Big Idea

The argument in one line.

YouTube shrank its homepage to three videos and reserved one suggested-tab slot for small channels, so growing now means packaging your video as a viral hit's sequel instead of a copy of it.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…

A creator running a channel under 10,000 subscribers whose videos are stuck under a few thousand views despite consistent uploads.
Someone still leaning on the old 'outlier theory' of cloning a viral video's title and thumbnail, and wondering why it stopped working.
A creator who wants a repeatable system for choosing video topics instead of guessing and hoping the algorithm notices.

SKIP IF…

You already have an established channel that regularly appears on the YouTube homepage. This is written for channels the algorithm doesn't trust yet.
You're looking for a full editing or scripting tutorial. The retention section only gives high-level principles, not a step-by-step walkthrough.

TL;DR

The full version, fast.

YouTube's homepage dropped from nearly twenty videos to three, and those slots go to big, proven creators because a bad first click risks losing the viewer. Small channels have to win through the suggested tab instead, where the algorithm reserves one low-view slot between viral videos. The old 'outlier theory' of copying viral titles and thumbnails is dying: AI makes copying trivial, and YouTube's new Semantic ID model detects derivative content and suppresses it. The replacement is the 'follow-up theory,' packaging a video as an unofficial sequel to one already dominating the feed, aiming for a 5-10% click-through rate and 30% average view duration, and using AI-assisted thumbnails plus real storytelling to hold viewers.

Free for members

Chat with this breakdown — free.

Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.

Create a free account →

Chapters

Where the time goes.

00:00 – 02:13

01 · The homepage shrank to three videos

Cold open plus Change #1: the homepage dropped from near twenty videos to three, and the algorithm only trusts big, proven creators to fill them.

02:13 – 03:15

02 · The suggested tab reserves a small-channel slot

Change #2: after a viewer finishes a big video, the sidebar shows big viral videos, one mid-size creator, then deliberately squeezes in one small-channel video with a few hundred views.

03:15 – 05:30

03 · The follow-up theory: become episode two of a viral video

A new tactic explained with the 'Why Bananas Are Good For You' example: package your video as the unofficial sequel or rebuttal to a video already dominating the feed.

05:30 – 07:44

04 · The outlier theory is dying

Copying a viral video's title and thumbnail worked for five to six years but is collapsing: AI makes copying easy, and YouTube's new Semantic ID model detects and suppresses derivative content.

07:44 – 10:39

05 · What 'good enough' means: 5-10% CTR, 30% retention

Real benchmarks for click-through rate and average view duration, including how both should scale with impressions and video length.

10:39 – 12:41

06 · AI thumbnails: the Higgsfield workflow

How the creator generates and finishes thumbnails with the Higgsfield AI tool plus Photoshop, including a thumbnail that drove 1.6 million views.

12:41 – 15:02

07 · Retention tactics: visuals and storytelling

The two levers for holding viewers past the hook: restrained use of visuals/B-roll, and turning flat facts into cause-and-effect storytelling.

Atomic Insights

Lines worth screenshotting.

YouTube cut its homepage from roughly twenty videos to three, and reserves those slots for proven, high-view content instead of new channels.
The suggested-video sidebar deliberately inserts one small-channel video, often under 1,000 views, between two or three videos from big creators.
Copying a viral video's title and thumbnail, the old 'outlier theory,' is dying because AI lets creators copy each other faster and more convincingly than ever.
YouTube's Semantic ID system analyzes a video's script, format, title, and music to detect when it's derivative of another video, then suppresses its reach.
Packaging a new video as an unofficial 'episode two' of a video the viewer just finished increases the odds it fills the suggested tab's small-channel slot.
A 10-minute video needs roughly a 5-10% click-through rate to reach 10,000 to 100,000 views.
Click-through rate naturally drops as impressions climb; a video at even a million impressions can still be healthy holding close to 5%.
A 10-minute video needs about 3 minutes of average view duration, a 30% retention rate, to consistently clear 10,000-plus views.
Retention benchmarks scale with length: a 30-minute video still needs roughly 30% average view duration, not a fixed number of minutes.
AI-generated thumbnails can now look indistinguishable from photographed ones, and one AI-assisted thumbnail drove 1.6 million views on this channel.
Overloading a video with graphics and B-roll can hurt retention as much as having none; the advice is to add visuals only when they serve the point.
Turning a flat list of facts into cause-and-effect narration is what keeps viewers watching, not the underlying content itself.

Takeaway

The suggested tab, not the homepage, is where small channels grow now

WHAT TO LEARN

YouTube's homepage now favors proven, big-creator videos, so growing a new channel means engineering videos that fill the one small-channel slot the suggested tab deliberately reserves.

01The homepage shrank to three videos

YouTube dropped homepage recommendations from near twenty videos to three because showing an unproven small channel risks losing the viewer entirely.
If your channel doesn't already have viewer trust, expect little to no homepage exposure; growth now has to come from elsewhere.

02The suggested tab reserves a small-channel slot

After a viewer finishes a big video, the sidebar shows two or three viral videos, one mid-size channel, then one small-channel video with as few as a few hundred views.
That reserved slot is a real, repeatable opportunity, not luck. The job is figuring out how to be the video that fills it.

03The follow-up theory: become episode two of a viral video

Frame your video as an unofficial sequel or rebuttal to a video already dominating the suggested feed, their 'why bananas are good for you' becomes your 'why bananas are bad for you.'
The viewer just spent real time on the original topic, so a video that feels like its next chapter earns a click more easily than an unrelated pitch.

04The outlier theory is dying

Copying a viral video's title and thumbnail almost exactly, the old 'outlier theory,' worked for years but is collapsing as AI lets everyone copy faster and better.
YouTube's Semantic ID system reads a video's script, format, and title to detect when it's derivative, then throttles its reach for lacking originality.

05What 'good enough' means: 5-10% CTR, 30% retention

Aim for a 5-10% click-through rate to clear 10,000 to 100,000 views; that rate should hold close to 5% even past a million impressions.
Aim for roughly 30% average view duration regardless of video length: about 3 minutes on a 10-minute video, about 9 minutes on a 30-minute video.

06AI thumbnails: the Higgsfield workflow

AI-generated thumbnails can now look identical to photographed ones and can drive millions of views; the tool isn't the problem, low-effort use of it is.
Test multiple AI-generated angles and backgrounds before picking a thumbnail, then finish it in an editor rather than shipping the raw AI output.

07Retention tactics: visuals and storytelling

Add graphics, B-roll, and animation only where they clarify a point. Overloading a video with visuals hurts retention as much as having none.
Convert flat lists of facts into cause-and-effect narration, for example they didn't eat so they had to hunt but couldn't find anything, to keep viewers watching.

Glossary

Terms worth knowing.

Semantic ID: YouTube's newer classification model that analyzes a video's script, title, format, and music to judge how original it is compared to other videos, then suppresses reach on content it deems too derivative.
Outlier theory: An older YouTube growth tactic of finding a video that vastly outperformed its channel's average and copying its title, thumbnail, and format almost exactly.
Follow-up theory: This video's proposed replacement tactic: framing a new video as an unofficial sequel or rebuttal to a video already popular in the suggested tab, rather than copying it outright.
Click-through rate (CTR): The percentage of people who see a video's thumbnail as an impression and actually click to watch it.
Average view duration: The average length of time viewers spend watching a video, often expressed as a percentage of total runtime, also called retention.

Resources

Things they pointed at.

11:08toolHiggsfield ↗

Quotables

Lines you could clip.

04:09

“The best way I can explain it is think of it as creating the episode two of someone else's video, right?”

names the whole video's core tactic in one line→ TikTok hook↗ Tweet quote

04:13

“Should you just post a random video and hope the algorithm will bless you? I don't like that. It feels like I have no control.”

relatable frustration before the payoff→ IG reel cold open↗ Tweet quote

10:11

“The average view duration that usually pushes my videos beyond 10,000 views... is three minutes.”

a real, specific number viewers can screenshot→ newsletter pull-quote↗ Tweet quote

13:16

“Does this thumbnail look like AI slop? It literally got 1.6 million views.”

proof-then-punchline against a common objection→ TikTok hook↗ Tweet quote

The Script

Word for word.

Read-along

Don't just watch it. Burn it in.

See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.

metaphoranalogy

All right, so in the last few months, YouTube has slowly changed forever. And it's hard to notice because every change has happened gradually. So viewers don't feel overwhelmed.

But if you create content, these changes are affecting you. And if your videos are getting stuck below 2000 views, these changes might be the reason why your videos are getting stuck. So if you have a small YouTube channel and you want to grow on YouTube, you have to listen to this video because inside this video, I'm going to show you all the changes and how to adapt to them to grow.

Now, one of the craziest changes that some people didn't even notice is that we went from having almost 20 videos on the homepage to only having three. And the other row below is now with short form videos. If YouTube can only show three videos on the homepage, it needs to show videos.

that are already performing well or that are created by big creators. YouTube can't risk showing a new small channel on the homepage because if the video isn't good enough, the viewer might leave the platform. And that's not what YouTube wants.

So your chance as a small creator of appearing on the homepage is extremely low. So how do you get views? How do you grow as a small channel?

Well, you have to start optimizing your videos for the suggested tab. Let me explain. When a viewer enters YouTube, they see three videos on their homepage.

Sometimes they even see an ad as their first video, dropping the total number of videos they see to only two. Now, below these videos, YouTube is showing short -form content. As I said, these two or three videos are probably already viral videos from big creators.

YouTube shows them because the algorithm knows the viewer will probably watch them. YouTube doesn't just want the viewer to watch one video. YouTube wants the viewer to stay on the platform for longer.

So after the viewer clicks on one of these three videos and finishes watching it, the algorithm tries its best to keep the viewer on the YouTube platform by suggesting similar videos on the right side of the video they just finished watching. This is also where the second change comes in. Recently, YouTube started showing more and more videos from small channels in the suggested tab.

But here's something most of you didn't notice about the suggested tab. So when you watch a video on YouTube, on the right side, the first suggested video will probably be from a big creator that already has a lot of views. The second suggested video might be another viral video from a big creator.

But now, the third suggested video will probably be from a mid -sized creator with a decent amount of views. But then... a video with only 300 views or 1000 views from a random small channel will appear.

And after suggesting the small video from the small channel, the algorithm keeps showing big videos, only viral videos below it. So YouTube tries to help small channels by squeezing in this brand new video from a small channel next to these big creators. And that is amazing news for small channels like you.

because this means every small channel now has an opportunity to grow. All you have to do is package your video the right way and make sure that whatever you show inside your video is interesting enough for viewers to watch. But the problem is that you probably have no idea how to package your video and how to optimize your video.

So let me help you out. Now, this is a brand new theory and it works with any niche. It doesn't matter what videos you post, what topics you post.

It works with any type of channel. Here's how it works. So pay attention.

As I said, a viewer enters YouTube and it's greeted with three big videos. Now, let's say one of these videos, it's called Why Bananas Are Good For You. So this viewer clicks on this video and watches the entire thing.

But on the right, there are suggested videos. And as I said, these suggested videos already have a lot of views. But there is one slot for one small channel.

And that's where you, as a small channel, need to learn how to sneak in into the viewer's eyes. Let me show you how to do it. So how can you appear as a small channel in this tiny slot?

What can you do? Should you just post a random video and hope the algorithm will bless you? I don't like that.

It feels like I have no control. That is why I found an interesting way to make sure I do appear there. So here's how this tactic works.

You basically have to create some sort of follow -up to the video the viewer is already watching. The best way I can explain it is think of it as creating the episode two of someone else's video, right? It sounds pretty interesting, right?

And I haven't seen other people talk about it, but they will after they see my video. For example, If the current video a viewer is watching is called why bananas are good for you, your video can be called why bananas are bad for you.

And this is why this is different from the outlier theory, because you're not recreating the same idea. Imagine a viewer has just watched a 20 minute video about why bananas are good for you. When they look at a suggested tab, they see your video there called why bananas are bad for you.

It feels like the second episode of that video. And since the viewer already watched a 20 minute video about why bananas are good for you, the chances of the viewer clicking on your video is very high because your video really feels like a follow up video to that video. And this is a great way to make a viewer interested in your content by using a video that's already popular.

And this is so important to understand because before this, the tactic everyone used to grow on YouTube was the outlier theory. So what creators used to do... And what they're still doing is this.

They see a popular video called Why Bananas Are Good For You. And then a hundred other channels create the exact same video. They just change the title a little bit.

They change the thumbnail a little bit, but they essentially make the same video. And this is what the outlier theory is. And it has been dominating YouTube for the past.

five to six years. It's what every single creator has been doing. But recently, because of all these new YouTube changes, the outlier theory has slowly stopped working.

And this is why a lot of channels have stopped getting views or feel stuck. Because the outlier theory is slowly dying. And there's a good reason for it.

The outlier theory relies too much on copying people. And now because of AI and with the help of AI, creators can copy each other faster and better. And this is a huge problem for YouTube because when a lot of content feels the same, the platform starts to feel boring.

And if the platform feels boring, a lot of viewers exit YouTube. And as I said in the beginning of my video, YouTube doesn't want that.

YouTube wants viewers to stay as much as possible on the platform because if viewers stay on the platform, that means YouTube can show more ads. can make more money right so if the platform becomes boring and viewers leave youtube makes less money and so this outlier theory has become a huge problem for youtube and that is why youtube introduced something into their algorithm called semantic id which basically detects your entire script your format your title and what you're talking about what music you're using inside your video this new ai algorithm from youtube understands your video on a much deeper level almost deeper than you do and because of this it can easily notice if your video is inspired from another video.

And if the algorithm detects that your video is a little bit too inspired from another video, it's going to stop sending viewers to your video because your video is not original. And this is also another change YouTube is going through. It wants more original content.

The old advice is slowly dying. That's why if you want to grow on YouTube nowadays, I'd suggest using what I'm going to call the follow up theory. And the reason why I like this so much is because as a small channel, this gives you a system to follow.

You don't just randomly post videos and hope to grow. You actually have a system you can follow. which in my opinion makes YouTube easier because the worst part about it is when you just post random videos and hope it works.

So if you want to grow faster, I would suggest using the system. But even if you appear in the suggested tab, you still need to make sure your video is interesting enough to click on and you need to make sure your video is interesting enough to watch. And that's where you probably fail.

And I think that's because you don't understand what good enough means. What's a good enough video on YouTube, right? What would viewers and what will the algorithm consider a good enough video?

Well, let me help you with that as well. I think the most helpful thing I can do for you as a small channel is to explain what is the average click through rate and the average view duration you should aim for. So I'm going to use a 10 minute long video because it's easier to calculate.

Let's say you post a video on YouTube. and it's 10 minutes long. For this video to get more than 10 ,000 views, 50 ,000 views or 100 ,000 views, you want a click -through rate of 5 % to 10%.

This does fluctuate a little bit depending on your niche and your topic, but overall, if you hit 5 % to 10%, that is good enough for your video to get even 100 ,000 views. But here's the confusing part. As your video gets more impressions, your click -through rate drops.

So how low should your click -through rate drop for you to know that your click -through rate is still good? Well, I think the best example I can show is my videos. For example, if a video reaches 100 ,000 impressions or 500 ,000 impressions, my click -through rate is still around 5%, sometimes 6, sometimes 4.

So even if you get 1 million impressions, you want your click -through rate to still be close to 5%. Now, if you start getting 2 million impressions or 10 million impressions, obviously the click -through rate is going to drop below 3 % or 2%, but you get the idea.

But as I said, click -through rate is not the only thing you want in order for your video to get views. You also need a good average view duration. So for how long on average should the viewers watch your video for your video to grow?

Based on my experience of creating content for more than 10 years and owning multiple channels, the average view duration that usually pushes my videos beyond 10 ,000 views or 50 ,000 views and sometimes even over 500 ,000 views is three minutes. So if the video is 10 minutes long, viewers watch my video on average for three minutes.

That is a retention of 30%. And honestly, that's not so bad. That's actually super easy to achieve.

Now, if your video is 20 minutes long and your average view duration is still three minutes, that is too low. As I said, 3 minutes only apply to a 10 minute video.

So the longer your video gets, the higher that average view duration needs to be. Whether your video is 15 minutes long or 25 minutes long or 30 minutes long, you need to hit a retention of 30%. But how do you do this?

How do you get a click -through rate of 5 % to 10 % and a retention of 30 %? Well, let's start with click -through rate. For that, you need a good thumbnail and a good title.

Now, the title is pretty easy to come up with. But what about the thumbnail? Well, this is where...

things get tricky. Recently, AI has gotten extremely good at creating thumbnails to the point where you can't even tell whether they were made with AI or not. I use AI to create thumbnails myself, and my recommendation is that if you want to stay ahead, you need to use them as well.

Not for all of your thumbnails, but you should incorporate AI when it comes to creating thumbnails. Now, to help you out and keep things as simple as possible, I'm just going to show you how I create thumbnails for my channel so you can simply just learn from me. So recently for my thumbnails, I've been using Higgs Field a lot.

Higgs Field is an AI platform where you can choose from different models to generate images and videos. And here's the entire process. I basically go to Higgs Field and under dashboard at the top, I click where it says image and then I click create.

I choose the model I want, which is usually the ChatGPT model, and then I just tell it what I want to create. Sometimes I can also add a reference photo with myself if I want to generate myself with different backgrounds or different angles. After that, I just send the prompt and within a minute, I have a thumbnail that looks pretty insane.

So for example, on the screen, this is a thumbnail I created with Hicksfield. It basically looks identical to me. It has perfect lighting, perfect shading.

And if I don't like it, I can just go ahead and create another one with a different background or a different angle. It's very fast and it's very efficient. And this is the reason why I recommend using it, because it speeds up the process quite by a lot.

And if you don't use it, some other creators like me are going to use it and we're going to stay ahead. And now, basically, after I have the final image from Hicks Field, I usually just drop it inside Photoshop. I add text.

I fixed a little bit of the lighting, and here's how the final thumbnail looks like. On the left side, you can see the image created with Higgs field, and on the right side, you can see the image after I added text and done some modifications to it. And to go one step further and show you why this works so well, check this out.

I used this thumbnail in this video, and this video got 1 .6 million views. So it's up to you if you want to create thumbnails like this or not. There's going to be a link down in the description where you can try Higgs field yourself and start creating better thumbnails.

But now we need to talk about retention. How do you get viewers to watch at least 30 % of your video? Well, based on my experience, there are two things that seem to help increase retention.

First, visuals. Different scenes, graphics, text, animations, B -roll, whatever you can add to your video to make the footage more interesting, add it. But don't overwhelm the video with too many graphics or too much B -roll.

Just add it when it makes sense. And you can still use AI to help you create some of these animations and motion graphics. For example, I also use Higgs field to create some 3D visuals for my video or to create specific motion graphics.

And it's very useful. A lot of creators are afraid to use AI in their videos because they're afraid to be categorized as AI slop creators. But the thing is, with AI, you can actually create some cool looking things.

I mean, just look at the thumbnail. Does this thumbnail look like AI slop? It literally got 1 .6 million views.

So the problem is not AI. The problem is the people who use AI to create these things. But now what's the other thing that can help increase retention to 30 %?

Well, that's storytelling. And honestly, I think it's the most important one because storytelling is what basically makes a viewer think, wait, what's going to happen next? No matter your niche, no matter your topic, you need to have some sort of storytelling in your video.

But a lot of you have no idea how to add storytelling to your videos. So let me show you a very basic example. Let's say you make a video about prehistoric humans.

And your script, your video, sounds like this. They hunted animals. They gathered food.

They made fire. They stayed in the rain. They were cold.

They were walking. They were dying. They were blah blah blah.

You see, you're just listing things. It gets boring. There's no storytelling inside your video.

Now watch what happens when you turn this into an actual story. They didn't eat all day. So they had to go and hunt.

But they couldn't find anything to hunt. So they were very hungry. The rain started.

It got worse. And because it got worse, a lot of them died. So I don't think you notice, but I'm basically saying the same thing.

I just use words to make it sound interesting. Now, if you're really serious about growing on YouTube and getting more views, I suggest you watch this video over here because inside this video, I'm going to show you how you can add storytelling the right way to your videos. It's basically episode two of this video.

The Hook

The bait, then the rug-pull.

YouTube quietly rebuilt its homepage and its suggested-video sidebar, and if you run a small channel, both changes are why your view count feels stuck.

Frameworks

Named ideas worth stealing.

04:09concept

The Follow-Up Theory

Find a video already popular in your niche
Package your video as its unofficial sequel or rebuttal
Let the suggested tab's small-channel slot surface it to viewers who just watched the original

A replacement for copying viral videos outright: position your video as the 'episode two' of a video the viewer already watched, so it feels like a natural next click in the suggested tab.

Steal forpicking a topic for the next video on a small channel with no existing audience

08:32list