Media Channel Taxonomy

AdCP media channel taxonomy: the 19 standardized channel types (CTV, OLV, DOOH, podcasts, retail media, etc.) used for media planning and budget allocation.

Media Channel Taxonomy

Status: Draft Specification Version: 1.0.0-draft Last Updated: 2026-01-23

This specification defines a standardized taxonomy for advertising media channels. It is designed for interoperability across media planning tools, ad tech platforms, and AI-powered advertising agents.

Motivation

The advertising industry lacks a standardized channel taxonomy. While IAB Tech Lab provides taxonomies for content, audience, and ad products, no equivalent exists for media channels. This leads to:

• Inconsistent channel naming across platforms
• Difficulty aggregating cross-channel campaign data
• Confusion between channels, formats, and buying models
• Friction in automated media planning workflows

This specification addresses these gaps by defining:

1. Media Channels - How buyers allocate budget (planning abstractions)
2. Property Types - Addressable inventory surfaces with verifiable ownership
3. Clear distinction between channels, property types, and formats

Design Philosophy

Channels represent how buyers plan and allocate budget, not where ads technically render.

This is a deliberate design choice. Buyers don’t say “I have $500K for web” — they say “I have $500K for display” or “I have $300K for OLV.” The channel taxonomy reflects this reality.

Key Principles

1. Planning-oriented: Channels match how agencies structure media plans
2. Lightweight tags: Channels help buyers express intent, not enforce precise classification
3. Multi-channel support: Properties may align with multiple channels (YouTube = olv, social, ctv)
4. Publisher-declared: Publishers indicate which channels their inventory supports
5. Agent-reconciled: Sales agents match buyer intent with publisher claims

Channels vs Technical Substrates

Channels deliberately avoid substrate-level concepts like “web” or “mobile app” because:

• Buyers don’t plan that way
• Most digital inventory would fall into both categories
• The same placement can be bought different ways

Instead, channels describe buying contexts: display, OLV, social, search, CTV, etc.

Terminology

The key words “MUST”, “MUST NOT”, “REQUIRED”, “SHALL”, “SHALL NOT”, “SHOULD”, “SHOULD NOT”, “RECOMMENDED”, “MAY”, and “OPTIONAL” in this document are to be interpreted as described in RFC 2119.

Definitions

Media Channel

A media channel describes how buyers allocate budget. It represents a planning abstraction that encodes assumptions about audience, environment, and buying approach.

Channels are defined by buying context, not by:

• Technical substrate (web vs app)
• How the ad looks (that’s a format)
• The specific technology serving the ad

Property Type

A property type describes a specific addressable inventory surface where:

• Ownership can be verified (e.g., via adagents.json)
• Ads can be programmatically served
• Identifiers exist for the property (domains, app IDs, device IDs)

Property types are technical classifications, distinct from channels. The same property may support multiple channels.

Format Category

A format category describes HOW an ad renders - its creative unit type. Examples: video, audio, display, native. Formats are orthogonal to channels.

Understanding Channels vs Property Types

Concept	Question It Answers	Determined By	Example
Channel	How do buyers allocate budget?	Planning context	retail_media
Property Type	Where does the ad technically render?	Addressable inventory surface	website

Why This Matters

Consider retail media: when a buyer allocates budget to “retail media,” they’re buying:

1. Sponsored products on retailer websites
2. Display ads on retailer apps
3. Off-site ads using retailer data
4. In-store digital screens

All of these are retail_media channel, but the property types vary (website, mobile_app, dooh).

Some names like dooh, podcast, radio, and streaming_audio appear as both a channel and a property type. The channel describes the buying context (how budget is allocated); the property type describes the technical surface (where the ad renders). For example, an in-store retail screen is retail_media channel / dooh property type, while a standalone digital billboard is dooh channel / dooh property type.

Multi-Channel Properties

Properties can align with multiple channels. Examples:

Property	Channels	Reasoning
YouTube	olv, social, ctv	Different buying contexts on same platform
ESPN App	olv, display	Supports video and display inventory
Amazon	retail_media, search, display	Multiple ad products

When to Use Each

Use channel when:

• Filtering products by budget allocation category
• Reporting on media mix
• Expressing buyer intent in product discovery

Use property_type when:

• Validating inventory ownership via adagents.json
• Technical ad serving decisions
• Platform-specific targeting

For audio inventory, execution and distribution details that sit below channel and property type belong on product or placement metadata. Use audio_distribution_types to distinguish music streaming services, FM/AM broadcast, podcasts, catch-up radio, web radio, video-game audio, and text-to-speech without renaming buyer-facing channels or adagents.json property types.

Media Channels

Channel Enum

The following 20 channels MUST be supported. Implementations MAY extend with additional channels using the ext field.

Channel	Description
display	Digital display advertising (banners, native, rich media)
olv	Online video outside CTV (pre-roll, outstream, in-app video)
social	Social media platforms
search	Search engine advertising
ctv	Connected TV and streaming on television screens
linear_tv	Traditional broadcast and cable television
radio	Traditional AM/FM radio broadcast
streaming_audio	Digital audio streaming services
podcast	Podcast advertising
dooh	Digital out-of-home screens
ooh	Classic out-of-home (billboards, transit)
print	Newspapers, magazines, print publications
cinema	Movie theater advertising
email	Email advertising and newsletters
gaming	In-game advertising
retail_media	Retail media networks and commerce marketplaces
influencer	Creator and influencer partnerships
affiliate	Affiliate networks and performance-based partnerships
product_placement	Product placement and branded content
sponsored_intelligence	Sponsored Intelligence — AI assistants, AI search, and generative AI platforms

Channel Definitions

display

Digital display advertising including banners, native units, and rich media across web and app environments.

Includes:

• Display banners on websites
• Display ads in mobile apps
• Native content units
• Rich media ads
• Interstitials (non-video)

Excludes:

• Video ads (use olv or ctv)
• Social platform ads (use social)
• Search results (use search)
• Retail media placements (use retail_media)

Typical Formats: display, native, rich_media

olv

Online video advertising delivered outside of CTV/television environments.

Includes:

• Pre-roll, mid-roll, post-roll on websites
• In-app video ads
• Outstream/in-feed video
• YouTube video ads (when not on TV screens)
• Video on news sites, sports sites, etc.

Excludes:

• CTV/streaming on TV screens (use ctv)
• Social platform video (use social)
• Linear TV (use linear_tv)
• Retail media video (use retail_media)

Typical Formats: video

Note: OLV (Online Video) is a distinct planning bucket from CTV. Agencies commonly budget separately for “OLV” and “CTV” campaigns.

social

Social media platform advertising, regardless of technical delivery surface.

Includes:

• Meta platforms (Facebook, Instagram, Threads)
• X (Twitter)
• TikTok
• LinkedIn
• Snapchat
• Pinterest
• Reddit
• YouTube (when bought through social-style targeting)

Excludes:

• Influencer content on social platforms (use influencer)
• Video ads bought for reach, not social engagement (consider olv)

Typical Formats: display, video, native

Note: Social is defined by BUYING CONTEXT (social platform ad tools) and audience engagement model.

search

Search engine results pages and search advertising networks.

Includes:

• Google Search ads
• Microsoft Bing ads
• DuckDuckGo ads
• App store search ads
• Shopping/product listing ads in search context

Excludes:

• Display ads on search engine properties (use display)
• Retail media search (use retail_media)

Typical Formats: text, display (shopping)

ctv

Connected TV advertising delivered through streaming applications on television screens.

Includes:

• Streaming service apps (Netflix, Hulu, Max, etc.)
• Virtual MVPDs (YouTube TV, Sling, etc.)
• Free ad-supported streaming TV (FAST)
• Smart TV native apps
• Gaming console streaming apps

Excludes:

• Linear broadcast/cable (use linear_tv)
• Video on phones/tablets/desktops (use olv)
• YouTube on mobile (use olv or social)

Typical Formats: video

linear_tv

Traditional broadcast and cable television advertising.

Includes:

• National broadcast networks (ABC, CBS, NBC, Fox)
• Cable networks (ESPN, CNN, HGTV)
• Local broadcast stations
• Addressable linear TV

Excludes:

• Streaming on TV screens (use ctv)
• TV Everywhere apps on mobile (use olv)

Typical Formats: video

radio

Traditional AM/FM radio broadcast advertising.

Includes:

• Terrestrial radio stations
• Satellite radio (SiriusXM terrestrial simulcast)
• HD Radio

Excludes:

• Streaming audio services (use streaming_audio)
• Podcasts (use podcast)

Typical Formats: audio

streaming_audio

Digital audio streaming services.

Includes:

• Music streaming (Spotify, Apple Music, Amazon Music, Pandora)
• Audio content platforms
• Digital radio streams (iHeartRadio digital, TuneIn)

Excludes:

• Podcasts (use podcast)
• Terrestrial radio simulcast (use radio)

Typical Formats: audio, display (companion)

podcast

Podcast advertising, including host-read and dynamically inserted ads.

Includes:

• Host-read sponsorships
• Dynamically inserted audio ads
• Podcast network advertising

Excludes:

• Music streaming (use streaming_audio)
• Video podcasts on YouTube (use olv or social)

Typical Formats: audio

dooh

Digital out-of-home advertising on electronic screens in public spaces.

Includes:

• Digital billboards
• Transit screens (subway, bus shelters, airports)
• Retail/mall digital displays
• Gas station screens
• Elevator screens
• Stadium/venue digital signage

Excludes:

• Static billboards (use ooh)
• Cinema screens (use cinema)

Typical Formats: display, video

ooh

Classic out-of-home advertising on physical (non-digital) surfaces.

Includes:

• Static billboards
• Transit posters
• Street furniture
• Wallscapes
• Wild postings

Excludes:

• Digital screens (use dooh)

Typical Formats: display (static)

print

Newspaper, magazine, and other print publication advertising.

Includes:

• Newspaper display ads
• Magazine display ads
• Newspaper/magazine inserts
• Trade publication advertising

Excludes:

• Digital versions of publications (use display)
• Direct mail

Typical Formats: display (static)

cinema

Movie theater advertising.

Includes:

• Pre-show advertising
• On-screen trailers and ads
• Lobby displays
• Concession advertising

Excludes:

• Streaming movie services (use ctv)

Typical Formats: video, display

email

Email advertising and sponsored newsletter content.

Includes:

• Sponsored email newsletters
• Email display advertising
• Dedicated email sends

Excludes:

• Transactional email
• CRM/owned email marketing

Typical Formats: display, native

gaming

In-game advertising across gaming platforms.

Includes:

• In-game display ads
• Rewarded video ads
• Playable ads
• Advergames
• Esports sponsorships
• Gaming influencer integrations

Excludes:

• Ads in non-gaming apps (use display or olv)
• Gaming content on streaming platforms (use ctv or olv)

Typical Formats: display, video, native

retail_media

Retail media networks and commerce marketplace advertising.

Includes:

• Retail media networks (Amazon Ads, Walmart Connect, Target Roundel)
• Grocery and delivery platforms (Instacart, DoorDash, Uber)
• Travel marketplaces (Expedia, Booking.com, Kayak)
• Financial services marketplaces
• Sponsored product listings
• On-site display and video on commerce platforms
• Off-site ads using retailer first-party data

Excludes:

• General display ads (use display)
• Social commerce (use social)
• Search ads on non-commerce platforms (use search)

Typical Formats: native, display, video

Note: Retail media is distinguished by its transactional context and closed-loop attribution capabilities.

influencer

Creator and influencer marketing partnerships.

Includes:

• Sponsored content creation
• Brand ambassador programs
• Affiliate creator partnerships
• User-generated content campaigns

Excludes:

• Ads placed on creator content by platforms (use social)
• Podcast host reads (use podcast)

Typical Formats: video, native, display

Note: influencer describes the BUYING MODEL (creator partnership) rather than where content appears.

affiliate

Affiliate networks, comparison sites, and performance-based publisher partnerships.

Includes:

• Affiliate networks (CJ, Rakuten, Impact, Awin, ShareASale)
• Comparison shopping engines (NerdWallet, Bankrate, The Points Guy)
• Review and recommendation sites
• Coupon and deal sites (RetailMeNot, Honey)
• Content commerce (editorial with affiliate links)
• Lead generation sites
• Cashback and loyalty programs

Excludes:

• Standard display ads on affiliate sites (use display)
• Influencer partnerships (use influencer)
• Retail media sponsored products (use retail_media)

Typical Formats: native, display, text

Note: affiliate describes the BUYING MODEL (performance-based, CPA/CPC/rev-share) rather than where content appears.

product_placement

Product placement, branded content, and sponsorship integrations.

Includes:

• Traditional product placement (film, TV)
• Virtual product placement
• Branded entertainment
• Event sponsorships
• Naming rights
• Branded content series

Excludes:

• Influencer partnerships (use influencer)
• Standard ad placements at sponsored events (use appropriate channel)

Typical Formats: native, video

sponsored_intelligence

Sponsored Intelligence — advertising within AI assistants, AI search results, and generative AI experiences via the reversed data flow.

Includes:

• Sponsored responses in AI assistants and chatbots
• Sponsored results in AI-powered search engines
• Display and video ads within generative AI experiences
• Brand experience handoffs via SI Chat Protocol

Excludes:

• Standard search engine ads (use search)
• Display ads on AI company websites (use display)
• Social platform AI features (use social)

Typical Formats: native, display, video

Note: sponsored_intelligence describes the BUYING CONTEXT where the user is interacting with an AI system, not traditional web or app surfaces. The channel uses the reversed data flow — buyers push data in, platforms generate ads with full context.

Property Types

Property types describe addressable inventory surfaces with verifiable ownership. They are used in adagents.json for authorization validation.

Property Type Enum

Property Type	Description	Example Channels
website	Web properties accessible via browser	display, olv
mobile_app	Native mobile applications	display, olv, social, gaming
ctv_app	Connected TV applications	ctv
desktop_app	Desktop applications (Electron, native)	streaming_audio, gaming
dooh	Digital out-of-home screen networks	dooh
podcast	Podcast feeds and episodes	podcast
radio	Radio station properties	radio
linear_tv	Linear television stations and networks	linear_tv
streaming_audio	Digital audio streaming properties	streaming_audio
ai_assistant	AI-powered conversational interfaces	sponsored_intelligence

Channels Without Property Types

The following channels do not have corresponding property types because they lack addressable, verifiable inventory surfaces in the traditional sense:

• ooh - Physical inventory lacks digital identifiers
• print - Physical publication inventory
• cinema - Theater inventory management systems
• email - Email list ownership differs from property authorization
• influencer - Creator relationships rather than property ownership
• affiliate - Performance-based partnerships, placements appear on partner websites
• product_placement - Content/event relationships rather than properties
• retail_media - Platform-managed inventory within retail ecosystems
• search - Platform-managed inventory
• social - Platform-managed inventory

Relationship Between Concepts

┌─────────────────────────────────────────────────────────────────┐
│                        MEDIA CHANNEL                            │
│         How buyers allocate budget (planning abstraction)       │
│                                                                 │
│   display, olv, social, search, ctv, linear_tv, podcast,       │
│   streaming_audio, radio, dooh, ooh, print, cinema, email,     │
│   gaming, retail_media, influencer, affiliate,                  │
│   product_placement, sponsored_intelligence                      │
└─────────────────────────────────────────────────────────────────┘

┌─────────────────────────────────────────────────────────────────┐
│                      PROPERTY TYPE                              │
│   Addressable inventory with verifiable ownership (technical)   │
│                                                                 │
│   website, mobile_app, ctv_app, desktop_app, dooh,             │
│   podcast, radio, linear_tv, streaming_audio, ai_assistant     │
└─────────────────────────────────────────────────────────────────┘

┌─────────────────────────────────────────────────────────────────┐
│                      FORMAT CATEGORY                            │
│         How the ad renders (orthogonal to channel)              │
│                                                                 │
│     audio, video, display, native, rich_media, text            │
└─────────────────────────────────────────────────────────────────┘

Implementation

JSON Schema

Channels are defined in the AdCP schema at:

/schemas/v3/enums/channels.json

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "$id": "/schemas/enums/channels.json",
  "title": "Media Channel",
  "description": "Standardized advertising media channels describing how buyers allocate budget",
  "type": "string",
  "enum": [
    "display",
    "olv",
    "social",
    "search",
    "ctv",
    "linear_tv",
    "radio",
    "streaming_audio",
    "podcast",
    "dooh",
    "ooh",
    "print",
    "cinema",
    "email",
    "gaming",
    "retail_media",
    "influencer",
    "affiliate",
    "product_placement",
    "sponsored_intelligence"
  ]
}

Usage in Product Discovery

When filtering products by channel:

{
  "filters": {
    "channels": ["ctv", "olv", "streaming_audio"]
  }
}

Usage in Property Definitions

Properties in adagents.json declare which channels they support:

{
  "properties": [
    {
      "property_id": "youtube_app",
      "property_type": "ctv_app",
      "name": "YouTube CTV",
      "supported_channels": ["ctv", "olv", "social"],
      "identifiers": [
        {"type": "roku_store_id", "value": "12345"}
      ]
    }
  ]
}

Usage in Product Definitions

Products declare which channels they are sold as. This typically inherits from properties but may be more specific:

{
  "product_id": "youtube_ctv_premium",
  "name": "YouTube CTV Premium",
  "channels": ["ctv"],
  "publisher_properties": [
    {
      "publisher_domain": "youtube.com",
      "selection_type": "by_tag",
      "property_tags": ["ctv_apps"]
    }
  ]
}

Even though YouTube properties support ["olv", "social", "ctv"], this product is specifically sold as CTV inventory.

Extensibility

Implementations MAY support additional channels beyond this specification using the ext field pattern:

{
  "channel": "display",
  "ext": {
    "sub_channel": "native_content"
  }
}

Versioning

This taxonomy follows Semantic Versioning:

• MAJOR: Removing channels, changing channel semantics
• MINOR: Adding new channels (append-only)
• PATCH: Clarifying descriptions, fixing typos

Adding new channels is a MINOR version change and MUST NOT break existing implementations.

Migration from Legacy Channel Values

Prior to this taxonomy, AdCP used a different set of channel values. The following table maps legacy values to the new taxonomy:

Legacy Value	New Channel(s)	Notes
display	display	No change, but now excludes video
video	olv, ctv	Split by viewing environment
audio	streaming_audio, podcast, radio	Split by audio type
native	display	Native is a format, not a channel
web	display, olv	Web is a substrate, not a planning bucket
mobile_app	display, olv, gaming	App is a substrate, not a planning bucket
dooh	dooh	No change
ctv	ctv	No change
podcast	podcast	No change
retail	retail_media	Renamed
commerce_media	retail_media	Renamed
social	social	No change
sponsorship	product_placement	Renamed and refined

Migration Guidance

1. Identify the planning context: Determine HOW the buyer allocates budget, not where ads technically render.

2. Split video by environment: If video was a single category, split into olv (desktop/mobile) and ctv (TV screens).

3. Remove substrate channels: If you had web or mobile_app as channels, map to planning-oriented channels (display, olv) based on buying context.

4. Update filters: When filtering products, use the new channel values.

Edge Cases and Ambiguities

YouTube Classification

YouTube spans multiple channels depending on context:

Scenario	Channel	Reasoning
YouTube video ad on phone/desktop	olv or social	Depends on buying approach
YouTube video ad on CTV	ctv	TV screen environment
YouTube Shorts	social	Short-form social context
YouTube Music	streaming_audio	Audio streaming context

Retail Media Complexity

Retail media may eventually warrant sub-channels:

Scenario	Current	Potential Future
Sponsored products on Amazon	retail_media	retail_media_search
Display on retailer site	retail_media	retail_media_display
Off-site using retailer data	retail_media	retail_media_offsite

With channels on products, buyers can distinguish retail media product types (sponsored products vs display vs offsite) without sub-channels. Use channels: ["retail_media"] combined with format_ids to narrow results. The two-axis approach (channel + format) is preferred over channel proliferation.

Gaming vs Display/OLV

Scenario	Channel	Reasoning
Rewarded video in mobile game via Unity Ads	gaming	Gaming-specific ad network
Banner in casual game via AdMob general pool	display	Standard mobile programmatic
Esports tournament sponsorship	gaming	Gaming audience context

Influencer vs Social

Scenario	Channel	Reasoning
Buying promoted posts through Instagram Ads	social	Platform ad tools
Contracting an influencer directly	influencer	Creator partnership
Platform-inserted ads around creator content	social	Platform ad tools

Future Considerations

The following channels may be added in future versions based on market evolution:

• messaging - WhatsApp Business, Telegram ads
• xr - VR/AR advertising

References

• IAB Tech Lab Taxonomies - Content, Audience, Ad Product
• OpenRTB/AdCOM - Placement types and media objects
• Planmatic Media Plan Schema - Media planning data structures
• RFC 2119 - Requirement level keywords

Changelog

1.1.0-draft (2026-03-13)

• Rename ai_media to sponsored_intelligence — Sponsored Intelligence is the umbrella for AI platform advertising (AI assistants, AI search, generative AI experiences)
• 20 channels defined

1.0.0-draft (2026-01-23)

• Initial draft specification
• 19 channels defined (planning-oriented approach)
• 10 property types defined
• Clear distinction between channel (planning abstraction), property_type (technical surface), and format
• Multi-channel support for properties
• Migration guide from legacy values