Ad Inventory Types Guide: Display, Video & Native

Reference

“Ad inventory” is the supply of advertising opportunities made available by publishers and platforms: a slot on a web page, an ad break in a video stream, a native placement in a feed, an interstitial in an app, or an audio break in a podcast. In day-to-day buying and selling, inventory is classified so that buyers can understand what they are purchasing (format, placement context, device environment, and measurement), and sellers can describe supply consistently across systems.

This overview provides a practical, standards-aligned way to describe common online advertising inventory types, with emphasis on how inventory is represented in modern ad-tech (especially programmatic), rather than on marketing outcomes. Where possible, terminology follows industry and technical standards such as IAB Tech Lab specifications (e.g., OpenRTB and native/video guidelines) and related technical work that references “advertising inventory” in a web context.

Scope: This page focuses on inventory types and placement contexts (where ads can appear), not campaign objectives, creative strategy, or vendor comparisons. Examples are illustrative and not exhaustive.

If you’re applying this taxonomy to real campaign planning, pair it with the channel-side guides for Google Display Ads and remarketing vs retargeting.

What “inventory type” means in practice

In online advertising, an “inventory type” typically answers a small set of technical questions:

Environment: web (browser) vs in-app vs connected devices (e.g., CTV).
Format: display, native, video, audio, rich media, and hybrids.
Placement context: in-stream video vs outstream video; in-feed native vs recommendation widgets; interstitial vs rewarded.
Transaction & delivery model: direct-sold placements, programmatic auctions, private marketplace deals, etc.
Measurement hooks: what counts as an impression, and what tracking and verification signals exist.

Specifications like OpenRTB exist because these distinctions must be communicated reliably between systems (buyers, sellers, exchanges, DSPs, SSPs). For example, OpenRTB defines structures for describing impressions and includes objects tailored to formats such as banner, video, audio, and native.

The rest of this article groups inventory by the environments and placements that are most commonly recognised across the market: desktop web, mobile web, mobile apps, video, native, email, feeds (including RSS-like distribution), and messaging-style environments.

Core inventory environments

1) Desktop web inventory (browser-based)

Desktop web inventory refers to ad placements rendered in a browser on laptop/desktop devices. The typical unit of inventory is a placement embedded in page layout: header, sidebar, in-article, footer, or overlays. In programmatic trading, these placements are usually represented as banner-type impressions (for display/rich media) or as video/native objects when those formats are rendered on the page.

Common desktop-web placements include:

Standard display placements: fixed areas in site layout that load a creative (image, HTML5, rich media).
In-article / in-content placements: units inserted within content flow, often with viewability constraints.
Overlays / interstitials: placements that cover a portion of the page or appear between navigations (subject to policy constraints on many properties).
Native placements rendered on web: in-feed or recommendation-style units that match the page design (covered further below).

Desktop web inventory is often easier to describe consistently because layout is relatively stable compared with mobile experiences, but it is also subject to significant variation in viewability, refresh behaviour, and ad-blocking prevalence.

2) Mobile web inventory (browser-based on mobile devices)

Mobile web inventory is served within mobile browsers. It resembles desktop web in that the environment is browser-based, but placement behaviour differs due to smaller screens, scrolling patterns, and the prevalence of sticky units and anchor placements. The same high-level formats apply (banner, video, native), but placement constraints tend to be tighter (e.g., minimum player sizes, interaction rules, and layout constraints).

Typical mobile-web placements include:

In-article units: display or native placements within content flow.
Sticky banners: anchored to top or bottom of the viewport.
Mobile interstitials: full-screen overlays (often constrained by policy and user experience guidelines).
Outstream video: video that plays outside a dedicated video player context (see “Video inventory types”).

Mobile-web inventory can be more sensitive to latency and user experience. As a result, many buyers and sellers treat mobile-web placement context (especially interstitial-like experiences) as a distinct class of inventory even when the underlying format is “banner”.

3) In-app inventory (mobile applications)

In-app inventory is delivered inside mobile applications (iOS/Android). The inventory is still typically described in terms of formats, display/banner, interstitial, native, video, and rewarded, but the rendering and measurement mechanisms differ from the browser. In programmatic systems, the impression is described similarly (banner/video/native objects), but app identifiers, SDK mediation, and device signals often shape how inventory is packaged and sold.

Common in-app inventory types include:

Banner: persistent or embedded display placements within app UI.
Interstitial: full-screen placements shown at transitions or natural breaks.
Rewarded video: opt-in video views that unlock an in-app benefit (most common in gaming apps).
Native: ad units integrated into feeds or lists within the app’s UI.
App open: placements shown when the app launches or returns to foreground (platform- and policy-dependent).

A key technical distinction is that in-app inventory is commonly mediated via SDKs and may be routed through multiple demand sources. This increases the importance of consistent classification, because the same user-visible placement can be filled by different demand paths.

Format families and placement contexts

4) Display and rich media inventory

“Display” inventory generally refers to image or HTML5 creatives rendered in a defined rectangle or container (sometimes called “banner” inventory in programmatic protocols). Rich media extends display with interactive elements, expansion, or additional assets. In OpenRTB, banner is one of the primary impression objects used to communicate this type of inventory.

In practice, display inventory is often sub-classified by placement behaviour:

Fixed placement: a predictable slot in layout (e.g., top leaderboard, sidebar rectangle).
In-content placement: inserted within reading flow, often scroll-dependent.
Sticky / anchor placement: persists in viewport while scrolling.
Interstitial / takeover: temporarily occupies most or all of the screen.

The reason this matters is not cosmetic: viewability, interaction, and load behaviour differ significantly across these classes. Two display placements can be the same pixel size yet behave very differently for users and measurement.

5) Native inventory (in-feed and content-integrated)

Native inventory refers to placements designed to match the look and feel of surrounding content. Rather than delivering a single “banner image,” native inventory typically passes structured components (headline, image, description, call-to-action text, advertiser name) that the publisher renders within its UI. The OpenRTB Native specification formalises this structured approach for native ad requests and responses.

Common native inventory placements include:

In-feed units: inserted into content feeds (news feeds, article lists, product lists).
Recommendation widgets: “related content” or “from around the web”-style placements.
Promoted listings: native-style items within marketplaces or directories (where policy allows).

Native inventory is often treated as a distinct category because the creative is rendered by the publisher, which changes measurement, creative approval processes, and brand-safety review. It also blurs boundaries: native can appear on desktop web, mobile web, and in-app environments.

6) Video inventory

Video inventory is commonly divided by whether the ad is served inside a video player experience (in-stream) or outside it (outstream). IAB Tech Lab guidelines describe in-stream video ads as formats served into a video player before, during, or after video content, and provide a shared terminology for describing placement context in digital video and CTV environments.

Within programmatic, video inventory classification is important enough that OpenRTB has added attributes over time to describe video placement more precisely (for example, updates intended to improve clarity around video placement signals).

Common video inventory types include:

In-stream (linear): pre-roll, mid-roll, post-roll within a video player (often with skippable/non-skippable variants).
In-stream (non-linear): overlays within a video player experience (e.g., companion overlays).
Outstream: video units that appear within page or feed content, not within a dedicated video player content stream.
Connected TV (CTV): video inventory delivered to TV-like devices and apps, typically full-screen, with different interaction and measurement constraints.

Practical note: Many buyers treat “video” as multiple inventory categories because in-stream, outstream, and CTV behave differently for completion rates, attention, user control, and measurement.

7) Audio inventory

Audio inventory covers ad placements in audio streams (music, radio-like streaming, podcasts). In OpenRTB, audio is represented as its own impression object, reflecting distinct parameters and playback contexts compared with display or video.

Common audio inventory contexts include:

Streaming audio: inserted during music or radio-style streams.
Podcast ads: dynamically inserted (DAI) or baked-in placements (where the “inventory” is an episode insertion point).

Audio is often sold with different assumptions about attention and interaction. For inventory classification, the key distinctions are typically the playback environment (app vs web), insertion method (dynamic vs baked-in), and whether the user can skip.

Distribution-based inventory: feeds, email, and messaging

8) Feed-based inventory (including RSS-style and content syndication)

Some inventory exists as part of syndicated content distribution rather than a traditional “page.” This includes placements in content feeds, aggregators, and syndication surfaces. In practical terms, many modern feed placements are implemented as native or display units inside an app or web feed, but it is still useful to treat “feed distribution” as an inventory context because it affects rendering and measurement.

Typical examples include:

News and content feeds: in-feed native or display units within continuous scrolling lists.
Syndicated widgets: content modules embedded on third-party pages where the inventory is sold or filled via a network.
RSS-style placements: ad units rendered within feed readers or email digests derived from feeds (less common today, but still present in some ecosystems).

Where feeds are rendered as UI lists, native inventory specifications (structured assets rendered by the publisher) are particularly relevant.

9) Email inventory

Email inventory refers to advertising placements within email messages (newsletter sponsorships, display-like modules, or native-style “promoted” blocks), as well as placements in webmail interfaces (which behave more like web inventory). The ad-tech and measurement constraints differ because emails are rendered by email clients with limited scripting, and because tracking is heavily constrained by client privacy features.

For inventory classification, the most useful distinctions are:

In-email placements: units embedded in the email content itself.
In-interface placements: ads shown in webmail or app interfaces around the inbox (closer to display/native inventory in a platform UI).

Because email environments have unique technical constraints, many buyers treat email sponsorships as a separate inventory type even when the “creative” resembles display or native.

10) Messaging and chat-like environments

“Messaging” inventory is a broad category covering ad placements within chat-like interfaces (messaging apps, community platforms, and other conversational UIs). The placements are often native-style (rendered as cards, suggested content, or sponsored placements in lists) rather than classic banners.

Classification is important here because user expectation and interaction patterns differ: messaging UIs are high-attention, personal contexts, which often leads to stricter policy controls and more limited placement options.

Publisher-sold vs network-aggregated inventory

Inventory types are often discussed as if they exist in isolation, but the same placement can be sold through multiple routes:

Direct-sold: a publisher sells a specific placement directly to an advertiser or agency.
Network-aggregated: inventory is pooled across many publishers and filled by a network or exchange.
Hybrid: premium demand is sold directly, while unsold (“remnant”) supply is made available to programmatic demand sources.

Programmatic standards like OpenRTB exist largely to enable this aggregation at scale by providing a consistent vocabulary and request/response model across participants.

Why this matters: Two placements may look identical to a user, but differ meaningfully in how they are packaged, disclosed, measured, and controlled depending on whether they are direct-sold or network-aggregated.

Ad inventory and measurement constraints

Inventory classification also needs to reflect measurement realities. For example, browser privacy changes, platform controls, and standards work around attribution and advertising measurement affect what signals are available in different environments. The W3C’s work on attribution explicitly references “sites that provide advertising inventory” as part of the web ecosystem, highlighting that measurement is tied to how inventory is provided and how outcomes are observed.

In practical terms, inventory types differ in:

Impression definition: when an impression is counted (render, viewable, video start, etc.).
Viewability and attention: how reliably an ad is in view or audible.
Interaction: click and engagement mechanics (especially limited on CTV and some audio environments).
Attribution signals: what downstream measurement is feasible, given privacy and platform constraints.

As a result, “inventory type” is not merely a taxonomy exercise. It is part of how buyers and sellers set expectations about delivery, verification, and measurement.

Why this taxonomy stays stable even as products change

The set of inventory environments has evolved over time (desktop-first web to mobile to app ecosystems to streaming and CTV), and product names change frequently. However, the underlying categories remain fairly stable because they correspond to durable constraints:

Rendering context: browser vs app vs streaming device.
Creative model: fixed creative vs structured assets rendered by the publisher (native).
Playback model: in-stream vs outstream vs audio-only playback.
Interaction model: direct click vs remote control vs limited input.

For this reason, standards bodies and technical specifications tend to organise around these durable concepts. OpenRTB’s separate objects for banner, video, audio, and native reflect that these formats and contexts require distinct parameters and measurement assumptions.

Summary: the most common inventory types

A compact classification that maps closely to how inventory is described in practice is:

Desktop web: browser-based placements on desktop/laptop sites (display, native, video).
Mobile web: browser-based placements on mobile devices (display, native, video, interstitial-like formats).
In-app: mobile app placements (banner, interstitial, native, rewarded, video).
Video: in-stream, outstream, and CTV placements (distinct playback and measurement contexts).
Native: structured-asset placements rendered in feeds and content modules (across web and app).
Audio: streaming and podcast inventory (distinct playback and interaction constraints).
Email: in-email sponsorship/placements and inbox-interface placements.
Feed/syndication: placements in continuous feeds and distributed widgets (often native-style).
Messaging UI: sponsored placements in chat/community conversational interfaces (typically native-style).

This classification is intentionally conservative: it prioritises durable technical differences (environment, rendering, playback) over fast-changing product labels.

Glossary terms

References

IAB Tech Lab. OpenRTB Version 2.6 (PDF). https://iabtechlab.com/wp-content/uploads/2022/04/OpenRTB-2-6_FINAL.pdf
IAB Tech Lab. Ad Format Guidelines for Digital Video and CTV (PDF). https://iabtechlab.com/wp-content/uploads/2022/03/Ad-Format-Guidelines_DV-CTV.pdf
W3C. Attribution Level 1 (Technical report). https://www.w3.org/TR/attribution/
IAB. Mobile Video Buyer’s Guide (PDF). https://www.iab.com/wp-content/uploads/2015/07/MobileVideoBuyersGuideFinal.pdf
IAB. OpenRTB Dynamic Native Ads API Specification Version 1.2 (PDF). https://www.iab.com/wp-content/uploads/2018/03/OpenRTB-Native-Ads-Specification-Final-1.2.pdf
IAB Tech Lab. Demystifying the OpenRTB Video plcmt Attributes. https://iabtechlab.com/demystifying-openrtb-video-plcmt-attributes/

Reference

Scope: This page focuses on inventory types and placement contexts (where ads can appear), not campaign objectives, creative strategy, or vendor comparisons. Examples are illustrative and not exhaustive.

If you’re applying this taxonomy to real campaign planning, pair it with the channel-side guides for Google Display Ads and remarketing vs retargeting.

What “inventory type” means in practice

In online advertising, an “inventory type” typically answers a small set of technical questions:

Environment: web (browser) vs in-app vs connected devices (e.g., CTV).
Format: display, native, video, audio, rich media, and hybrids.
Placement context: in-stream video vs outstream video; in-feed native vs recommendation widgets; interstitial vs rewarded.
Transaction & delivery model: direct-sold placements, programmatic auctions, private marketplace deals, etc.
Measurement hooks: what counts as an impression, and what tracking and verification signals exist.