How YouTube counts results
YouTube ads are bought through Google Ads, so a YouTube recap uses Google’s attribution rules. The top of the report is usually views, view rate and watch time. Those tell you how the ad played. The results come further down, and they can include people who watched part of an ad without clicking. Google calls some of these engaged-view conversions. They work like a view-through conversion: credit for watching, with no click.
The other thing to know is that a view on YouTube can mean different things. A paid view, a view on your channel and a view on a creator’s channel aren’t the same. Ask whether the view count on a video includes paid views before you compare it with your organic posts.
Find the window in the report. Video is often seen days before someone acts, so the window decides how much later activity YouTube can take credit for.
Where results get counted twice
- Watching, then searching. Someone watches a trailer on YouTube, then searches your title and clicks a search ad. YouTube and search can both take credit for the result, sometimes inside the same Google account.
- Creator videos. A creator’s video with your brand in it shows up in your creator report. If paid ads also ran on that video or that channel, your paid-media report can count the same results. Two reports, one post.
- Long windows on big reach. Video reaches many people, and a long window lets YouTube take credit for anyone who later acts for any reason.
- Mixed campaign types. Some Google campaign types spread across YouTube, search and more. Ask for YouTube on its own.
A simple first step: ask for one month of YouTube results split into clicks and watches, then line them up against your own numbers by day. If results rise on the days a trailer or ad drops, that’s a good sign. If the reported number stays high on quiet days, the window is probably doing a lot of the work.
A checklist for your next YouTube recap
| Look at | What to look for | Why it matters |
|---|---|---|
| Clicks vs. watches | Results split by click and by watched-without-click | Watch-based credit is easy to overcount |
| Window | How many days after a view a result still counts | Video is often seen long before action |
| Paid vs. organic views | Whether view counts include paid views | Paid views can make organic posts look stronger |
| Watch time | How long people kept watching | Tells you whether the ad held attention |
| Overlap with search | Results counted by both YouTube and search | Same account, two credits |
| Comments | What viewers said under the video | The fastest read on how it landed |
Questions worth asking
Ask them of whoever runs the account, in-house or a partner.
- How many of these results came from people who clicked, and how many from people who only watched?
- What window does this report use, and can we see a shorter one?
- Does the view count on our trailer include paid views?
- Do any of these results also show up in our search reports?
- Which videos kept people watching longest, and did those drive more results?
Ask these in your next meeting rather than over email. Answers in the room tend to be clearer, and you’ll hear which ones can be answered on the spot.
Comments are worth a look too. In a real Arcus audience report for a streaming service, fans talked about the shows and almost never about the service itself. Our guide on watched vs. shared explains why views alone can mislead.
What a regions-on, regions-off test looks like on YouTube
YouTube ads can be targeted by location, so the test works the same way as on other channels. Pick regions that look alike, run YouTube ads in some and pause them in others for a few weeks, and compare your own numbers: tickets, sign-ups, streams or sales. That’s a regions-on, regions-off test, also called a holdout test.
Keep search steady during the test, since the two interact. If you pause YouTube and search results fall in the same regions, that tells you something useful about how they work together.
Google also offers lift studies for YouTube, including surveys that measure recall. A brand lift study is good evidence for brand work. It measures what people say in a survey, so it’s strongest when it lines up with your own numbers later.
Want to see what worked for your brand? Try it free: pick a question, and we answer it with your own numbers and what people say online. No card.