Digital video advertising has quietly become a Tower of Babel. A “video” impression on a connected TV app, a six second social clip, and a FAST channel pre roll get planned, bought, and measured as though they are the same product, when in practice each behaves nothing like the others. The Interactive Advertising Bureau is trying to fix that, and its answer says as much about where media buying is headed as it does about video itself.
The industry has been arguing over definitions for years
On July 9, IAB and IAB Tech Lab released the Redefining Media Types (RMT) Standard for public comment, an attempt to build one shared taxonomy for digital video across connected TV, online video, social video, FAST channels, video podcasting, and retail video. The comment period runs through August 8, with a finalized version expected later in 2026.
Jamie Finstein, VP of IAB’s Media Center, said the project grew out of something mundane: confusion over terminology in IAB’s own video ad spend reports. As the same questions kept surfacing across separate working groups on measurement, streaming, and investment, standardizing the underlying definitions became a priority the trade group could no longer defer.
What the standard actually defines
Two layers, not one label
RMT is built as a two layer framework rather than a single tag. The first layer sorts video into four viewing experience categories, described as Lean Back, Personal Screen, Passive, and Communal, which group inventory by how audiences actually engage with it rather than by the device or platform selling it. The second layer adds impression level operational attributes: sound, skip capability, full screen presentation, addressability, measurability, device type, and ad format.
That structure matters because it separates “what kind of viewing moment is this” from “how is this specific impression instrumented.” A buyer can ask for Lean Back inventory with guaranteed sound-on delivery and addressability, for instance, without needing a separate glossary for every publisher’s version of “premium video.”
Built by the people who have to use it
The working group behind RMT started in January and pulled in measurement specialists, streaming product leads, investment officers from major holding companies, and members of the Media Rating Council. IAB Tech Lab developed the technical specification and will maintain it going forward, with an explicit design goal of feeding into the IAB Measurement Center’s Project Eidos taxonomy work and informing the OpenRTB and AdCOM specifications that already run most programmatic video transactions.
That last detail is the one worth sitting with. A definitions document is only as useful as the systems that enforce it, and by wiring RMT into OpenRTB and AdCOM, IAB is trying to make consistent video terminology a property of the bid request itself, not just a shared vocabulary humans agree to use in meetings.
Why this is really about agents, not glossaries
IAB’s own framing points at the real trigger. Finstein argued that as agentic AI becomes a more prominent part of the media buying process, a common language stops being a nice-to-have and becomes a hard requirement: if the humans planning a campaign are not speaking the same language, there is no way to train the models that increasingly execute on their behalf. That is the industry-first admission this initiative makes. Standardized taxonomy has been a low-grade irritant for programmatic buyers for years; it becomes a blocking issue the moment software, not people, is expected to interpret “video” consistently across a bid stream. The programmatic supply chain’s push toward naming every party that touches a bid request was the transparency layer; RMT is the semantic layer sitting underneath it, and both are converging on the same conclusion: agentic buying cannot scale on inconsistent metadata.
The timing lines up with where budgets are actually moving. CTV spend, cross-platform video planning, and programmatic video investment have all continued to climb, and measurement problems that were once vendor-specific footnotes, like the ad repetition and frequency issues Omnicom’s research has quantified on streaming platforms, are now industry-wide symptoms of the same root cause: nobody agrees on what is actually being measured.
What this means for the marketing leader
For a CMO or head of media, RMT is not a standards-body curiosity. It is a preview of how buying briefs, reporting dashboards, and vendor contracts will need to be rewritten once agentic tools sit between strategy and execution. Three practical implications follow:
First, media plans built on today’s inconsistent labels will need to be re-mapped once RMT lands, so any dashboard or MMM model that ingests “video” as a single line item is a candidate for near-term rework. Second, procurement teams evaluating CTV, social video, or retail media partners should start asking vendors now whether their inventory taxonomy will align to RMT’s viewing-experience categories, since misalignment here is exactly the kind of gap that breaks automated buying later. Third, any team piloting agentic media-buying tools should treat inconsistent video definitions as a live risk, not a future one: an agent optimizing across platforms that define “video” differently will misallocate budget with total confidence.
What to do before the comment period closes
Marketing and media teams do not need to wait for the finalized standard to act. Reviewing current video buying briefs against RMT’s two-layer structure now will surface exactly where internal definitions diverge from the direction the industry is heading, and submitting feedback during the public comment window, open through August 8, is the cheapest way to influence a taxonomy that will likely become a contractual reference point within a year. Waiting for the finished standard means inheriting whatever definitions get locked in without having had a say.
Source: IAB
