media Types" (RMT) standard — a proposed taxonomy for categorizing digital video inventory by viewing environment and attributes. The episode is primarily a standard explainer: useful for practitioners trying to understand why the industry needs a shared vocabulary for CTV, streaming, DOOH, and social video, but light on hard data or controversy.

What was covered

RMT standard origin and structure. The IAB released the Redefining Media Types standard in mid-July; a public comment period runs until August 8th. The standard creates three top-level viewing environment buckets: lean-back large screen, personal screen, and public ambient display (i.e., digital out-of-home/DOOH).
Why CTV ≠ airport/bar screens. A core deliverable of RMT is formally separating public ambient displays (screens in airports, restaurants, bars) from connected TV (CTV), closing a definitional gap that DOOH vendors have exploited when pitching CTV budgets.
Consumer-first framing. Finstein says the IAB structured the standard top-down from consumer experience (how viewers think about content) before layering in bottom-up operational/technical attributes — including addressability and data signals — to avoid an industry-only lens.
AI agents as the urgency driver. Finstein cites agentic media buying — AI systems transacting on buyers' and sellers' behalf — as the primary reason to codify definitions now. Without shared language, she argues, mismatched definitions will cause costly errors at scale in programmatic execution.
Edge cases acknowledged. Hotel-room TVs, a bar showing an NFL game (sound-on vs. sound-off), and FAST channels delivered over the internet but consumed like linear were all flagged as unresolved edge cases the working group debated but did not fully resolve.
Overlap with measurement and planning tools. Finstein distinguishes RMT from IAB's parallel "Project Etos" measurement workstream but notes RMT attributes (e.g., duration of viewing session) can feed agency planning tools and downstream media quality conversations — even though IAB explicitly declines to define "premium."
Early feedback themes. Initial public comments have clustered around duration/length definitions and how linear TV should be baselined; CTV-vs.-streaming definitional confusion (CTV as device vs. content category) is anticipated as a likely friction point.

Notable claims & predictions

Jamie Finstein on AI agents: "If we as humans can't agree on the same language, how are we going to expect these agents that are transacting on our behalfs to be able to speak the same language and know what is what?" She frames RMT as essential pre-training infrastructure for agentic ad buying.
Finstein citing agentic transaction scale: At a Denver conference in June, a panelist estimated only "3% to 5% of total inventory" is currently being transacted by AI agents — but Finstein warns that could reach meaningful scale "within a year, two years."
Tim Peterson on the practical risk: Without RMT, "you tell an agent, okay, I want CTV inventory for this campaign and they end up running in an airport terminal display because technically it's a connected TV."
Finstein on ambiguity having incumbents: She implies (without naming names) that "there is something to be gained from ambiguity for some" — a tacit acknowledgment that DOOH companies and others benefit from definitional blurring.
Finstein on CTV vs. streaming: The RMT working group defined CTV strictly as the connected TV device, meaning Netflix watched on an iPad is not CTV under this standard — a meaningful departure from how most buyers and platforms currently use the term.

Fact check

Claim (Finstein): Only "3% to 5% of total inventory" is currently being transacted by AI agents. This figure comes from an unnamed panelist at an unnamed Denver conference in June, relayed second-hand. It is unverified — no source, methodology, or inventory universe is defined. The number may refer to a specific company's inventory, a specific format, or a loosely defined concept of "agentic" transactions. Listeners should treat it as anecdotal color, not an industry measurement.
Claim (Finstein): The RMT standard defines CTV strictly as the connected TV device, not the content/service. This is consistent with IAB's technical definitions historically, but it runs directly counter to how The Trade Desk (DSP software advertisers use to buy digital ads), Netflix, Amazon, and most buy-side practitioners currently use "CTV" interchangeably with streaming services regardless of device. Finstein herself acknowledges this mismatch will likely generate pushback in the comment period. True as a definitional choice; contested as an industry norm.
No claims rise to clearly false.

Why this matters for ad-tech operators

Buy-side targeting hygiene. RMT creates a formal definitional firewall between CTV and public ambient display/DOOH. Buyers who have been unknowingly or knowingly letting DOOH inventory bleed into CTV line items now have an industry document to reference in vendor negotiations and RFPs — and a framework to encode into DSP deal parameters.
Agentic buying prep. If AI agents executing programmatic buys proliferate within 1–2 years (per Finstein's timeline), the absence of a shared ontology becomes a literal execution risk: misclassified inventory signals mean misallocated spend at machine speed. RMT is positioned as the definitional layer that makes agent instructions unambiguous — making early adoption by SSPs (sell-side platforms that help publishers sell ad inventory) and DSPs a de-risking move.
Planning tool integration. Finstein notes that agency planning tools are an immediate practical use case. If major holdcos (Omnicom, Publicis, WPP, GroupM) embed RMT attributes into their proprietary planning systems, it could accelerate adoption faster than a traditional IAB standards cycle — but that alignment hasn't been publicly confirmed.
Impact is indirect and longer-dated for most operators. This episode describes a standard still in public comment, without confirmed platform adoption or enforcement mechanism. Ad-tech operators should monitor the final standard (post-August 8th feedback) and any subsequent SSP/DSP implementation announcements before adjusting workflows. The immediate relevance is strategic awareness, not operational change.

Full analysis

The IAB wants to redraw the map of digital video. Its new "Redefining Media Types" standard, explained on the Digiday Podcast by IAB VP Jamie Finstein, sorts video inventory into three buckets by where you watch it: lean-back big screen, personal screen in your hand, and public ambient display like the screen over an airport gate. The load-bearing move is a definitional wall between connected TV and digital out-of-home. The stated urgency is AI agents that will soon buy media on our behalf and need a shared vocabulary to not screw it up.

This is a Type 2 decision for most operators. Nothing is being enforced. It is a public comment draft with feedback open through August 8th. What's actually being decided is whose definition of "CTV" wins, and whether the industry codifies that before agents start transacting at scale.

The Skeptic. The whole thing rests on one bet: that AI agents will buy enough inventory, soon enough, that a missing dictionary becomes a real cost. Finstein's own evidence for that is a secondhand number from an unnamed panelist at an unnamed Denver conference, guessing 3% to 5% of inventory is agent-transacted today. Call it what it is: a hallway rumor, not a measurement. And here's the tension the standard walks straight into: it defines CTV as the physical device, so Netflix on an iPad isn't CTV. Nobody on the buy side talks that way. Trade Desk doesn't, Netflix doesn't, buyers don't. A standard that contradicts how the entire market already speaks doesn't reduce confusion. It adds a second language nobody asked for. For a non-specialist: the IAB is writing a grammar book for a language everyone already speaks their own way.

The Market Analyst. Follow who benefits from the fog. Finstein all but says it: "there is something to be gained from ambiguity for some." Translation, DOOH vendors have been pitching airport and bar screens into CTV budgets because "technically it's a connected TV." A hard CTV/DOOH wall takes that arbitrage away. So the winners of adoption are the pure-play CTV sellers and the SSPs that route that inventory honestly. The losers are DOOH shops that priced against CTV RPMs. For measurement firms, VideoAmp and iSpot and the like, a stable taxonomy is upstream plumbing they'd love to have settled. In plain terms: a few vendors have been selling cheaper screens at premium prices by blurring a line, and this draft draws the line.

The Operator. Fine, but who implements it Tuesday? A standard changes nothing until a DSP encodes it into deal parameters and an SSP tags inventory to match. That's the part the podcast waves past. Right now a trader who wants "CTV, not airport terminals" already writes exclusions and vets deal IDs by hand. The honest sellers already sell it straight. The dishonest ones will keep mislabeling until a buyer's contract cites chapter and verse. So the real value here is having an IAB document to point at in an RFP dispute, full stop. That's useful. It is not a workflow change. And the edge cases they punted on, hotel TVs, a bar showing the NFL game sound-off, FAST channels over IP, are exactly where the mislabeling money lives.

The Customer / End User (the buyer). Agencies are the tell. Finstein flags agency planning tools as the immediate use case. If GroupM, Publicis, Omnicom, and WPP bake RMT attributes into their proprietary planners, adoption happens fast and everyone downstream follows. If they don't, this sits in comment-period limbo. Buyers do want the CTV/DOOH firewall, they've been burned by it. What they don't want is to relearn "CTV means the device," because their whole media plan is organized around streaming service, not hardware. For the informed outsider: the people spending the money like half of this and will quietly ignore the other half.

The CFO. Cost to engage is near zero, and that's the point. Reading the draft and filing a comment by August 8th is cheap insurance. Re-tooling systems around a standard with no adoption commitment and no enforcement is spending ahead of demand. The payback only exists if agentic buying actually scales and if the buy side accepts the device-based CTV definition. Two unproven ifs. So you spend the comment-period hours, you don't spend the integration budget.

Where the council splits. First, does the agent timeline justify acting now? The Skeptic says the urgency is built on a rumor; the Market Analyst says taxonomies get set early and you don't want to be relitigating definitions once machines are transacting. Second, the device-based CTV definition: technically defensible, commercially at war with how everyone already talks. That single choice decides whether RMT becomes the industry's language or a footnote the buy side routes around.

What it hinges on. Two things. Whether the holdcos adopt RMT attributes into planning tools, which is the real adoption engine, and whether the final standard bends on the CTV-as-device definition after the comment period. If the IAB holds the line and the agencies balk, the standard stalls. To de-risk: file a comment, watch the post-August 8th revision on the CTV definition, and watch for any DSP or SSP that announces it's encoding RMT into deal parameters. That announcement, not the draft, is the signal.

Prediction: When the IAB publishes the revised RMT standard after the August 8th comment period closes, the strict "CTV equals the physical device" definition will be softened or carved out with exceptions, not held as written.

Confidence: Medium. Finstein herself predicts pushback, and the definition fights universal buy-side usage.

Why: The draft defines CTV as the device, so Netflix on an iPad isn't CTV, and Finstein openly expects that to draw fire in comments. Standards bodies revise the specific clauses that generate the most public friction, and this one collides with how Trade Desk, Netflix, and every buyer already use the word. When a definition contradicts settled market language, the body almost always adds carve-outs rather than force the whole industry to relearn its vocabulary. The opposite outcome, the IAB holding a definition the buy side rejects, would make the standard dead on arrival, which nobody in the working group wants.

Revisit by 2026-10-15: We're right if the revised RMT standard adds exceptions, softening, or a content-based alternative to the device-only CTV definition. We're wrong if it ships with the strict device-only definition intact.

Listen to the episode →