[0:00] Everyone says AB testing your thumbnail
[0:02] and titles is the smartest thing you can
[0:04] do for your YouTube growth. YouTube
[0:05] literally built the feature for you, so
[0:08] of course you should use it, right?
[0:09] Wrong. Because if AB testing is supposed
[0:11] to help your video perform better, then
[0:14] why did six out of seven of my videos
[0:16] completely tank the moment I turned it
[0:19] on? I was just like you. I saw the big
[0:21] creators raving about it. Mr. Beast uses
[0:24] it. So, I thought, "Okay, let's me go
[0:26] all in." I ran AB testing on seven brand
[0:29] new videos back-to-back. I trusted the
[0:32] tool and one by one I watched those
[0:35] videos die. But, what if AB testing
[0:37] itself isn't broken? What if YouTube is
[0:40] quietly punishing you every single day
[0:43] the test is still running and you don't
[0:46] even know about it? Here's what I found.
[0:48] When YouTube runs your test, it splits
[0:50] your traffic equally between all your
[0:51] variations. So, if you're testing three
[0:54] titles, 66% of your audience is seeing
[0:57] an experimental title that might be
[0:59] terrible. And YouTube is judging your
[1:02] video the entire time this is happening.
[1:05] So, in this video I'm going to show you
[1:06] the actual math behind why this tool can
[1:09] destroy your video's momentum and who
[1:12] should and honestly should not be using
[1:15] it. But first, let's me quickly break
[1:17] down how AB testing actually works in
[1:19] 2026 because YouTube just made some big
[1:22] changes to it. As of 2026, YouTube is no
[1:26] longer just AB testing thumbnails. You
[1:28] can now test both titles and thumbnails.
[1:31] You can test up to three different
[1:33] titles, three different thumbnails, or
[1:36] three combinations of titles and
[1:38] thumbnails. How cool is that? One thing
[1:40] most people don't understand is YouTube
[1:42] doesn't pick the winner based on the
[1:43] click-through rate or CTR. Instead,
[1:46] YouTube picks based on watch time and
[1:49] average view duration. That's quite
[1:51] interesting, isn't it? I mean, it is
[1:52] called AB testing, so wouldn't they want
[1:55] to track clicks instead? The truth is
[1:58] YouTube is doing this because YouTube
[1:59] doesn't want clickbait. For example,
[2:02] take a look at these three thumbnails.
[2:04] If version A gets 10% clicks, but
[2:06] everyone leaves in 5 seconds because it
[2:08] is clearly clickbait, and version B gets
[2:11] 2% clicks, but they watch for 50% of the
[2:14] video, then version B wins even though
[2:17] the CTR is significantly lower. And
[2:20] here's what I discovered when running AB
[2:22] testing on seven newly launched videos.
[2:24] Based on my tests, all six out of my
[2:28] seven videos of mine underperformed,
[2:30] failing to gain 1,000 views. You see,
[2:33] the biggest issue with YouTube AB
[2:35] testing is if it can't find a clear
[2:37] winner, it drags the test up to a
[2:40] maximum of 14 days. That's actually a
[2:42] trap because it damages your channel
[2:45] each day it drags. Let me explain to you
[2:47] why. Check out this test I did for one
[2:50] of my videos. Nothing fancy, just a
[2:52] title test. Based on the final result,
[2:55] it is clear my top title is the winner
[2:58] with over 46%
[3:00] watch time share. But you see, if you
[3:01] squeeze your eyes real good, you'll
[3:04] notice that this test took over 11 days
[3:07] to conclude. So, just by thinking of
[3:09] this logically, the math doesn't add up.
[3:11] You see, in order to conclude which
[3:13] title is the best, YouTube needs to test
[3:15] all three of them, meaning my traffic is
[3:18] split 33 33 33. If I have one clear
[3:22] winner, then that makes the other two
[3:24] experimental titles. And those two
[3:26] experimental titles are shown to 66%
[3:30] of my audience. So, let's do a simple
[3:32] logical walk-through here. YouTube
[3:34] starts with the top one, found that it
[3:36] works, has over 40% watch time. But,
[3:39] YouTube doesn't know if this is the best
[3:41] yet, so it pauses the top one, then
[3:44] tests the second title. Oh, no, the
[3:46] second one did not perform well after 3
[3:48] days. So, let's move to the third one.
[3:51] Again, oh, no, the third one failed,
[3:53] too. So, that makes the first one the
[3:56] clear winner. And the entire test took
[3:59] 11 days. So, tell me, if you knew the
[4:01] first one is the winner, then you would
[4:04] stop the test and just use the first
[4:06] title, right? But, you can't do that if
[4:08] you are AB testing and I think that's a
[4:11] massive flaw in this feature. Moreover,
[4:13] I have a very bad feeling that when
[4:15] YouTube is testing my experimental
[4:17] titles and both underperformed, YouTube
[4:20] might have received the signal that the
[4:22] video is not a good one. Therefore, when
[4:25] it decides that the first one is the
[4:27] winner, the damage has already been
[4:29] done. It is already too late. YouTube
[4:31] thinks this video is bad because of your
[4:33] experimental titles. The video has
[4:35] already lost its initial first 48 hours
[4:38] boost, so YouTube decides to not give
[4:41] any more impressions to it and the video
[4:44] fades away. I definitely do not have
[4:45] concrete proof of that, but that's what
[4:48] I saw in my AB testing and I hope that
[4:51] isn't true. That's not to say this is
[4:52] entirely YouTube's fault. There's
[4:54] certainly an angle where it doesn't make
[4:56] sense to use AB testing, especially for
[4:59] small YouTube channels. And when I say
[5:01] small YouTube channels, I don't mean by
[5:03] subscriber count. I mean by view count.
[5:06] You have probably heard that AB testing
[5:09] has been a game-changer, that it helped
[5:11] many videos boost their CTR and
[5:14] subsequently more views. What you might
[5:16] be aware of as well is that those
[5:18] sayings often come from big channels.
[5:21] So, what this means is that you need a
[5:22] big amount of views in order to conclude
[5:25] your AB testing effectively because if
[5:28] you don't have sufficient views, then
[5:29] your test will not only drag on, but it
[5:32] will also fail to come to an accurate
[5:35] conclusion. From my own experience and
[5:37] testing, I highly suggest using AB
[5:39] testing only if you can average 1,000
[5:42] views per newly launched video,
[5:45] preferably about 1,000 views in the
[5:47] first hour. The more, the merrier, of
[5:49] course. The reason is because you want
[5:51] to have enough views to really run the
[5:52] AB test to the fullest and also to have
[5:55] AB testing conclude as soon as possible,
[5:58] preferably within the first day. Heck,
[6:00] even 1,000 views is me being optimistic.
[6:03] I'm actually thinking about 10,000 views
[6:05] average. So, if you're getting less than
[6:07] 1,000 views per video, I suggest you
[6:09] stop using AB testing and instead use
[6:12] this strategy pivot. My first strategy
[6:15] is if I choose to use AB testing, I will
[6:18] let it run for the first 24 hours only.
[6:21] Then, I will evaluate the test so far
[6:24] and manually select the one that is the
[6:26] clear winner. I will not let it drag on
[6:28] for more days. My other strategy is to
[6:30] avoid AB testing altogether and do
[6:32] manual swaps. So, if I am testing the
[6:35] first title at launch, I will check it
[6:37] again after 24 hours. If it fails, I
[6:40] will manually swap to the second title
[6:43] and test it. After that, I will pick and
[6:45] go with my gut feelings based on which
[6:47] gets more views. The new video boost
[6:49] usually lasts for the first 48 hours or
[6:52] so, so this to me is quite a good test
[6:55] provided you are disciplined enough to
[6:57] do it manually. And also, if you don't
[6:59] have the traffic to support your tests,
[7:01] then avoid testing on brand new videos.
[7:04] You have already put in a lot of effort
[7:06] and time into it, so don't let the video
[7:08] underperform because of YouTube's tool.
[7:10] Another strategy I tried is to look out
[7:13] for underdogs in your channel, such as
[7:15] videos with high impressions, but low
[7:17] CTR. Videos like this actually signal
[7:20] that YouTube sees a market in it, but no
[7:22] one is clicking because of a bad
[7:24] thumbnail or title. And I think AB
[7:27] testing does help when it comes to this.
[7:29] In fact, I actually tested this on one
[7:30] of my videos. It is rather old, a few
[7:32] years old, so I just did some tests on
[7:34] it just to see what happens. But so far,
[7:37] it doesn't look like I managed to revive
[7:38] it, although I haven't left it long
[7:40] enough yet, so maybe YouTube still needs
[7:42] time to follow this one up, so we'll
[7:44] see. So, yes, I am very disappointed
[7:46] with YouTube's AB testing. I really wish
[7:48] it worked for everyone, big or small
[7:50] channels, but it is very clear to me
[7:52] that I should stop doing any more AB
[7:55] testing for now and instead rely on my
[7:58] instinct. Seven videos used to test AB
[8:00] testing is a lot of videos, but hey, I
[8:03] think I gained quite a lot from it and I
[8:06] hope you do, too. Also, I run YouTube
[8:07] channel audits as well for my viewers,
[8:09] where I will audit and give my most
[8:11] honest feedback to growing YouTube
[8:13] channels, help give them a direction,
[8:15] and help them grow their channel. If
[8:17] you're interested, check out the link in
[8:19] the description. So, this concludes my
[8:21] AB test. The next test I will do is
[8:23] going to be what happens if I were to
[8:25] upload seven videos without AB testing.
[8:28] Is my gut feeling more accurate than
[8:30] some YouTube tool? That's what I'm going
[8:31] to test next, so when the video is
[8:33] ready, it will appear somewhere over
[8:35] here. So, be sure to check it out. All
[8:38] right, thanks for watching and I'll
[8:39] catch you in the next one.