Competition Ratio: How to Pick a YouTube Niche with Data

Three niche examples used to sit in this post as proof that the competition ratio works: 95,000 monthly searches against fewer than 2,000 active channels, called 18.5x. 450,000 against 3,500, called 14.9x. 450,000 against 9,200, called 4.9x. Divide them yourself. The formula those numbers are meant to demonstrate gives 47.5x, 128.6x and 48.9x. Not one of the three survives its own arithmetic.

The instinct underneath is still right: compare demand against supply before you spend a year on a topic. The decimal point is what was fake. Here is what you can actually measure, using what YouTube publishes about itself, and how to choose between two niches without a number that cannot be sourced.

Why does the competition ratio fall apart when you try to compute it?

Because the numerator does not exist. The formula asks for monthly search demand divided by active channels, and YouTube does not publish monthly search counts for a topic. What it publishes is a relative band. A band cannot be divided by a channel count, and the division is exactly where the false precision enters.

YouTube's own research surface says this plainly. The Trends tab in YouTube Studio reports an audience interest level that "ranges from very low to very high", built from "videos that have been watched over 1,000 times per week and from the last 28 days". That is an ordinal signal over a 28-day window. It is not a monthly search volume, and YouTube never converts it into one.

So when a tool hands you "95,000 monthly searches", it is handing you a model's estimate, not a figure YouTube discloses. That is not automatically worthless — estimates can rank things usefully — but it stops being a measurement the moment you divide it and quote the result to one decimal place. There is a second tell. The framework circulates in incompatible versions — a 4x threshold in one place and 20:1 in another, one dividing by active channels and another by videos carrying the keyword in the title. Thresholds that disagree by a factor of five are not competing readings of a measurement; they are the sign that nobody is reading one. Seen that way, the three broken examples above are not a typo. A ratio built on an unsourceable numerator lands wherever the person quoting it needs it to land.

What does YouTube actually give you instead of search volume?

Three signals, all relative and all anchored to your own channel. None is a market-wide count, and that limit is the honest part of the answer. They tell you which way interest is moving and who is already being served, not how many people typed a phrase last month.

• Audience interest level. Studio, then Analytics, then Trends. A very low to very high band on topics your audience already watches, refreshed over a 28-day window.
• Saved searches. Terms you mark and track, so what you read is a direction over time rather than one snapshot you happened to catch.
• Your own traffic sources. Whether YouTube search or Browse is already bringing viewers to anything you have made near that topic.

Notice what all three have in common: they are comparisons, not totals. That is the shape of every honest answer available here, and it is enough to pick between two niches.

How do you compare demand against supply without inventing a number?

Rank instead of divide. You cannot compute a defensible ratio, but you can order candidate niches against each other with reasons attached — and ordering is all that choosing between two niches ever required. The output is a ranked shortlist, not a score.

Step 1: Read demand as a direction

For each candidate, record the interest band from the Trends tab and whether it is rising, flat or falling across the window. A falling band on a big topic is worse than a steady band on a small one, because you will be arriving after the audience starts leaving.

Step 2: Count the supply properly

This is the half you can do honestly, by hand. Search the topic on YouTube, filter to channels, and count only those posting on that specific topic within the last 90 days. A channel that covered it once two years ago is not competition. Ten channels shipping weekly is a harder room than a hundred dormant ones, and a raw channel count hides that difference completely.

Step 3: Rank, then write down why

Put your candidates in order using both columns — interest direction and active supply — and write one sentence per niche saying what decided it. When you revisit the list in three months, that sentence is what tells you whether you were wrong about the niche or wrong about your own consistency. A number would have told you neither.

This is also the line between niche research and keyword research, which answers a different question: keywords tell you how to title one video, while this decides what the channel is about.

Does a thin niche make YouTube push your video harder?

There is no documented mechanism that does this. The claim is intuitive — few answers, hungry system, bigger test — and YouTube's description of its own recommendations does not contain it. What the documentation names is narrower, and more useful once you stop expecting a bonus.

YouTube sorts its signals into two groups: "Viewer Personalization", which covers "the signals that help us understand a user's preferences", and "Content Performance", which is "all about how well the video performs when it's offered to viewers". Neither is a scarcity credit. That page does not mention competition between channels at all.

Satisfaction is real, but it is not a recent switch. YouTube states it works to "maximize long-term viewer satisfaction" and ranks search results on "how well the title, description, and video content match the viewer's search" and "what videos drive the most engagement for a search". A dated changeover from watch time to satisfaction appears on no YouTube page. This site used to attribute that changeover to creator-tool blogs; we took it out of the post that existed to explain it, and it does not belong here either.

What a tight niche actually buys is the thing YouTube does name: the viewers who finished your last video have the interest affinity your next one matches, so your back catalogue becomes your warmest distribution. That is audience overlap, not a cluster score, and it compounds without anything having to favour you.

When is a niche worth walking away from?

Four warning signs, and only one of them is about the size of the audience. Each is something you can check before recording, which is the whole point of doing this in advance rather than after fifty uploads.

• Interest is falling, not flat. A high band today means little if the direction is down across the window. You would be entering a room the audience is already leaving.
• Supply is active, not merely large. A hundred dormant channels is an opening. Ten shipping weekly on your exact topic is not, whatever the totals say.
• No monetization path. YouTube defines CPM as "how much money advertisers are spending to show ads on YouTube" and RPM as your own revenue "per 1,000 engaged views" — different denominators, so they are not interchangeable — and it publishes no per-niche breakdown of either. Anyone quoting you a niche CPM range is quoting a third-party sample. Our own breakdown of what CPM by niche can and cannot tell you keeps that distinction.
• You cannot sustain fifty videos. Demand opens the door and consistency keeps you in the room. If fifty ideas do not exist without repeating yourself, no ranking saves the channel.

Monetization has a hard floor worth knowing before you pick: the Partner Program asks for 1,000 subscribers with 4,000 qualified watch hours in the last 12 months, or 1,000 subscribers with 10 million qualified Shorts views in the last 90 days, and YouTube has said updates to the programme arrive on February 1, 2027. A niche whose viewers watch three minutes at a time reaches that floor very differently from one whose viewers watch twenty, which is the watch-hour arithmetic behind niche selection.

What should you check before you commit?

Six questions, none of which needs a number YouTube does not publish. Answer them for each candidate niche and the ranking usually writes itself — and where two candidates tie, the tiebreaker is almost always the last question rather than the first.

• Direction: is the interest band rising, flat or falling over the window?
• Active supply: how many channels posted on this exact topic in the last 90 days?
• Depth of supply: are the incumbents shipping weekly, or dormant?
• Revenue surface: is there an ad, affiliate, sponsorship or product path, and does the audience arrive in a buying mindset?
• Content depth: can you list thirty video ideas right now without repeating yourself?
• Your edge: what do you bring that the active incumbents do not — expertise, access, format or production?

If you cannot narrow far enough to answer the second and third questions cleanly, narrow again. "Cooking" is not a niche. "Healthy cooking" is barely one. "Meal prep for night shift workers" is a topic whose active supply you can actually count, which is the test that matters.

Key takeaways

• The competition ratio has no sourceable numerator. YouTube reports an audience interest level "from very low to very high" over 28 days, not monthly search counts, so any ratio quoted to a decimal place is a model's estimate dressed as a measurement.
• The three niche examples this post used to cite failed the post's own formula — 95,000 over 2,000 is 47.5x, not 18.5x.
• Rank candidate niches on interest direction and active supply instead of dividing. Ordering is all a choice between two niches ever needed.
• No YouTube page describes a scarcity bonus for thin niches. It names Viewer Personalization and Content Performance, and never mentions competition between channels.
• Satisfaction is documented as a long-term goal, not as a dated replacement for watch time. Count only channels active on your exact topic in the last 90 days — dormant ones are not competition.

Choosing where to concentrate is the decision underneath all of this, and it happens before you have a single analytic of your own to read. Gleam scores demand against competition for a niche using public YouTube data, so what you compare is measured options rather than two hunches with a decimal point attached. See what your niche scores.