Contents
Key findings
- Two YouTube buys running against the same audience definition shared only 5% of their unique reach, measured by Nielsen Digital Ad Ratings.
- On-target rate was 46% higher on the contextual buy than on the audience-targeted buy.
- The audience-targeted buy came in below the unexposed control on on-target rate — it reached fewer in-target people than a random sample of the population would contain.
- The two methods delivered against visibly different content: home and property videos on one side, celebrity news, gaming and streamer clips on the other.
Advertisers who already run audience-targeted YouTube campaigns ask a reasonable question before adding a contextual buy: won’t the two compete for the same people? The concern is that a second line item simply re-buys reach you already own, adding frequency instead of audience. We ran both against the same brief and measured the overlap.
How much audience overlap is there between contextual and audience-targeted YouTube buys?
Very little. Cross-media reach for the two buys was measured with Nielsen Digital Ad Ratings, which deduplicates unique users across campaigns. Overlap came to 5% of total unique reach. The other 95% of people reached by the contextual buy had not been reached by the audience-targeted one.
That is a larger result than it first appears. Both campaigns ran on YouTube, in the same market, at the same time, with the same creative and the same intended audience. The only variable was how inventory was selected. Two selection methods pointed at the same audience definition landed on almost entirely different people.
For planning, that reframes the question. A contextual line is not a substitute for an audience line competing for the same impressions; in this test it behaved as incremental reach.
Which buy reached more of the intended audience?
The contextual buy, by 46%. On-target rate — the share of people reached who fall inside the intended audience — was measured by a third-party research service across three groups: those exposed to the contextual buy, those exposed to the audience-targeted buy, and an unexposed control.
| Group | On-target rate | vs unexposed |
|---|---|---|
| Contextual buy (GP) | ~29% | +21% |
| Audience targeting (TrueView) | ~20% | −18% |
| Unexposed control | ~24% | — |
The control group is the part worth pausing on. It represents the natural incidence of the target audience in the population — what you would get by reaching people at random. The audience-targeted buy scored below it.
We are not claiming that audience targeting systematically underperforms random reach. This is one campaign, one category, one market, and the differences are within the range that survey-based measurement can produce by chance. The result that matters is the relative one: on this brief, selecting inventory by content reached more of the intended audience than selecting it by audience signal.
What did the two buys actually run against?
Different content, which is the mechanism behind the overlap figure. The brief was a category with everyday purchase relevance and a clearly defined persona — home and moving.
| Top placements included | |
|---|---|
| Contextual buy | Custom-home room tours, small-footprint house builds, moving-day vlogs, storage and decluttering, apartment tours, home-goods hauls |
| Audience-targeted buy | Celebrity news and gossip compilations, forum-thread narration channels, music covers, game streams, VTuber clips, fitness stunts |
Both lists are defensible in their own terms. The audience-targeted buy found people whose behavioural profile suggested they were in market and served them wherever they happened to be watching, which for a general audience is entertainment. The contextual buy served people at the moment they were watching content about the category.
Those are two different propositions, and the 5% overlap is the arithmetic consequence of them. Whether the moment matters more than the profile depends on the objective — but you cannot get both from one line item, which is the practical argument for running them together.
What this changes in a plan
- Treat the contextual line as incremental, not competing. If overlap is in single digits, splitting budget across both methods buys audience rather than frequency.
- Measure overlap rather than assuming it. Deduplicated cross-media measurement is the only way to know; platform-side reach reports will not tell you.
- Include an unexposed control in on-target measurement. Without a baseline, an on-target rate is a number with nothing to compare it to.
- Read the placement report as evidence, not admin. The content each method selected explains the overlap figure better than the overlap figure explains itself.
Frequently asked questions
Does contextual YouTube advertising cannibalise existing audience-targeted campaigns?
In this test, no. Deduplicated cross-media measurement put the overlap in unique reach at 5%, meaning 95% of the people reached by the contextual buy had not been reached by the audience-targeted campaign running alongside it.
What is on-target rate and how is it measured?
It is the share of people a campaign reached who fall inside the intended audience definition. It is measured by a third-party panel or data provider rather than by the ad platform, so that the party selling the impressions is not also grading them.
Why would audience targeting reach fewer in-target people than an unexposed group?
Audience segments are inferred from behavioural signals, and inference degrades — shared devices, stale intent, and broad segment definitions all dilute it. That said, a single campaign scoring below its control is not proof of a general pattern; it is a reason to measure your own.
Should contextual replace audience targeting on YouTube?
The overlap result argues the opposite. If the two methods reach different people, replacing one with the other trades audience for audience. Running both extends unique reach, which is usually the point of adding a second line.
Method
- Design: two YouTube campaigns run in parallel with identical creative, identical audience definition and identical budget. The only variable was the inventory selection method — audience targeting (TrueView) against contextual placement selection (GP).
- Category: a high-frequency purchase category with a clearly defined persona (home and moving).
- Reach overlap: Nielsen Digital Ad Ratings cross-media measurement, deduplicated unique users.
- On-target rate: third-party measurement across exposed and unexposed groups. Absolute values in the table are approximate; the 46% relative gap is the reported figure.
- Limits: one advertiser, one category, one market, one flight. Directional, not a generalised benchmark. Survey-based on-target measurement carries sampling error that the published figures do not state.
Source: GP Inc., parallel-delivery test, Japan, 2025.
Related reading
Adding a contextual line to an existing YouTube plan. GP selects YouTube inventory at video level and reports what it bought, so overlap and on-target rate can be measured rather than assumed. Talk to your local team.
