Unclassified inventory: what should a buyer do with unlabeled placements?
Part of Brand suitability
The short answer
Unclassified inventory is ad space a classifier has not rated yet, because the site is new, its URLs are hidden, its language is not covered or there was too little data. It is unknown, not unsafe. Buyers handle it with a written rule per cause: hold it, review sample pages by hand, or allow it with a note, then recheck.
Key takeaways
- Unclassified means a classifier made no call, so it says nothing yet about what the page contains.
- The cause decides the rule: a new site, a hidden URL, an uncovered language and a reporting gap each need a different response.
- A strict brand can exclude unknowns, but that choice belongs in the written tolerance with its reach cost stated, not in a default checkbox.
- No public source sets a normal share of unclassified inventory, so each buyer measures its own baseline from report dimensions and watches the trend.
What counts as unclassified inventory?
Definition
- Unclassified inventory
- Placements that a brand safety or suitability classifier has not assigned to any rating, so they carry an unknown status instead of a safe, suitable or unsuitable one.
The Media Rating Council's 2018 supplement on content level brand safety (opens in a new tab) says that content which does not meet a verifier's minimum evidence rules "must be classified as unknown or not yet classified" and that "buyers should have the ability to prevent ad serving to unknown or unclassified content." Platforms put that status into their settings under their own names.
| Source | Name used | What the source says it means |
|---|---|---|
| Google Ads content suitability (opens in a new tab) (Display Network and Google video partners) | Not yet labeled | Content "that hasn't been labeled yet because they haven't yet completed the classification process" |
| Display & Video 360 digital content labels (opens in a new tab) | Not yet labeled | "Content that has not been labeled", with causes listed on the same page |
| Display & Video 360 brand suitability (opens in a new tab), third party verification settings | Unrateable | "Impressions that are unrateable can't be described as being brand safe or not" |
| MRC 2018 supplement, sections 3 and 3.1 | Unknown or not yet classified | Content below a verifier's minimum thresholds, or properties it cannot measure directly |
Why is unclassified content not the same as unsafe content?
An unsafe rating is a finding about content. An unclassified status is the absence of one. The MRC supplement treats it as a disclosure duty, saying a verifier that cannot measure part of a property "may list the property or portions of the property as uncategorized or unknown with clear disclosure of capabilities or limitations" so a buyer can avoid it if it chooses.
Impressions that are unrateable can't be described as being brand safe or not.
Why does a placement end up unclassified?
Each cause has its own fix and its own risk. Display & Video 360 lists three causes for Not yet labeled, the MRC supplement adds two, and the last row is a reporting gap that only looks like one.
| Cause | Source wording | What it says about the page |
|---|---|---|
| New site | DV360: "when a site is new, we don't have enough data on the site to classify it"; new sites normally receive a label a month later | Nothing yet; recheck after the next monthly update |
| Hidden URL | DV360: some "publishers don't make the URL of their impressions transparent" | The buyer cannot see which page ran, which is a supply question as much as a content one |
| Uncovered language | DV360: "A site is in a language not supported by digital content labels" | Nothing about the content; a fluent reader can judge it |
| Too little evidence | MRC 3.1: content not meeting "baseline guardrails" is unknown or not yet classified | The verifier saw too few observations to make a reliable call |
| App without measurement access | MRC section 6: apps without API support for content measurement "may be classified as 'unknown'" | The app may be fine; the verifier cannot see inside it |
| Removed video in a report | Google Ads: a deleted video shows as "not available" in the placement report, with impressions still counted | A reporting gap after the fact, not a missing suitability rating |
How do I handle an unlabeled or unclassified placement?
Write one rule per cause before launch, so a planner applies a decision already agreed with the client instead of improvising one. Three rules cover the ground. Hold keeps the placement out until something changes. Review sends a sample of its pages to a person. Allow keeps it in the plan with the gap labeled and a recheck date.
| Cause | Brand with a low tolerance | Brand with a medium or high tolerance |
|---|---|---|
| New site | Review sample pages; hold until the review is done | Allow with a note; recheck after the next label update |
| Hidden URL | Hold; ask the seller for a transparent source or a named deal | Review the inventory source; hold if the seller cannot name the site |
| Uncovered language | Review by a fluent reader before any spend | Allow after a fluent reader samples recent pages |
| Too little evidence | Review by hand | Allow; the gap closes as observations build |
| App without measurement access | Hold unless the app is reviewed by hand | Review the app listing and recent content |
Each rule has a cost. Hold is simple to audit but its cost falls on new and small publishers. Review puts a person's judgment on the page but is the slowest. Allow keeps reach but accepts a known gap, so it needs the note and the recheck date.
Should a strict brand block unclassified inventory?
It can, and the tools allow it. Display & Video 360 notes that one verification partner's settings let buyers "block unrated URLs from being targeted," and Google Ads offers Not yet labeled as a label to exclude on the Display Network and Google video partners. Google says that as of September 2024 YouTube ads no longer use digital content labels, and that sensitive content exclusions apply to the Display Network and Google video partners, so the Not yet labeled choice does not reach YouTube.
A blanket block suits a brand with a low tolerance in several categories and reach to spare. It is a weak call when the plan depends on niche, local or new publishers, the sites the causes table points to. If the brand blocks unknowns, the tolerance document should say so and name who reviews the blocked list. Google's help pages gave two paths when checked in October 2026, Tools > Content suitability and Admin > Account settings; confirm in the account.
How much unclassified inventory is normal?
No standard, platform page or trade body report checked for this guide publishes a normal share of unclassified inventory, so ask for the method behind any quoted norm. The useful number is the buyer's own baseline, measured the same way each month.
Measure an unclassified baseline in four steps
- Pull impressions by label. Display & Video 360 reports include a Digital Content Label dimension, defined as "Classification of content according to brand safety," alongside Domain and Inventory Source.
- Split out reporting gaps. Remove placements that are missing for reporting reasons, such as removed videos shown as not available or placements below Google's minimum reporting thresholds.
- Group the rest by cause. Count unknown impressions by inventory source and domain, so a single opaque source does not hide inside the total.
- Track the trend. Compare the share month to month and flag a rise to the team that owns supply and check first for a new inventory source or a change in URL transparency.
How do you reduce unclassified inventory?
Ways to shrink unclassified inventory
- Sources that pass hidden or partial URLs have been asked for full URL transparency, or moved to a deal limited to named sites.
- New publishers on the plan are rechecked after the next monthly label update instead of being dropped on the first pass.
- Sites in languages the classifier does not cover have a fluent reviewer and a recorded decision.
- The verification partner has disclosed where it cannot verify content and why, as TAG's Brand Safety Certified Guidelines require of certified content verification services.
- An inclusion list of reviewed sites carries the core of the budget, so unknowns sit at the edge of the plan rather than its center.
- Publishers who believe they were rated wrongly have a route to ask for review, which certified verification services must offer.
TAG's Brand Safety Certified Guidelines (opens in a new tab), version 5.0 of July 2026, section 4.15, say a certified content verification service "must disclose to their clients instances where content verification is not possible as well as the reason(s) why content verification was not possible," and must disclose "the frequency it classifies and reclassifies content." Sites that cannot be checked at all are covered in the guide to unverifiable sites in a media plan.
What do unclassified rules look like on a niche retail plan?
Fictional example
A model railroad and hobby retailer in northern New England
A fictional model railroad and hobby supply retailer sells through two stores and online to customers in Maine and New Hampshire, around Portland, Bangor and Manchester. Its plan leans on layout building blogs, railroad history sites and club pages, in English. The first label report shows a sizable group of these domains as Not yet labeled, a label the previous agency had excluded by default.
The planner groups the unknowns by cause. Several blogs are new, so they stay in with a note and a recheck after the next monthly update. One inventory source passes no page URLs; it is held, and the seller is asked to offer the sites as a named deal. Two club sites run open member forums, so a reviewer reads recent threads before they return. The tolerance document now records the rule for each cause and the person who owns the monthly recheck.
How does PlacementGraph treat a site it could not check?
Frequently asked questions
Does the Not yet labeled setting apply to YouTube campaigns?
No. Google says that as of September 2024 YouTube ads no longer use digital content labels, so excluding Not yet labeled affects the Display Network and Google video partners only. On YouTube, buyers rely on the inventory type, excluded content themes and placement exclusions.
Is an unclassified site more likely to be a made-for-advertising site?
The sources on this page do not say so. Unclassified describes a gap in a content classifier's coverage, while the ANA's 2023 definition of made-for-advertising sites rests on ad density, refresh, paid traffic and templated content. A new independent blog and an arbitrage site can both be unlabeled, so check site quality separately.
Are apps more often unclassified than websites?
The MRC supplement explains why they can be. The MRC notes that many applications do not support API functionality for content measurement, and says apps without that support may be classified as unknown. Classifying an app from its store description is allowed with disclosure, but is not content level measurement.
Can a publisher ask a verification service to classify its site?
Under TAG's Brand Safety Certified Guidelines, a certified content verification service must have an objective review process for claims of wrongful classification as brand unsafe. A publisher stuck at unknown can ask the buyer which service rated it, request a review there and share full page URLs.
Sources
- MRC Supplement to IAB Guidelines for the Conduct of Ad Verification: Enhanced Content Level Context and Brand Safety (opens in a new tab)Media Rating Council, September 2018
- About content suitability (opens in a new tab)Google Ads Help
- Digital content labels in Display & Video 360 (opens in a new tab)Display & Video 360 Help
- Brand suitability (opens in a new tab)Display & Video 360 Help
- Dimensions in reports (opens in a new tab)Display & Video 360 Help
- About the Content suitability report (opens in a new tab)Google Ads Help
- Brand Safety Certified Guidelines, version 5.0 (opens in a new tab)Trustworthy Accountability Group, July 2026
Researched and written by PlacementGraph Research. Read how we research, source and update these pages. Corrections: diana@placementgraph.com.