---
title: 'A/B testing on YouTube'
source: 'https://youtube.com/watch?v=jKRiib3Xzoo'
video_id: 'jKRiib3Xzoo'
date: 2026-06-14
duration_sec: 253
---

# A/B testing on YouTube

> Source: [A/B testing on YouTube](https://youtube.com/watch?v=jKRiib3Xzoo)

## Summary

YouTube's A/B testing feature allows eligible creators to test up to three video titles and/or thumbnails to determine which combination generates the most watch time. The feature is available on YouTube Studio for public, long-form videos and provides statistical results to help optimize video performance.

### Key Points

- **Introduction to A/B Testing** [0:00] — YouTube's A/B testing lets creators choose up to 3 titles and/or thumbnails to test with viewers to pick the winner.
- **How to Start a Test** [0:12] — Eligible creators can use the feature via YouTube Studio desktop when publishing a new video or on an existing video's Video Details page.
- **Test Types** [0:26] — Creators can choose between Title only, Thumbnail only, or Title and Thumbnail combination tests.
- **Title and Thumbnail Limits** [0:35] — Up to three titles (under 100 characters each) or three thumbnail images can be tested. For combinations, three title-thumbnail pairs can be tested.
- **Availability and Start** [0:55] — A/B testing is only for public, long-form videos. The test starts once the video is published.
- **How the Test Works** [1:02] — YouTube shows the options evenly across the audience and selects a winner based on which generates the most watch time share.
- **Monitoring the Test** [1:17] — Creators can check the test status on the Video Details page under 'View test report', showing time left and performance.
- **Test Duration** [1:30] — Results may take a few hours to a couple of weeks, depending on how similar the options are and the video's traffic.
- **Test Results** [1:52] — If statistically significant, a 'Winner' label appears and YouTube automatically updates the video. If 'Performed Same', the first option is selected.
- **Manual Control** [2:20] — Creators can always manually select the title and thumbnail, or run a new test (which deletes the old experiment).
- **Best Practices** [2:33] — Titles and thumbnails must follow Community Guidelines. They should work together and help viewers understand the video content.
- **Testing Strategy** [3:10] — Start with a limited selection of older videos, then test recent videos for meaningful insights. Test thumbnail variations with different layouts and elements.
- **Thumbnail Tips** [3:34] — Use balanced images with a focal point, high contrast, minimal text, and consider depth of field and blank space.

### Conclusion

YouTube's A/B testing feature helps creators optimize titles and thumbnails by providing data-driven insights. By testing variations, creators can improve click-through rates and overall video performance.

## Transcript

Not sure which title or thumbnail to use?
Try “A/B  testing,”  a feature that lets you 
choose up to 3 video titles and/or  
thumbnails to test with viewers to help 
pick your winner. Here’s how it works.
Eligible creators can use the “A/B testing” 
feature through YouTube Studio desktop  
when publishing a new video, or with an 
existing video. On the Video Details page,  
click the icon under the ‘Title’ section 
or ‘Thumbnail section. From here,  
choose if you want to run a Title 
only test, a Thumbnail only test,  
or a Title and Thumbnail test.
For title only, you can add up  
to three different titles to test. 
Titles must be under 100 characters. 
For thumbnail only, you can upload three 
different images to learn which viewers  
prefer. For title and thumbnail, you can add 
three different title and thumbnail combinations
After you’ve added your titles, thumbnails or 
combinations of both, click Done. A/B testing is  
only available for Public, long form videos, and 
the test will start once you publish your video.
Here’s how the test works: 
YouTube will show your chosen titles and/or 
thumbnails evenly across your video’s audience,  
and will select a winning thumbnail based on 
which one generates the most watch time share 
You can check the status of your test at 
any time by going to the Video Details  
page and clicking “View test report” 
under the Title or Thumbnail section 
You’ll see how much time is left for the 
test and how they’re performing. You can  
also choose to stop the test and manually 
set one of the options as the winner 
If you let the test run, the results 
may take a few hours or up to a couple  
of weeks to be finalized.
How long it takes to get  
the results is affected by factors like:
How similar your titles and/or thumbnails  
are (the more they differ, the 
faster your test will finish) 
And your video traffic (the more 
impressions that your video receives,  
the faster your test will finish)
Once the test is finished,  
the test report will show the results
If there’s a statistically significant result,  
you’ll see a “Winner” label. YouTube will 
automatically update your video title and/or  
thumbnail to this winner when the test is done.
You might also see a “Performed Same” label.  
This result indicates that all test options had 
similar watch time performance. If there’s no  
clear winner, or if the options “Performed 
Same,” the first title and/or thumbnail  
will be selected and shown to your audience.
Remember that in either case, it’s ultimately  
up to you – you can always manually select 
the video title and thumbnail you want to use. 
You can also choose to run a new test, but 
note it will delete your old experiment
Now let’s walk through some things 
to note and best practices for  
using the “A/B testing” feature:
Titles and Thumbnails must follow  
our Community Guidelines, and the tool will 
not work on videos that are designated as  
made for kids, for mature audiences, or live.
Note that great titles and thumbnails serve  
an important purpose beyond getting viewers to 
click. They help a viewer understand what the  
video is about so that they don't waste 
their time clicking on the wrong videos. 
Remember that your title and thumbnail should 
work together as a team. We highly encourage  
testing title and thumbnail combinations 
to find which pairing works best to help  
viewers understand what your video is about.
We suggest testing titles and thumbnails on  
a limited selection of older videos for 
initial guidance, then a selection of more  
recent videos you think will provide the most 
meaningful insights for your content strategy. 
We also recommend testing thumbnail 
variations that explore different  
layout compositions and changes to thumbnail 
elements (including background, text overlays, and  
object placement) to learn which resonates most 
with your audience. Some tips to keep in mind: 
Ensure your images are balanced, with a focal 
point to direct viewers attention towards 
Utilize high contrast, allowing 
your subject to stand out against  
the background in both light and dark mode.
With text, remember that fewer words can be  
impactful, while too much text can be 
difficult to process while scrolling. 
And finally when it comes to depth of field, keep 
your background in mind and play with blank space
We hope the A/B testing feature will help you 
gain insights when it comes to making titles  
and thumbnails, and help your overall video 
performance in the long run! For more info,  
check out the links below. Thanks for watching!