How to find a YouTube niche
A niche is not a topic. It is an audience with attention and money attached to it, and you find one by reading what is already working inside it rather than by picking the subject you most feel like making videos about.
Pick three or four candidate subjects. For each one, find the channels that dominate it and the small channels beating their own size. If both exist, the niche has demand and it has room. If only the dominant ones exist, it is closed. If neither exists, there is no audience there at all.
Why "pick a niche you love" is bad advice
It is offered constantly and it skips the only question that matters, which is whether anybody is watching. Enthusiasm keeps you uploading for about four months. It does not create an audience that was never there.
The opposite advice is just as bad. Chasing whatever is largest puts you against channels with a decade of head start, and a new channel does not out-resource an established one. What you are actually looking for is the narrow gap between the two: a subject with proven demand where the current occupants are beatable.
A niche is an economic market of attention. Treating it as a hobby category is why most channels never find one.Frame 117
The two channel types that tell you everything
Every niche worth entering contains both of these. Their presence, and the gap between them, is the read.
An authority channel is one of the dominant channels in a subject. It tells you the format the audience has already accepted, the length they tolerate, and the ceiling. It is your map of what works, not your competition.
A breakthrough channel is a small channel pulling far more views than its subscriber count should allow. This is the important one. A breakthrough channel is proof that the audience will still watch somebody new, which is the single thing you need to know before committing a year to a subject.
Authority channels but no breakthrough channels means the audience is loyal to who already has it. Breakthrough channels but no authority channels usually means a fad with no durable audience underneath it. Both present means the niche is open.
Judge by outliers, never by subscriber counts
A subscriber count tells you what a channel achieved in the past. It says nothing about whether the subject is working now.
What tells you that is the outlier: a video that beat its own channel's normal performance by a wide margin. An outlier is the only signal on YouTube that separates the idea from the audience that was already there. A 2 million view video on a channel that averages 3 million is a quiet day. A 200,000 view video on a channel that averages 8,000 is the idea doing the work.
Which means the question to ask of a candidate niche is not "how big are the channels here", it is "are new ideas still breaking out here, and how recently".
Size is not the same as opportunity
Not every audience can produce a million views. Some subjects have a hard ceiling, and a channel serving 40,000 people well can be worth considerably more than one serving 2 million badly, because the advertising rate and the buying intent are not the same across subjects.
This is why a raw view count is a poor comparison tool across niches, and why it is worth measuring performance relative to the channel rather than in absolute numbers. A video doing 100,000 views in a small, high-value subject can be a far bigger event than a million in a broad one.
What to do with the shortlist
| Step | What you are looking for |
|---|---|
| Collect | Three or four candidate subjects, chosen for demand rather than preference |
| Map | The authority channels in each. Format, length, upload rate, what the audience already accepts |
| Test | Whether breakthrough channels exist. If nobody new is breaking in, neither will you |
| Date it | How recent the outliers are. A niche that last broke out two years ago is a closed one |
| Commit | One subject. Not three. Splitting a year across three niches produces three dead channels |
Where this fits
Everything above is reading, and reading is the part that is free. It is also the part almost nobody does, which is why most channels are started on a hunch and abandoned inside six months.
Metarch exists because this reading is slow by hand. It tracks outliers across channels and niches continuously, scores each one against the channel it happened on rather than against the platform, and shows the authority and breakthrough channels in a subject side by side. It is free for one channel.