I've built four faceless YouTube channels in the last eight months. The first three are dead. The fourth is at $3,400 a month in ad revenue and climbing.
Everyone will tell you the difference was luck. It wasn't. The difference was one specific thing I got wrong on channels 1-3 and right on channel 4. Not the AI clip generator. Not the voice cloning. Not the thumbnail template. The niche. That was the whole game.
I want to save you the four months I lost. This is what I actually do now to find a niche, why the first three failed, and the checklist I run before I let myself post a single video.
One thing before we get into it. I spent the first three months learning faceless YouTube off other YouTube tutorials. Every one of them said "pick a niche you're passionate about" or "just post consistently and you'll find your audience." Both are wrong for faceless. The right answer is uncomfortable and structural, not vibes-based. A structured course walked me through the actual method in about ninety minutes and I'll link to it below. YouTube tutorials are fine once you already know what you're looking for. They are a maze when you don't.
One more thing. There are three methods for finding a niche and I'll walk through all three. The third one is the least talked about and the one I use most now. Don't skim past it.
The checklist that decides whether a niche is even worth trying
Before I test any new channel idea, I run it through six questions. If it fails any of them, I don't touch it. This alone would have killed all three of my dead channels before I made a single video.
- Are there at least three small channels in the niche pulling more views than they have subscribers? That's the demand-vs-supply signal. If videos consistently outrun the subscriber count, the algorithm is pushing that content to a hungry audience. If views track subscribers 1-to-1, the audience is saturated.
- Are there any channels over 100,000 subscribers already dominating the niche? If yes, walk away. YouTube trusts an established channel more than it will ever trust yours in month one. You cannot outrank a 100k channel in the same lane.
- Are there small channels in the niche that are failing? If most of the small players are stuck at 200 views per video, that's the market telling you the algorithm doesn't want more of this. Don't try to be the exception.
- Is the niche younger than six months? Older niches have already sorted their winners. New niches (under six months old) still have algorithmic space.
- Can you make longer videos than your competitors? Aim for at least 1.5x the average length in that niche. Longer content earns higher RPM and gives the algorithm more surface to grab watch time.
- Is the niche monetizable at a decent RPM? Some niches (kids content, ambient sleep music) monetize at $1-2 RPM. Others (finance, tech, some history) hit $8-12. A $10 RPM niche pays five times more per view than a $2 one. Same effort, different bank account.
My first three channels failed at least two of these each. Channel 1 was in a niche with three 500k-subscriber channels already dominating. Channel 2 was in a niche older than three years. Channel 3 was in a $1.20 RPM lane where the math never worked.
Method 1. Retrain your YouTube homepage into a niche-discovery engine
This is the method I use for the first pass. Twenty minutes of setup, and your YouTube homepage becomes a machine that shows you faceless niches you never would have thought of.
Create a new Google account. Open YouTube in an incognito window and sign in. Your homepage is empty because YouTube has no viewing history on this account yet. That's the point.
Now search for anything broad you're interested in. Not a niche yet, just a topic. History. Cars. Finance. Sports. Click the first faceless video you find in results. While it plays, like the video, subscribe to the channel, and leave a comment. Do the same for two or three more faceless videos in the sidebar. Now scroll YouTube's homepage. Every channel it shows you from here on out is faceless because that's the only signal it has to work with.
Do this for thirty minutes. By the end, your homepage is a list of small-to-medium faceless channels you've never heard of. Most of them will be in niches you didn't know existed. Sort each one against the six-point checklist above.
I found the niche my fourth channel runs in on the third day of doing this. It was a small channel with 5,000 subscribers whose average video was pulling 55,000 views. Ratio was 11-to-1. Fresh niche, no big players, good RPM lane. Two hours of research to confirm the checklist, then I committed.
Method 2. Transfer a proven niche to an untapped variant
Second method. I use this when I want a channel adjacent to something I already know works.
Take a proven channel in one country and rebuild the same concept for a different one. There's a channel called "Unfolded United States" that does state facts. The same concept can be run for Canada, Australia, the UK, or any country the original doesn't cover. The audience exists, the format is proven, and the specific execution slot is empty.
Or take a proven vertical and transfer it to a sibling category. A car-storytelling channel becomes a boat-storytelling channel or a motorcycle-storytelling channel. Same narrative structure, different subject matter. Sometimes it doesn't take. When it does, you get a channel with the ceiling of a proven format and the space of an untapped market.
The rest of the method (script structure, thumbnail template, upload cadence) I got from a short course on Mindwand that breaks faceless YouTube into 15-minute daily lessons. Coursera and Skool have similar programs if either fits better. Pick one and finish it. Building the muscle for this kind of pattern-transfer takes practice and having a course to reference cut months off my learning curve.
Method 3. The X leaks method (my most-used, least talked-about)
This is the one nobody teaches. Everyone treats X (Twitter) as a distraction. It's actually a leak site for niche ideas if you know where to look.
Go to X search. Type "faceless YouTube" or "niche ideas" or anything similar. Then, and this is the important part, sort by Latest and scroll past the viral posts. You are NOT looking for the posts with 80,000 views. Those are already public knowledge. You are looking for posts with 30 or 40 views. Someone who thought they had a good idea, tweeted it into the void, got no traction, and moved on. That's a niche nobody has picked up yet.
I found two of the niches I now test-post in from posts that had under 50 views on X. Both were tweets from someone who mentioned a specific niche in passing, got zero engagement, and never followed through. I picked one up and it became channel 4.
The mental shift is that low-engagement posts on X are unclaimed ideas. Not because they're bad. Because whoever posted them didn't have the discipline to actually build. You do.
The honest math
- Channel 1 (motivational, dominated niche), 47 videos, $8 total, 4 months in
- Channel 2 (old niche), 22 videos, $2 total, 6 weeks before I quit
- Channel 3 (bad RPM lane), 18 videos, $37 total, 5 weeks
- Channel 4 (niche found via homepage method), 34 videos, $3,400/month current, 3.5 months in
Total spend across all four channels, about $300 in tool subscriptions. Total time, roughly 400 hours. Effective hourly across the four is $9/hr. Effective hourly on just channel 4 in the last two months is closer to $65/hr, and the trend is up.
If you were expecting a $10k/month screenshot, that's not what this is. Real numbers from someone who is 8 months in, not 8 years. Real numbers grow. Fake screenshots don't.
Anti-hype close
The reason I'm writing this is I keep seeing people grinding away on their fifth video in a dead niche wondering why nothing is working. It's the niche. It's almost always the niche. Better editing won't save the wrong niche. A better thumbnail won't save it. Paying for a course that teaches "the algorithm" won't save it either if you skipped the niche step.
If you're stuck, the highest-ROI thing you can do this week is retire your current channel and spend a Saturday running Method 1 above. That's the honest answer. The Mindwand course covers all six checklist items plus the full pipeline downstream. Not the only option out there, just the one that mapped closest to what actually worked for me. Same advice as always. Pick one, finish it.
I hope some of it lands.