Now you can. Here's how.
Or make up a bunch of new titles
Run tests for as long as you want.
Your test runs automatically.
We'll update every thumbnail or title once per day until the end of the test.
See which thumbnails and titles won, and how much you improved your click-through rate.
You've got a ton of old videos, why not try a few different titles? You never know, they just might go viral.
What our early users have said
If I've learned anything from ten years at Netflix, it's that no matter how hard you try, you never really know what's going to make someone click.
I love this tool. Much more affordable & easier to use than TubeBuddy, and just as effective (if not more) for A/B tests.
We store all the thumbnails you upload. Every day at midnight, we change the thumbnail on your chosen video to the next thumbnail in the list. At the end of the test, we look back at the click-through rate each thumbnail got and use that to calculate which thumbnail performed the best. That thumbnail (or title) is marked the winner.
Normally, an A/B test is when you show 100 people something, 50 see A and 50 see B, then you judge if A or B is better based on their reactions. So, a thumbnail A/B test means we show thumbnail A for one day, then thumbnail B for one day, and repeat that for a couple weeks. At the end, we see which thumbnail more people clicked on.
This tool is not magic, it only runs the test. This tool can make you more powerful only if you use it properly. Make different thumbnails, come up with new titles, and then this tool can do the heavy lifting to tell you what’s working. The best thing you get out of this tool won’t be views or click-through rate, it will be understanding what your audience actually wants deeper than ever before.
Nope! YouTube has announced “plans” to implement A/B Testing in the next year or so, but they’ve been saying that for like 5 years so it’s impossible to know what’s really gonna happen. But, worst case, this will be a useful tool until they make it :)
Click-Through Rate is #views/#impressions - so out of all the people that saw your video, how many of them clicked on it? This is extremely important, as this allows us to do calculations even though different days have different numbers of impressions or views, because we can judge based on the percentage.
Average View Duration is how long the average viewer watched your video.
We use this stat based off of recent research from the discord/twitter thumbnail communities that has shown this to be the most reliable metric of video success. We are constantly looking out for new stats that can give more insight, so if you have more data or research to show please reach out to us!
As many as you want. The only limit is one active test per video.
We recommend no more than 5, but you can upload as many as you want. All thumbnails will be tested including the original.
Try to make very unique thumbnails. If all you do is change one color in the image, it’s much less likely to have a big impact, and you’ll learn less.
We judge Thumbnails/Titles based on CTR*AVD (Click Through Rate * Average View Duration). We use this stat based off of recent research from the discord/twitter thumbnail communities that has shown this to be the most reliable metric of video success.
You can only run one test per video at a time. You can run a thumbnail test, a title test, or a combination thumbnail+title test.
If you find that your click-through rate is way higher the first few days, we recommend adding a 3-day delay to your test to ensure you’re running with the least bias possible. This should get around any big CTR spikes and ensure a true test of each thumbnail.
As many as you want! All titles will be tested including the original.
As many tests as you want! The only limit is one test per video at a time.
As long as you want! But tests will always be a multiple of the number of thumbnails/titles. For example, if you upload 3 thumbnails + the original, your test will run for a multiple of 4 days (4, 8, 12, 16, ….) to ensure every thumbnail gets an equal amount of test time.
After each thumbnail has been run for at least one day, you can upload partial stats and see what thumbnail is currently winning.
Almost! We have automated all the thumbnail/title swapping throughout the test. The only manual part is that at the end of the test, you have to upload your analytics from YouTube studio. This takes about 15 seconds.
Pick a video, upload your thumbnails/titles, hit start test. Wait until the test is over, upload your analytics one time, and boom you’re done!
When you sign up, you give the tool access to your basic analytics. We’ll use this connection as much as we can, but at the end of the test you will still need to upload your analytics for that specific video so we can give an accurate result.
1280x720, less than 2MB
Each thumbnail/title must be tested for at least one day. The number of thumbnails/titles you upload, plus 1 for the original, determines the minimum test length.
Partially. You are giving us access to look at some of your YouTube analytics, as well as change thumbnails/titles on your videos.
At the moment, the ThumbnailTest.com costs $19/month or $15/month paid annually.
You can start the test at midnight GMT any day. By default, we will start swapping thumbnails as soon as the clock hits midnight GMT after you start your test.
We collect the minimum data needed to figure out which thumbnail/title converted the best. This includes impressions, views, CTR, and AVD for every day the test ran. We also store a token used to update your thumbnails and titles overnight.
This is dependent on the channel. Soon we’ll have a channel-wide stats import that will help you know what number to shoot for.
Me! Rox - roxcodes on Twitter and Twitch. I build all my products live on stream, so if you have any questions feel free to drop by.
We’re still doing research on this, but for now we estimate a few hundred views per video. Below that it’s hard for data to be significant.
100% safe, the only risks with tools like this is if the thumbnail is changed super fast (like 10 times a second or something crazy). Since this tool is swapping thumbnails at most once per day, you are at absolutely no risk.
Yes. This tool is literally updating your live video’s thumbnail. This is the only true way to test which thumbnail is better - see what the audience clicks on :D
This is a tough one, but really the only bias is if suddenly one of your videos blows up overnight and gets a ton of views - this could skew the data and make it tough to make an accurate call. But if you want to be super safe, delay the test 2-3 days after publish, and as long as you don’t suddenly go viral, the test will be accurate. And if you do go viral, who cares about the test 😀
If you aren’t A/B testing thumbnails, you’re playing a guessing game. By testing multiple thumbnails, you remove the chance that your video flops because of a bad thumbnail - you’ll always know you have the best thumbnail you could make. On top of this, you’ll get to learn more about your audience and what they like by seeing how they react to different thumbnails and titles.
High CTR = Great Thumbnail + Titles. High AVD = Great Video. High CTR + High AVD = Viral!
No. Thumbnails will always switch at midnight GMT as this is when YouTube analytics update.
The tool will always include the original thumbnail or title in the experiment to ensure that you know if your original was better than your new variations.
Not yet, but if more people ask for this we can add it :D
Probably, haven’t tried yet. If you want to try this, reach out to Rox#074 on discord and let’s see what we can do :D
Click “Ongoing”, then press “Stop Test” in the top right.
For now, please ask all questions in the discord. If you’ve paid for the tool and haven’t gotten an invite yet, please contact Rox#0074 on discord.
Click “Past” on the sidebar
Click “Ongoing” on the sidebar
Click “New” on the sidebar. Select your type of test, either Thumbnail or Title at the top of the screen. In step 1, select the video you’d like to test. In step 2, upload your thumbnails or add your new titles. In step 3, set test length and any delays, then hit “Start Test”.
For now, please report all bugs in the discord. If you’ve paid for the tool and haven’t gotten an invite yet, please contact Rox#0074 on discord.
For now, please give all feedback in the discord. If you’ve paid for the tool, click the Discord link in the sidebar. Otherwise, please dm Rox#0074 on discord :)
Yes! We give affiliates 60% of lifetime revenue from everyone you bring in.
If you don’t love it in 30 days, you can get a full refund of your first payment. After that, no promises :)
I saw the opportunity that TubeBuddy was forcing people to pay $50 a month for a million tools when 90% of people were just using the A/B testing tool. Felt like I could make a better A/B Testing Tool, and charge less for it, and convert a lot of users over :)
Yes, with the team plan you can have up to 3 total team members (yourself included). For more team members, please reach out to Rox#0074 on discord for an enterprise plan.
Yes, with the team plan you can have up to 3 total channels. For more channels, please reach out to Rox#0074 on discord for an enterprise plan.