Home / Ad testing / YouTube ad testing
YouTube ad testing before launch
YouTube ad testing shows your pre-roll, bumper or Shorts ad to real viewers before it runs and measures how they respond, second by second. On skippable in-stream ads, viewers can skip after five seconds, so the opening matters more than anything else in the edit. EchoDepth plays your ad to viewers matched to your audience, 15 or more per audience segment, each watching at home on their own webcam after giving consent. We code their facial response against 44 FACS Action Units on every frame and plot it second by second as an Emotional Response Curve of valence, arousal and dominance. You see whether the hook earns attention before the skip button appears, where interest drops, and how people react when the brand shows up. Testing two or three openings side by side is the most useful YouTube brief, because the first five seconds decide how many people see the rest.
Updated 17 October 2026. Platform details checked against Google Ads Help: About video ad formats.
YouTube formats you can test
Every format is shown to viewers at its native aspect ratio and length.
- Skippable in-stream
- Viewers can skip after five seconds, according to Google. Test whether your opening holds them past that point.
- Bumper ads
- Non-skippable and six seconds at most. With no time to recover, every second of the curve counts.
- Non-skippable in-stream
- Google lists these as 60 seconds or shorter. Watch for the point where response flattens, because viewers can't leave.
- Shorts ads
- Vertical ads that Google says feel like organic Shorts. We show them at native aspect ratio.
What you can learn
Every finding traces back to a moment in your ad and the facial movements behind it.
- The hook
- Whether the first five seconds lift arousal and valence, or leave faces flat.
- The skip point
- How response looks at the moment the skip button would appear, and whether the ad has earned the next second.
- Brand in the first five seconds
- How viewers react when the logo or product appears early, which matters for viewers who skip.
- The drop-off
- Where the curve falls away in longer edits, so you know what to cut.
- Edit comparison
- Which of two or three openings performs best, with each viewer seeing only one.
How an EchoDepth test works for this format
Remote and browser-based, managed for you from brief to debrief.
Send the cuts
Your full edit plus any alternative openings or bumper cut-downs.
Match the audience
Testers from our panel matched to your targeting: age, gender, region and interests.
Watch at home
Viewers play the ad in a browser at YouTube's aspect ratio, webcam on, after consent.
Read the curve
Second-by-second Emotional Response Curve with the skip point marked, plus a debrief call.
Pre-test or live test?
They answer different questions. Most teams need both.
| Pre-test with EchoDepth | Live test in market | |
|---|---|---|
| When it happens | Before you spend on media | After the campaign is live |
| What it tells you | Why a moment works or fails, from facial response | Which ad got more views, clicks or completed views |
| Who watches | A matched sample who have agreed to take part | Real users scrolling, in a real feed |
| Limits | Viewers know they're in a study, so it isn't a live feed | Paid media is spent before you learn anything |
A webcam test is not a live feed. Viewers know they are in a study and they don't have a skip button, so we measure response to every second rather than real skip behaviour. That is a strength for diagnosing an edit and a limit for predicting view-through rate. Use the pre-test to choose and fix the cut, then use YouTube's own reporting once it is live.
Questions
How the measurement works is set out in the methodology. For other formats, see TikTok, Reels and Shorts ad testing and TV ad pre-testing, or the ad testing overview.
How do you test a YouTube ad before it runs?
Show it to a matched sample of viewers before launch and measure their response. EchoDepth plays your ad to 15 or more consenting viewers per audience segment on their own webcam, codes their facial response on every frame and reports a second-by-second curve, so you can fix the edit before you spend on media.
Why do the first five seconds matter so much on YouTube?
On skippable in-stream ads, Google lets viewers skip after five seconds. If the opening doesn't earn attention by then, many people never see the rest of the ad or the brand.
Can you test YouTube bumper ads?
Yes. Bumpers are non-skippable and six seconds at most, so the test focuses on whether the single idea lands and the brand registers in that time.
Can you test YouTube Shorts ads?
Yes. Shorts ads are vertical and shown at native aspect ratio in the test. See the TikTok, Reels and Shorts page for more on vertical formats.
Does a pre-test predict view-through rate?
Not directly. Viewers in a study don't have a skip button, so we measure their response to every second rather than real skip behaviour. Use the pre-test to choose and fix the edit, and YouTube's reporting to measure performance once it's live.