What does content adjacency mean on a page, in a video and in a feed?
Part of Brand suitability
The short answer
Content adjacency is how close an ad appears to a specific piece of content. On a web page it can mean the page around the ad; in video, the host video and those before or after it; in a feed, the posts above and below. The MRC measures it as physical distance in disclosed units.
Key takeaways
- The MRC defines adjacency as the physical distance between an ad and specific classified content, and counts content as directly adjacent when it can be on screen with the ad at the same time.
- On display, adjacency is judged at the page or URL in the MRC's site context model, so the classification of the whole page stands in for what sits beside the ad.
- In video, the host video and the videos that play before and after, including autoplay, are the adjacent content; thumbnails count too.
- In feeds, adjacency is counted in discrete units such as posts, tiles or screens, and the MRC notes that blocking on it may be possible only through the ad server, which may be the platform itself.
- An adjacency report is only as useful as its stated unit, distance basis and coverage gaps.
What is content adjacency in advertising?
Definition
- Content adjacency
- The degree of physical distance between an ad and specific content that has been classified for context or brand safety, within a site, page, app or platform. The closer the content, the more a viewer is likely to read the two together.
A simple illustration is a car ad printed beside a story about a fatal crash. Digital media made the question harder, because the content beside an ad can change with each load, scroll and autoplay. The Media Rating Council's Enhanced Content Level Context and Brand Safety supplement (opens in a new tab), issued in September 2018, gives the MRC's definition, and it explicitly names newsfeed, continuous scroll or swipe, and continuous play environments as the places that need finer proximity measurement.
| Term | Meaning | Where it comes from |
|---|---|---|
| Measurement unit | The discrete piece of content distance is counted in, such as a post, tile, pane, screen or video | MRC 2018 supplement; each platform must disclose its unit |
| Directly adjacent | Content zero units from the ad, or able to be viewable on screen at the same time as the ad | MRC 2018 supplement, using the 50 percent of pixels viewability test |
| Adjacency range or score | A defined band of distance beyond direct adjacency, such as two to five units away | Optional in MRC guidance; must be clearly defined |
| Temporal threshold | A time limit after which nearby content is treated as having less effect | Encouraged, not required, by the MRC supplement |
| Minimum adjacency rule | A platform rule that ads are not placed within a set number of units of some content | Platform monetization controls the MRC encourages platforms to disclose |
Adjacency is one dimension of suitability, not a separate set of categories. The brand still decides what kinds of content it avoids; adjacency decides how far away that content has to be. Our guide to brand safety categories covers the first question, and this glossary covers the second.
How is adjacency measured for display ads?
For display ads on the open web, the measurement unit is the page or the domain. A verification service or a planning tool classifies the URL the ad loads on, or the whole domain, and treats everything on it as adjacent. The MRC's 2012 Guidelines for the Conduct of Ad Verification (opens in a new tab) built that model: site context is judged from attributes of the web page or site the ad is served to, against parameters the advertiser sets.
- Domain level: one label for the whole site. Fast and broad, but it treats a recipe page and a crime report on the same news site the same way.
- URL level: one label for the page the ad loaded on. The 2018 supplement asks content level services to classify text, images, video and audio at this level.
- On screen proximity: whether flagged content is visible at the same time as the ad. The MRC treats content that can be viewable alongside the ad, meaning 50 percent of its pixels on screen, as directly adjacent.
- Rolled up labels: when a URL is not yet classified, some services apply the domain's label. The MRC's October 2025 policy requires property level services to disclose when they do this.
What counts as adjacent content in video and feeds?
Video and feeds move the content past the viewer rather than leaving it on one page, so the MRC asks for different units. In video playlist environments, the 2018 supplement says adjacency should be measured as the number of video units that play before and after the content, "including content that automatically begins playing without additional user interaction." It also asks services to consider thumbnail images, especially where a thumbnail misleads about the video beneath it.
| Environment | Adjacent content | Unit |
|---|---|---|
| In-stream video | The host video the ad plays in, before, during or after | The video, or a segment of a long video |
| Video playlist or autoplay | Videos that play before and after the ad's host video, including autoplay | Number of video units away |
| Vertical short video feed | Videos shown before and after the ad as the viewer swipes | One video per screen |
| Scrolling social or news feed | Posts above and below the ad, including images and comments | A post, tile or pane |
| Display page | The page or the content visible beside the slot | The URL, or the screen |
Long videos add a further wrinkle. The MRC notes that one video can contain segments with different classifications, so a single label for a long video may hide a short unsuitable section. For feeds, the supplement says adjacency must account for scrolling and swiping on mobile and should be reported as a count of units above or below the ad, with the basis disclosed.
Google's Content suitability report (opens in a new tab) is one public example of feed adjacency reporting. It shows placement data for content shown before and after YouTube Shorts ads. Google also states that the report does not support Watch Next feed adjacency or the Home feed, that counts may not be deduplicated for Shorts, that it may include invalid traffic, and that its impression counts are not the impressions billed.
Why does adjacency risk differ by ad format?
Three things change from format to format: how tightly the ad is bound to the content, how fast the content changes, and who is able to see both at once. The table sets them side by side for the formats in a mixed plan.
| Format | How tightly ad and content are linked | How fast the content changes | Who can measure and block |
|---|---|---|---|
| Display on a publisher page | Loose; the ad sits beside the content | Slowly; the article stays the same once published | The buyer's verification or planning tools can read the page |
| In-stream video | Tight; the ad plays inside the content | Not at all for one video, but each play is a different video | Classification of the host video, which may rely on the platform's own signals |
| Autoplay and playlists | Medium; the ad sits in a sequence | With each autoplay step | Mostly the platform, which decides the next video |
| Scrolling feeds | Medium; the ad sits between posts | With each scroll and refresh | The ad server, which may be the platform itself, per the MRC supplement |
The last column is the key one for feeds. The MRC supplement observes that third party verification services may have to rely on adjacency signals supplied by the platform, and that "only the ad server, often times the platform itself, can allow blocking, covering or otherwise avoidance based on adjacency." It requires either direct third party measurement or validated first party data for content level adjacency accreditation. Platform controls also differ by surface: Google states that its content exclusions may give limited coverage of YouTube in-feed ads and YouTube Shorts, as the content suitability help page (opens in a new tab) explains.
How should adjacency be reported to a brand?
An adjacency report should let a brand answer one question: how often did our ads sit close to content we said we avoid, and how close? That is only possible when the report states its rules. The MRC says the measurement unit and the basis for reporting must be clearly disclosed, and that any ranges or scores must be clearly defined.
What an adjacency report should state
- The measurement unit for each environment, such as URL, post, screen or video.
- How distance is counted, and what counts as directly adjacent.
- Whether adjacency signals came from the platform or were measured independently.
- Which surfaces are not covered, such as a feed the report does not support.
- Whether any time threshold was applied to nearby content.
- Which suitability categories and tolerance levels the adjacency counts refer to.
- Whether counts are deduplicated and whether they match billed impressions.
Fictional example
Adjacency rules for a regional airline
An unnamed regional airline flying out of Seattle and Portland plans a summer fare campaign across news sites, YouTube in-stream and YouTube Shorts in Washington and Oregon. Its profile accepts general travel and local news but excludes content about air crashes and aviation incidents.
The agency writes the rule three ways. On display, any URL classified as aviation incident coverage is excluded, and the airline accepts that other stories on the same news site remain eligible. On in-stream video, the host video must not be about an aviation incident, and long news videos are reviewed by segment where the partner can do so. On Shorts, the agency explains that the brand cannot pick the videos before and after the ad, reads the before and after placements in Google's Content suitability report each week, and records in the plan that the Home and Watch Next feeds are not covered by that report.
How does PlacementGraph look at the content around a placement before the buy?
Frequently asked questions
Is content adjacency the same as context?
No. Context is what a piece of content is about and how it treats its subject. Adjacency is how close an ad sits to that content. A brand needs both answers: which contexts it avoids, and whether those contexts must be absent from the whole page, the same screen or only the post directly beside the ad.
Is adjacency the same as competitive separation?
They are related but different. The 2012 MRC and IAB ad verification guidelines list competitive separation as its own service line, about keeping an ad away from a competitor's ads. Content adjacency is about distance from editorial or user content, measured against the brand's suitability categories.
Can a brand choose the videos that play before and after its ad?
Not in feeds and playlists that the platform sequences. The platform decides the sequence, so a brand's controls are the platform's own settings, any minimum adjacency rules the platform discloses, and after the fact reports of content shown before and after the ad. Channel and video inclusion lists narrow the host video, not its neighbors.
Do comments under an article count as adjacent content?
They can, when they appear near the ad slot. The MRC supplement treats unmoderated user generated content as a distinct challenge, and comments change after classification. Ask whether a partner's page classification includes comments, and consider placements where comments sit far from the ad or are moderated.
Sources
- MRC Ad Verification Supplement: Enhanced Content Level Context and Brand Safety (September 2018) (opens in a new tab)Media Rating Council
- Guidelines for the Conduct of Ad Verification (February 14, 2012) (opens in a new tab)IAB and Media Rating Council
- Policy for Property Level Ad Verification Representations (October 18, 2025) (opens in a new tab)Media Rating Council
- About the Content suitability report (opens in a new tab)Google Ads Help
- About content suitability (opens in a new tab)Google Ads Help
Researched and written by PlacementGraph Research. Read how we research, source and update these pages. Corrections: diana@placementgraph.com.