Cross-Platform Video Production: A Viewing-Condition Matrix

A video can travel across channels, but changing its aspect ratio at export does not make it cross-platform. The version should change when the viewer’s relationship with the screen changes: whether sound starts on, whether the ad can be skipped, how much of the display it occupies, and what the viewer can do next.

On September 28, 2026, IAB published the final A New Standard for Classifying Video Advertising. It first organizes video by consumer viewing environment—large-screen communal viewing, personal screens and public ambient displays—then describes the impression through viewing and operational attributes. The production implication is practical: a platform name is not a creative specification.

The framework does not prove that every platform, seller or measurement provider has implemented one classification. Brands can still use its language now to write clearer briefs and stop treating every channel adaptation as an export task.

Why “CTV, social and retail media” is an incomplete brief

A CTV placement may be full-screen in a living room, sound-on by default and completion-required. A personal-screen placement may sit inside a mobile feed where the viewer can scroll away. Retail media may mean a clickable product video inside a shopping app or a public screen in a store. All are video, but the audience does not encounter them in the same way.

IAB’s framework says a single platform can distribute more than one media type. Classification belongs to the viewing environment and inventory attributes at delivery, not to the platform logo. A production team therefore cannot infer the shot design from “YouTube version,” “Amazon version” or “Meta version.” It must ask where the ad appears, what the sound default is, whether it fills the screen, and which other interface elements compete with it.

The two layers in IAB’s framework

Layer one: how the viewer encounters the ad

Large-screen communal viewing describes a primary household screen with co-viewing potential and an extended viewing session.
Personal screen covers smartphones, tablets, computers and wearables, ranging from deliberate viewing to incidental exposure in feeds, games and retail surfaces.
Public ambient display covers venue screens and digital video billboards where multiple people may see the media in a public or semi-public setting.

These are neither aspect-ratio labels nor platform lists. They describe the physical and behavioural relationship between audience and screen.

Layer two: what the impression can do

The standard further distinguishes sound default, skip-enabled versus completion-required delivery, full-screen versus partial-screen presentation, and whether audience decisioning operates at an individual, household or unit level. Six operational attributes add the execution detail: signals, content context, device type, ad format, screen size and surface, and content source and supply type.

Those fields do not write the script for the brand. They change the creative decision. A product claim can unfold through performance and music on a sound-on living-room screen. The same claim in a silent mobile feed needs an earlier visual cause and result. On a public display, small copy and a story that depends on continuous viewing are weaker choices.

The ONCE viewing-condition card

A production agreement does not need to reproduce the full IAB taxonomy. For every priority placement, ONCE recommends a five-line card:

Viewing relationship: communal large screen, personal screen or public ambient display; deliberate session or incidental encounter?
Attention constraints: sound default, ability to skip or scroll, and full-screen or competing interface?
Evidence that must survive: product appearance, operating action, performance result, offer condition or disclosure?
Available action: watch only, click, scan, shop, pause or open a destination?
Acceptance package: aspect ratios, durations, captions, safe areas, audio stems, poster frames, commerce overlays and platform previews?

If two placements produce nearly identical answers, they can probably share a master. When the viewing relationship and attention constraints differ, the production should capture separate modules before post-production.

How one shoot can serve three viewing environments

Capture evidence modules first. Give the core product action, result state and limiting condition their own shots so the facts do not depend on one voiceover rhythm. Silent mobile edits, narrative CTV cuts and public-screen loops can then draw from the same evidence library.

Capture attention modules next. Preserve character relationships, spatial blocking and sound design for communal screens. Shoot closer product actions, vertical composition and an earlier value cue for personal screens. For public displays, create distant-readable silhouettes, short loop points and state changes that do not rely on small copy.

Reserve interaction space. Shoppable video and retail media need safe areas for prices, buttons and product cards. Our article on dynamic TV creative separates a stable evidence master from a dynamic transaction layer. The viewing-condition card works one step earlier, deciding which scenes should never be forced into the same master.

Treat sound-on and silent viewing as distinct storytelling modes. A silent version is more than the original mix muted under a wall of captions. Action, composition and state change should carry the causal story; captions carry language. For caption files, descriptive text and player QA, see our production workflow for Europe’s accessibility baseline.

Compare suppliers by approved viewing package

Counting delivered videos mixes masters, simple exports and real scene redesigns. A more useful internal model is:

Cost per approved viewing package = (shared strategy and production + environment-specific picture and sound modules + adaptation and platform QA + rights and version records) ÷ viewing-condition packages that passed acceptance.

This is a decision model, not an industry price benchmark. A “ten-platform package” that exports one film ten times creates many files without adding viewing situations. A shoot that captures stable product evidence, personal-screen action, communal-screen storytelling and a public-display loop can reduce future reshoots even when the initial asset count is smaller.

When one master is still enough

A brand film, launch-event opener or website hero may need only one master and a few framing adaptations when the viewing environment is concentrated and interaction is simple. A short test can also begin with one core execution, then fund environment-specific production after the proposition shows promise.

When a campaign spans living-room TV, silent feeds, shoppable retail pages and public displays—and each setting performs a different job in the buying journey—the viewing-condition card belongs in pre-production. IAB provides common language; product truth, audience task and real inventory still determine the creative choice.

A minimum action for the next campaign

Take one video currently running across channels and describe every placement with the five-line card. If the team can name platforms but cannot state the sound default, exit behaviour, screen occupancy, required evidence and available action, version planning is still happening at export. Resolve those questions before deciding how many films to commission.