Contents
Key findings
- Using affinity targeting on Shorts, 80% of impressions were reported as “Other” — a placement the advertiser cannot inspect. With topic targeting it was 60%.
- Buying the same inventory by placement instead reduced unverifiable impressions to about 1% across 9 million impressions.
- The single most-served placement in the affinity test was a video that had since been removed or made private.
Short-form inventory on YouTube is bought the same way as long-form, but the reporting behaves differently. We ran a test to answer a question advertisers ask and rarely get an answer to: when you buy Shorts, where does the ad actually appear? For most of the budget, the honest answer turned out to be that nobody can tell.
A minority. In the affinity-targeted test, roughly 65,000 impressions were delivered and about 51,000 of them — 80% — were grouped under “Other” in the placement report. Under topic targeting, about 2,100 of 3,500 impressions, or 60%, fell into the same bucket.
| Targeting method | Impressions | Reported as “Other” | Share |
|---|---|---|---|
| Affinity | ~65,000 | ~51,000 | 80% |
| Topic | ~3,500 | ~2,100 | 60% |
| Placement targeting (GP) | 9,000,000 | ~90,000 | ~1% |
Once an impression is grouped as “Other” there is no way to find out what it ran against. That is not only a brand-safety question. It means the largest share of the budget produces no information that can be used to optimise the next flight.
Unverifiable placements occur on long-form inventory too. In this test the share was higher on Shorts than in comparable long-form affinity campaigns we have run.
What did the visible placements look like?
Not what the targeting implied. The test assumed a beauty and cosmetics advertiser, with affinity and topic segments set to match. The most-served placement under affinity targeting was a video marked unavailable, meaning it had been deleted or made private after delivery. Others included sensational clips and unauthorised re-uploads of anime and drama.
Topic targeting delivered more content that was genuinely about beauty. Its most-served placement was a video of a man reacting in pain to hair-removal laser treatment. Unrelated content also appeared, including footage of a crocodile in a lake.
None of this indicates the platform malfunctioning. Topic targeting improved subject relevance over affinity, as designed. The gap is between “related to the topic” and “somewhere a beauty brand would choose to appear”.
Why does Shorts produce this kind of inventory?
Because of what the format rewards. Most Shorts run under sixty seconds and can be skipped instantly, so creators optimise for immediate attention. That selects for content designed to arrest a scrolling viewer, and a portion of it is later removed by the platform for guideline breaches or rights claims — which is why a deleted video can still appear as a top placement in a report.
How do you buy Shorts with visibility?
By selecting the videos rather than the audience. Placement targeting delivers only against a specified list of video IDs, so the report shows what was bought. In one Shorts campaign run this way, about 1% of nine million impressions fell into the unverifiable bucket, against 80% under affinity targeting.
- Check the “Other” share before judging a Shorts campaign. If most impressions are unverifiable, the placement report is not describing your campaign.
- Treat deleted placements as a signal. A removed video at the top of the report indicates the kind of inventory the targeting is selecting.
- Separate topic relevance from suitability. Topic targeting improved subject match here and still produced placements a beauty brand would reject.
- Report Shorts separately from long-form. The two behave differently enough that combining them hides the problem.
Frequently asked questions
Can you see where YouTube Shorts ads are shown?
Often not. In this test 80% of affinity-targeted impressions and 60% of topic-targeted impressions were grouped as “Other” in the placement report, with no way to retrieve the specific video. Placement targeting reduced that to about 1%.
What does “Other” mean in a YouTube placement report?
It is a bucket for impressions whose specific placement is not disclosed. The advertiser cannot see which video the ad appeared on, cannot exclude it, and cannot use it to optimise later flights.
Is brand safety worse on Shorts than on long-form YouTube?
This test found a higher share of unverifiable placements on Shorts than in comparable long-form affinity campaigns, and the visible placements included sensational and unauthorised content. The measurable difference is transparency; suitability follows from it, because you cannot exclude what you cannot see.
Is topic targeting better than affinity targeting on Shorts?
On subject relevance and transparency, yes: more of the placements were about the intended topic, and 60% were unverifiable against 80%. Neither reached a level where the advertiser could see most of what they bought.
Method
- Test design: a hypothetical beauty and cosmetics advertiser, with affinity and topic segments set to cover comparable subject matter.
- Scale: approximately 65,000 impressions under affinity targeting and 3,500 under topic targeting. The placement-targeting comparison is drawn from a separate campaign of 9 million impressions.
- The three figures are not a controlled comparison. They come from different campaigns at different scales and should be read as an indication of the transparency gap, not as a matched test.
- “Other” is the platform’s own reporting category, not a classification applied by us.
Source: GP Inc., YouTube Shorts placement test, March 2025. Campaigns run in Japan.
Related reading
- Contextual and audience-targeted YouTube buys shared only 5% of their reach
- Keyword targeting vs contextual targeting on YouTube
Buying video-level placements on Shorts. GP extracts videos by ID and delivers against that list, so the placement report shows what was bought. Talk to your local team.
