I have been waiting for this for so long! Youtube has finally released their thumbnail A/B testing tool feature for all content creators now on the platform. You can test up to three thumbnails and the algorithm will decide which thumbnail has the best watch time share (meaning click-through rate and average view duration combined) and will automatically use that thumbnail. Amazing isn’t it!
At last, YouTube released the A-B testing tool for free on the platform for every creator. I’ve been waiting for this feature for so long just because I’m a small creator and every impression counts.
I recording this video immediately after I saw that the A-B testing tool is out and I know bigger creators have had access to the A-B testing tool for several months now just because YouTube gave bigger creators access just to test it out I guess. But now everyone can use it!
This is really great news for us smaller creators, everybody has access. And this was pure coincidence, I found that the A-B testing tool was out not from a reddit post or video but just because I was in YouTube studio doing some thumbnail changes.
Let me show you what it looks like in this video.
How did I find out? I was in this channel doing a thumbnail for a new video, well actually an old video… Long story short, I was splitting my old channel into three different channels but that’s another story. I was doing a new thumbnail and then I noticed, you know how you previously got thumbnail recommendations, YouTube would do screenshots from the video as it was playing, whereas now I saw like this little dotted line.
And then I wondered what’s what’s happened and funny thing is I was doing studio for a few hours that day, I think the A-B testing tool just got released because I was seeing my previous options in Youtube studio as usual.
When you click there on the new option icon-thing, you’d have “change”, “download auto-generated” and then you can see also “test and compare”. This is the A-B testing tool.
When you click there you might get a message saying “build your channel history”. Don’t worry! This is because your channel is probably a brand new channel. I also got this message but I was just preparing and launched the channel a few days before seeing this new AB testing option. That video history will build up.
In order to use the A-B testing tool you actually need to have “feature eligibility” ticked.
When you go to “settings”, go to “channel” and then you go to “feature eligibility” and you need to have this “advanced features” ticked. You need to be eligible to use test and compare.
This has been added just now as an advanced feature for the final category for eligibility. In order to be eligible, you just need to have channel history, so basically upload several videos and wait a few weeks.
On one of my channels, I don’t have this ticked just because this is a brand new channel. I have like around 50 videos on the old channel so we do have the feature eligibility ticked over here. This channel is seven or eight months old so it has roughly 50 videos so we do have access to the thumbnail compare.
Let’s try it out!
Now every creator on YouTube has the chance to test three different thumbnails and the platform, the algorithm, will find out the best one. The algorithm will present different people with a different version of the thumbnail and then based on the click-through rate the algorithm will automatically choose the best one and then tell you which one is the victorious one and it will keep that one.
I haven’t done much research but I think it’s not only click-through rate. I actually think it’s based also on average watch time as well because the algorithm works like that. It checks the click-through rate and then the average watch time because when you put a really cool thumbnail that’s really clickable but has nothing to do with the video itself then people will drop off. That’s not really a good thumbnail because it’s not representative of the video content itself so it’s considered misleading.
I think what the platform will do is that it will choose the thumbnail that’s the best representative of the video with the highest click-through rate.
Basically, with the highest click-through rate while also having a highest average view duration combined.
This is how it works:
- Upload images that you want to test.
- YouTube will see which one performs the best.
- Wait for the results up to two weeks to see the report.
- Page viewers will see the winning thumbnail.
- Youtube automatically updates your video when the test is done.
Right, remember it’s up to you. You can keep the winning thumbnail or choose a different one. They are obviously giving you the freedom to choose whatever you like. You don’t have to use this tool every time and it is quite a bit of work to create three thumbnails every time.
When you start a test, it says “running”. You can click “view test report” but if there is not enough clicks it will say “not enough information, please check back later”.
It says “running time two weeks”.
Watch time share
How is the winning thumbnail selected – the winning thumbnail will be the one that generated the highest share of the video’s watch time.
Okay, so it’s the thumbnail that didn’t get the highest click-through rate but it’s the thumbnail that the viewers clicked and watched the video the longest for combined.
Why is watch time share used – great thumbnails don’t just get viewers to click. They also help viewers to understand what the video is about so that they can make informed decisions on what to watch by selecting based on the watch time share. The winning thumbnail is the one that gets viewers both – to click and stay to watch.
Right, so I got it spot on earlier when I was thinking how it works.
Continue Youtube instructions citation:
“Additionally, YouTube search and discovery systems consider both when ranking videos. Choosing thumbnails based on watch time share will benefit both you and the viewers.”
Cool, the viewer will see the best thumbnail because it’s the best representative of the content. So the viewer will decide if that’s what they want to watch and then we get more clicks and more watch time based on the thumbnail.
Thing is, if you set up an AB Thumbnail test on a video that does not get impressions or clicks, the test will say that it is “inconclusive”. If you don’t get a lot of impressions, then it can say “oh, I couldn’t test it out enough or people didn’t click enough for two weeks so I can’t tell you what the best thumbnail is”.
Good luck with your AB tests!
Hope this was helpful!
