A video idea from 2021 can go viral again in 2026
A lot of "new" trends are old ideas in new packaging
Viewers like it that way: familiar enough to trust, fresh enough to click
The catch is picking one your niche hasn't seen in a while
In a Mittalmar collection, set the upload date to before 2023 and sort by outlier ratio
That's a list of ideas that already worked
Then check if anyone has made one recently
Most creators research their niche by studying the top videos
That’s how you end up copying the wrong thing
We scored 3,158 videos in one niche against each channel’s own normal
What we found was not what we expected 🧵
Here’s how to find what’s working in your niche
1. Create a collection in Mittalmar
2. Add the channels in your niche
3, Export it as a CSV
4. Feed it to Claude
5. Ask what actually outperforms, and why
Minutes instead of weeks of scrolling competitors
GTA 6 Is Going To Print Insanely Hard
If you aren't betting on this already you're missing out
Channels have positioned themselves months ago to benefit from the huge demand wave
Create GTA 6 thumbnails with Mittalmar today
YouTube Diary #33
Working on building the best possible Claude Project for the channel atm
It ties directly into the current lackluster angles/ideation on my channel
Plan is to create god-tiered context that feeds the engine
Currently feeding 10K videos from niche + adjacent niche directly to Claude through Mittalmar
A self-correcting engine that learns over time and incorporates all the data is the key here
Scripts, ideation, titles, all run based on superior datasets
Going to be super fun to test this in practice going onwards
Current bottleneck is ideation so let's upgrade it 100x
YouTube Diary #32
AVD locked in good, but views lackluster on recent videos
Been building some assets that serves at bottom to feed die-hards, but now it is time to bring in some liquidity into the system
Going to test a new format now, we have a lot of data on the current format so makes sense to test something new now
Going to base my next project on an outlier I tracked from a competitor channel in Mittalmar
Found multiple outliers on the topic only a month old
Going for a more in-depth approach with 10x quality
Let's see how it goes!
If you ever feel a little stuck
It can be really useful to look into data
There's a hundred patterns you can't see
But it all exists
Use Mittalmar to analyze your niche like never before
YouTube Diary #31
Beginning YouTube Community Posts this week to keep audience warm
No idea if it helps or not, but with a large focus on building die-hards, it makes sense as a strat beside everything else
Will start posting BTS content
If anyone got any experience, let me know
Every YouTube Niche Has It's Hidden Rules
What is winning grammar in one niche
Is thrash grammar in another niche
With Mittalmar you can check what really works in your niche
@fyreinteractive I create videos about battles
Before I went through everything that happened after the battle
Now I end it at the peak (peak-end rule)
Outro = no longer loss of retention
Very important point
YouTube Diary #29
How Dead Videos Can Suddenly Wake Up
So I bet you haven't seen this that often
A video that got 2 engaged views on average for the last 5 years just woke up
The video had 70% audience still watching after 30 seconds, and 50% AVD, but were completely stuck at abysmal views
The video had terrible packaging for years
After returning from a 5 year absence, I I swapped the packaging
There was no change immediately
But the moment one of my other videos got traction, the video suddenly woke up
YouTube tested the new packaging because liquidity had arrived
It had an audience and an environment to test the dead video in
Goes to show that conditions sometimes needs to be right for a video
Which means sometimes a video just needs a little nudge
Not talking big big numbers here, but for this video it was a big difference
Creds to @mittalmaryt for identifying the issue with the thumbnail and for generating the new thumbnail too
Really cool to see that a video is never truly dead
It's the packaging that is dead
YouTube Diary #27
Found a sick YouTube hack yesterday
After my return to my channel, I havent really looked into my niche, what is working/not working
So yesterday I added 20 big channels from my niche into a folder in @mittalmaryt
Then I exported all the data as a CSV file
Then I gave that CSV file to Claude
It then ran an analysis of all the 4600 videos
That allowed me to get a snapshot of the working formats, recognition anchors, best working topics, etc
It took me under 20 minutes
I have now uploaded all this into my YT channel brain
Which means every video going onwards will have this data in mind when ideating
The best thing is that I can just update the collection every now and then and track the newest developments
Bullish on this!
YouTube Diary #28
Newest video's audience retention is in, and I am really happy with the results
Will definitely be reduced with more views, but the fact that we are 10% higher than other videos with the same views shows that all the changes we made have payed off
So fun to learn new things! The next thing to improve upon is definitely ideation.
And of course, we want to try and maintain this level if we can!
YouTube Diary #26
Audience retention keeps improving with a record breaking 47.3% yesterday on a 15 minute long video (still early data and will likely fall, but incredible result for me this far)
After my return 7 weeks ago, audience retention has climbed, with some videos deviating of course
The biggest change from video 1 to 2, was cutting drawn out outros
All outros now lead directly to the next video in a natural fashion so that the viewer don't even realize the next video is being pitched
This has improved the drop-off at the end drastically
From video 3, I started ending my videos with the peak-end rule proposed by @MarioJoos. That too has been a successful intervention.
I expect video 3 to perform really well eventually. Even though it flopped views-wise, I know this has great quality. I will give it more time, and the AVD here is very strong too for such a lengthy video
Video 4 was a big flop, because it took over 10 minutes to deliver on what was promised in the hook with no payoffs that completely killed momentum. But I learned a lot from this mistake, so very glad I made it!
Video 5 went viral, so it's AVD will be somewhat lower because it got pushed hard, but it also began at around 40%
Video 7 graph not in yet, but currently sitting at 47.3%
Bias: More die-hard subs = higher AVD too
In general, after pivoting all my videos to one recipe the AVD has started increasing which just goes to show that @wono_strategy recommendation of having one recipe for one audience works brilliantly. His framework of introducing one new thing for each video is also the model I have based everything on
What my videos still lack is definitely strong angles and remarkable ideas. Yes this is an evergreen niche and I am building a catalogue that will serve long compilation videos, but still, would like for some videos to break through hard from the get go like video 5
@RichardKistnen@vidsummit@MarioJoos Very cool! Yes a lot of the lessons on YouTube are universal but should still be tailored to the unique creator and their goals. Not easy, but very cool!
YouTube Diary #26
Audience retention keeps improving with a record breaking 47.3% yesterday on a 15 minute long video (still early data and will likely fall, but incredible result for me this far)
After my return 7 weeks ago, audience retention has climbed, with some videos deviating of course
The biggest change from video 1 to 2, was cutting drawn out outros
All outros now lead directly to the next video in a natural fashion so that the viewer don't even realize the next video is being pitched
This has improved the drop-off at the end drastically
From video 3, I started ending my videos with the peak-end rule proposed by @MarioJoos. That too has been a successful intervention.
I expect video 3 to perform really well eventually. Even though it flopped views-wise, I know this has great quality. I will give it more time, and the AVD here is very strong too for such a lengthy video
Video 4 was a big flop, because it took over 10 minutes to deliver on what was promised in the hook with no payoffs that completely killed momentum. But I learned a lot from this mistake, so very glad I made it!
Video 5 went viral, so it's AVD will be somewhat lower because it got pushed hard, but it also began at around 40%
Video 7 graph not in yet, but currently sitting at 47.3%
Bias: More die-hard subs = higher AVD too
In general, after pivoting all my videos to one recipe the AVD has started increasing which just goes to show that @wono_strategy recommendation of having one recipe for one audience works brilliantly. His framework of introducing one new thing for each video is also the model I have based everything on
What my videos still lack is definitely strong angles and remarkable ideas. Yes this is an evergreen niche and I am building a catalogue that will serve long compilation videos, but still, would like for some videos to break through hard from the get go like video 5
@RichardKistnen Basically for me, my videos contain battles. The trick then, is to end during the peak of the battle. Then use only 2-3 sentences after to finish the videos
Before, I would typically spend 2 minutes on the post-effects of the battle. But viewers would drop-off like crazy tbh