Skip to Main Content
YouTube test report

Rene’s Top Five on YouTube: June 14, 2024 edition

I get it, I get it! No preamble! Thumbnail Test & Compare (A/B Testing) is rolling out and that’s all most every creator cares about so let’s just get right to it!

Who? Thumbnail Test & Compare is rolling out to any and all channels with Advanced Features enabled. If you’ve been growing a channel for a while, it probably already has those features enabled. If the channel is newer and fresher, you might need to go into settings and work on enabling them.

What? Thumbnail Test & Compare let’s creators choose up to 3 thumbnails that YouTube will then split-test among viewers. Each variant will be tested against roughly the same sized audience segment (half for two thumbs, a third for three thumbs), and each viewer will only ever see the same thumbnail variant on the same account, so no one will be confused into clicking on the same video twice (unless they’re switching accounts or seeing them linked on different social networks, for example). The test will run until the system can determine one of the variants has achieved a significantly higher percentage of watchtime, or 2 weeks go by without a meaningful result. You can run tests on new videos, old videos, and re-run them on the same videos any time.

Where? Thumbnail Test & Compare is currently available to channels around the world as part of the desktop uploader, exactly where you’d previously select a custom thumbnail for your video. Just now, instead of picking a frame or uploading a single image, you can upload 2 or 3 images for testing.

When? Thumbnail Test & Compare’s rollout is starting now and will continue over the coming weeks. If you have multiple channels, you may see it active one channel before another. Just check back every now and then and it’ll pop up!

Why (Watch Time)? Thumbnail Test & Compare returns watch time rather than separate metrics on click-through rate (CTR) and retention (AVP), because watch time includes both! You have to click to watch and you have to retain to build up time. If you over-index on CTR, it could become click-bait, which could tank retention, and hurt performance. This way, the tool helps build good habits — thumbnails that make a promise and videos that deliver on it!

If at first you don’t get insights, take notes, maybe try larger differences between thumbnails, and keep experimenting and learning.

Bonus Top Five Tips!

🕰️ Test immediately to see how your core audience is reacting. When your initial test completes, test again 72+ hours into a video for insights into the wider, more casual audience. Test any time — weeks, months, years! — to see how updated branding, different focuses, and other factors affect evergreen videos.

😮 Start testing with bigger differences between thumbnails, like the layouts, different image elements, etc. and, if you feel you need to, work your way to more granular differences like clothing colors, background elements, and facial expressions.

😬 Be patient and be curious. Not all tests will end with statistically significant results, especially for channels that aren’t pulling a ton of impressions and views, or for thumbnails that perform about the same for your audience. If at first you don’t get insights, take notes, maybe try larger differences between thumbnails, and keep experimenting and learning.

🤔 It’s not just about one video’s performance but about learning how the audience reacts to your thumbnails over time to grow your skills and lift your whole channel(s). Analytics don’t just tell us about how we’re doing in the now, they help educate us so we can do better in the future.

🎉 Audiences and trends change over time. Orange can pop until everyone does orange, then purple might pop. Have fun! Test. Compare. Iterate. Repeat!

Test & Compare can and will evolve and expand over time, so get with the contenting and let us know what you think!