By Adam Ivy. I have been on YouTube since 2008, spent eight years doing it wrong, then grew past 300,000 subscribers and generated over 3.5 million dollars from my content. Everything below comes from that scar tissue.
The YouTube algorithm is not a gatekeeper that rewards big channels and ignores small ones. It is a preference engine, a matchmaker that connects each video to the viewers most likely to watch it. YouTube is not building a profile of you. It is building a profile of the person watching you, and your job is to make it obvious who that person is.
YouTube isn't guessing. They're not rolling the dice. They're not favoriting the big channels and ignoring your small channel. I spent eight years stuck around 50,000 subscribers believing the deck was stacked against me, when the real problem was that I kept confusing the matchmaker.
What is the algorithm actually optimizing for?
It is optimizing for viewer satisfaction, not creator effort. The algorithm is simply a preference engine. It watches what people click, how long they stay, and what they watch next, then it serves more of what worked. When creators say "the algorithm changed," what usually changed is viewer behavior.
That means you never beat the algorithm. I stopped trying to game the system and I started to work with it, and that is when things moved.
How does YouTube decide who sees your video?
Through a tryout. Every upload gets tested on a small audience whose watch history suggests they might like it. Pass the test and the circle widens. Fail and the video dies quietly. Three signals dominate that test: whether your videos stay on a consistent topic, whether the same kinds of viewers keep showing up, and how your retention compares to other videos competing for the same audience.
This is why confusion kills reach faster than bad quality ever could. If your last five videos are about five unrelated topics, YouTube cannot figure out who to test your next video on. It is like walking up to somebody on the street and asking for directions, but refusing to tell them where you are trying to go.
What happens when you change topics?
You force the system to start learning again. Every pivot deletes the audience data YouTube has gathered about your channel and restarts the matching process. I spent years "experimenting" with different content styles and wondered why nothing compounded. I was resetting my own progress every month.
That does not mean you can never evolve. It means you evolve deliberately, one adjacent step at a time, instead of whiplashing between niches.
Are you building an audience or earning an introduction?
You are earning an introduction. The audience you want already exists, already watches videos in your topic, and already has viewing habits YouTube understands. Your job is to be so clear about who you serve that YouTube can confidently introduce you to them.
Here is the mindset shift in one table:
| Fighting the algorithm | Working with the preference engine | |
|---|---|---|
| Mental model | Gatekeeper that picks favorites | Matchmaker that connects viewers to videos |
| Strategy | Chase hacks and upload tricks | Make one audience's preferences obvious |
| Topic choice | Whatever might go viral | Consistent lane the system can learn |
| When views stall | Blame the platform | Check packaging and retention signals |
What should small channels do with this?
Get clear, stay in your lane, and let the data accumulate. Real growth compounds slowly, then it happens all at once. Many creators quit one or two videos before YouTube would have pushed them, not because they were failing, but because they were staring at the wrong numbers while the system was still gathering data.
Still have questions?
Does YouTube favor big channels? No. Big channels have more accumulated viewer data, so the matchmaker takes fewer risks matching their content. A small channel with a clear audience and strong retention gets tested and pushed the same way.
Do shadowbans exist? In my experience the shadowban is a story creators tell themselves to avoid accountability for content that failed its tryout. Fix the packaging and the retention before you blame a conspiracy.
How long until a niche change stops hurting? Expect the system to relearn over a stretch of uploads, not days. Every consistent video in the new lane rebuilds the viewer profile. The more decisive and consistent you are, the faster the introduction gets made.