Finding Candidate Channels

You can't reverse-engineer a channel you've never seen. So before any of the judging happens, you need a pile of names in front of you, and most people get about four of them by remembering channels they already watch. That isn't research. That's your own viewing habits handed back to you.

So let's source properly.

Start with the browser, because this one decision poisons everything downstream if you get it wrong. Do all of it from a private window, logged out. The YouTube you use every day has months of your own behaviour cooked into it, and it will happily serve you a version of the niche that's really a version of you. You want the space as a stranger sees it. Clean session, or don't trust what comes back.

Then the manual sweep, which is dull and still the most dependable hour you'll spend. Put your niche keywords into search. Anything that looks even vaguely like what you're trying to build, middle-click it and let the tabs stack up. Don't stop to assess. Assessing while you gather is how you finish the hour with four channels and no list. Gather wide, throw things out later.

You can hand part of this to AI too. Give it the niche, give it a subscriber range and a views floor you actually care about, and ask for a shortlist. Then push it a step further and have it read across everything you've collected: what do these channels all do the same way, and what's nobody doing. It'll be wrong about some of them, and that's fine. It's generating candidates, not making the call.

If you've got a tool like NexLev, there's one filter worth knowing by heart. Pull the channels similar to one you like, then sort for the smallest subscriber counts sitting next to the highest average views per video. Huge audience with huge numbers tells you nothing. A tiny audience with huge numbers means the content itself is carrying it, and that's a channel you've found before the rest of the niche has.

Now the part almost nobody does, and it's the one that keeps working after you close the laptop.

Train the account you research from. Whatever profile you're using for this, subscribe to your niche and nothing else. Watch your niche and nothing else. No music, no sport, no whatever you had on last night. Give that two or three weeks and your home page stops being a home page and turns into a competitor dashboard. Uploads from the channels you care about arrive on their own. Topics catching fire in your space show up before you'd have thought to go looking for them.

The same trick works on shorts and it works faster. Sit in the feed, like the ones from your target niche, drop a comment on the channels you want more of, swipe straight past everything else. The system reads that as appetite and buries you in more of it. A week of feeding it deliberately and you've got an endless scroll of exactly the format you're studying.

None of this tells you which channels are worth modelling. That's the next step, and most of what you've collected won't survive it. What this gets you is a list long enough that the survivors are actually worth having.

Chase them down once. Then let them come to you.