How to Use YouTube Test & Compare for Titles and Thumbnails
YouTube Test & Compare runs up to 3 titles or thumbnails at once and keeps the one with the highest watch time share. Here's how to set up and read a test.
By Sam Parham, Co-Founder and CEO, 1DS · Published · Updated · 5 min read
YouTube Test & Compare lets you upload up to 3 titles, thumbnails or both for 1 video. YouTube shows the options to different viewers at the same time, then keeps the one that earned the highest share of watch time. You start a test from the title box or the thumbnail section in YouTube Studio on a computer, and it should finish within 2 weeks.
Every rule below comes from YouTube's own help pages, checked in November 2026. The tool keeps changing, so check eligibility again before you plan a round of tests.
What is YouTube Test & Compare?
It's YouTube's built-in A/B test for packaging, meaning the title and thumbnail a viewer sees before clicking. You give 1 video up to 3 options, and the one with the highest watch time becomes the version everyone sees.
YouTube's blog announced the rollout of Thumbnail Test & Compare on June 14, 2024. Its help page for the tool now covers titles as well, and the button in YouTube Studio reads A/B Testing.
You pick 1 of 3 test types:
- Title only, with up to 3 titles.
- Thumbnail only, with up to 3 thumbnails.
- Title and thumbnail, with up to 3 combinations.
YouTube may hold back a small share of viewers as a control group. Those viewers only see the default title and thumbnail, and their viewing is left out of the result.
Who can use YouTube Test & Compare?
Any channel with advanced features turned on can use it in YouTube Studio on a computer, on long-form videos. Several formats and video settings are ruled out.
- Device: the tool only works on computers, through YouTube Studio.
- Advanced features: YouTube says active creators with enough channel history will likely qualify automatically, and you can verify with an ID or a video to get access sooner. Check yours under Settings, then Channel, then Feature eligibility (YouTube's feature access page).
- Formats: Shorts, scheduled live streams and Premieres can't be tested. Live archives can, and a Premiere becomes eligible after it ends and converts to a long-form video.
- Settings: videos made for kids, made for mature audiences or set to private can't be tested. Age-restricted content has to be marked 18+, or you can lose access to the tool.
Shorts catch people out. YouTube treats square or vertical uploads of 3 minutes or less as Shorts, so a vertical 2-minute video can't be tested, and a video that converts to a Short loses its earlier tests.
How do you set up a test?
Open the video in YouTube Studio, click A/B Testing in the title box or under the thumbnail, add your options and click Done. Setup takes a few minutes once your options are ready.
- Sign in to YouTube Studio on a computer.
- Upload a new video, or open an existing one from Content.
- Click A/B Testing in the Title box or under Thumbnail.
- Choose Title only, Thumbnail only, or Title and thumbnail.
- Add up to 3 options, then click Done.
Leave the title and thumbnail alone while the test runs. YouTube says any change to either stops the test, and you'd have to start again.
Upload every thumbnail at 1280 x 720 or higher. If even 1 option is smaller, YouTube downscales all of them to 480p (854 x 480).
To check progress, open the video's Analytics, go to the Reach tab and click Manage test under "How your A/B test is going." You can stop a test from there at any time.
How does YouTube pick a winner?
On watch time share. The option that earned the biggest share of the video's watch time during the test wins, which favors packaging that brings in viewers who stay.
YouTube's help page says it optimizes tests "for overall watch time over other metrics, like click-through-rate." Its reason is that good titles and thumbnails help viewers understand what a video is about, "so that they don't waste their time clicking on the wrong videos."
That's also why YouTube's verdict can differ from a third-party tool's. YouTube shows every option in the same period, while many outside tools run options in sequence, and YouTube says those tools "often only optimize for click-through rate, which may determine a different 'winner' than measuring by watch time share."
A thumbnail that oversells can win the click and lose the test. Our guide on how to read your CTR explains the clickbait signal YouTube describes: high CTR, low average view duration and lower than expected impressions.
What do the results mean?
Every test ends with 1 of 3 labels. Only Winner means YouTube found a difference it calls statistically significant.
| Result | What YouTube says it means | What to do next |
|---|---|---|
| Winner | The option clearly outperformed the others on watch time share, and YouTube is sure the result is statistically significant | Keep it, note what changed, and try the same kind of change on a similar video |
| Performed Same | All options performed about the same, with no clear winner | Pick the option you prefer, and make the next test's options more different |
| Inconclusive | No strong statistical difference in engagement between the options, so the first title and thumbnail you uploaded becomes the default | Check how many impressions the video got, then retest with bigger differences or on a video with more traffic |
YouTube's help page says "It's normal not to receive a 'Winner' test result." It gives 2 reasons: the options were too alike to make a measurable difference, or the video didn't get enough impressions.
Results can shift between runs on the same video too. YouTube compares that to flipping a coin, and adds that early impressions lean toward viewers who already know your channel, while later ones reach more people who don't.
Does YouTube Test & Compare work?
Yes, as a measuring tool. It shows which option earned more watch time from viewers who saw the options in the same period, though it can't tell you why, or whether the same change will win on your next video.
That makes it more reliable than swapping a thumbnail by hand and comparing weeks. YouTube's CTR FAQ warns that when several titles or thumbnails are tried on 1 video, differences "might be due to traffic sources, rather than the title or thumbnail." The A/B tool shows every option in the same period instead.
A plan that gets more from each test:
- Start with older videos that still get steady impressions. YouTube recommends this to reduce the impact on your channel's overall views.
- Change 1 big idea per test, such as a face against no face, or a searchable title against an intriguing one. YouTube's title tips describe those 2 title types.
- Let each test run to its result, which can take up to 2 weeks.
- Log every test, including the ones that end Performed Same or Inconclusive.
Then make sure the opening keeps the promise the winning package makes. Our guide on how to write a hook for a YouTube video covers the first 30 seconds, and our guide to repurposing content across platforms covers adapting a winning idea for TikTok and Reels.
YouTube Growth, 1 of our 10 services, covers long-form strategy, packaging and retention for the creators and athletes 1DS Management represents. If you want a team planning and reading tests like these with you, apply for representation.
Questions creators ask
Can you use Test & Compare on YouTube Shorts?
No. YouTube says A/B testing isn't available for Shorts, and it treats square or vertical uploads of 3 minutes or less as Shorts. If a video converts to a Short, you lose access to its earlier tests.
How long does a YouTube A/B test take?
YouTube says a few days or up to 2 weeks, depending on impressions, how recently the video was published and other factors. Videos with more views are more likely to produce a winner.
Why didn't my test pick a winner?
YouTube names 2 common causes: options too similar to make a measurable difference, or too few impressions. Make the next set of options more distinct, or test a video that gets more traffic.
Does an A/B test hurt a video's views?
Some options will do worse than others while a test runs. That's why YouTube recommends testing older videos first, to reduce the impact on your channel's overall views.
Can you test a title and thumbnail together?
Yes. Choose Title and thumbnail, upload up to 3 combinations, and YouTube keeps the one with the highest watch time. A combined test won't show whether the title or the thumbnail made the difference, so test them separately when you want to know which one matters.
Why does a third-party tool show a different winner?
YouTube says many outside tools test options in sequence and score them on click-through rate. Its own tool shows every option in the same period and scores on watch time share, so the 2 can disagree.
Written by
Sam Parham
Co-Founder and CEO, 1DS
Want a team negotiating your deals?
1DS Management represents more than 100 creators and athletes across brand partnerships, content strategy and monetization.