SIRGOVIRAL CREATOR GUIDE

How to Test YouTube Thumbnails and Read the Results

Test up to three designs in YouTube Studio and use the result to plan your next thumbnail.

To A/B test YouTube thumbnails, open the video in YouTube Studio on a computer, choose A/B Testing → Thumbnail only, and upload two or three designs. Keep the video title fixed so your test can answer one question about the thumbnail. When the result is ready, check YouTube’s result label alongside each option’s watch time share.

Choose a video and decide what you want to learn

Start with a long-form video that still earns impressions. Pick one decision you can reuse, such as whether viewers respond better to the finished result or to a price shown up front. For a Nigerian cooking channel testing a budget jollof recipe, compare the finished plate with a photo of the ingredients and their total cost. A third option might show the creator holding the finished plate.

Make every option deliver the same promise as the video. If you show “₦5,000,” explain what that amount covered and when you priced it. Use a legible crop, keep the important detail large enough for a phone screen, and export every design at 1280 × 720 or higher. YouTube says one design below 720p can lower the resolution of all versions in the test.

Run a thumbnail-only test in YouTube Studio

  1. On a computer, open YouTube Studio → Content and select the video you want to test. You can also set up a test during a new upload.
  2. Find A/B Testing in the title or thumbnail area. Choose Thumbnail only. The older thumbnail tool was called Test & Compare.
  3. Add your current thumbnail first as a baseline, then upload one or two alternatives. Select Done and finish saving the video details if Studio prompts you.
  4. Give the test time. Results can take several days or up to two weeks. Changing the title or thumbnail during the test stops it.
  5. To review the test, open the video’s Analytics → Reach, find the test card and choose Manage test.

Read the watch time share and result label together

Watch time share shows how the viewing time attributed to the experiment is split across your options. Click-through rate measures how often an impression becomes a view. A design can earn clicks and still deliver less watch time if viewers leave early. Use the test result to judge how well each design attracts people who stay to watch.

Example watch time share for a budget jollof video
  • A · The finished meal · 38%
  • B · The ingredients and total cost · 34%
  • C · The creator holding the finished meal · 28%

These figures are illustrative. A leads by four percentage points, but the report label tells you whether YouTube found a clear winner.

  • Winner. One option clearly led on watch time share. YouTube applies it when the test finishes. Record what changed and test that idea again on a similar upload before treating it as a channel-wide rule.
  • Performed Same. The versions performed similarly. Choose the clearest design at phone size, then try a different idea on a future test.
  • Inconclusive. The test did not establish a clear difference. YouTube defaults to the first option. If the video still receives impressions, test more distinct concepts or run the next experiment on a video with steadier traffic.

Why the A/B Testing option may be missing

Check that you’re using desktop Studio and that Advanced features is enabled in the channel’s feature eligibility settings. YouTube excludes Shorts, private videos, made-for-kids videos and age-restricted videos. A scheduled Premiere needs to end and become a long-form video before it can be tested.

Turn the result into your next thumbnail brief

Keep a small log with the video title, thumbnail versions, result label and your next test idea. For example, “The finished dish won on the budget recipe. Next test: finished dish with and without a price label.” This gives you a specific design decision to test on the next comparable upload.

For more creator tutorials, browse Sirgoviral’s Social Media Guides. If you are separately comparing promotion options, the Sirgoviral YouTube views guide explains package details. Judge the thumbnail with the data in your Studio experiment.

Platform reference: YouTube Help on A/B testing titles and thumbnails.