YouTube RPM: Why Some Channels Earn 10x More Per View

If your RPM is stuck under $8, the fix is almost never more views, more uploads, or better production. RPM is set by who watches your videos, not by what you make, and changing the audience you attract can double or triple it without changing your output at all. That’s the core argument of Eddie Eizner’s RPM breakdown, where he opens the analytics on his own channels to show the same effort earning $2 per thousand views on one channel and $30 on another. The numbers below are his, shown on screen in the video.

What RPM actually is (and why it isn’t CPM)

RPM, revenue per mille, is what you earn per 1,000 views after YouTube’s cut. CPM is the advertiser-side number: what companies pay for 1,000 ad impressions before the split. YouTube’s official share on long-form ad revenue is 45%, with 55% going to the creator, and in one of Eddie’s examples the gap is visible in his own dashboard: advertisers paying $13 per thousand while he receives $5, since unmonetized views and the revenue share both come out of the advertiser number.

The distinction matters because creators chase the wrong lever. CPM is set in an auction you don’t control directly. What you do control is which auction your audience gets entered into.

The auction that sets your rate

Advertisers aren’t buying views; they’re buying customers. Companies pay YouTube to reach a specific audience, and when several companies want the same audience, they bid against each other. A niche’s RPM starts at, say, $5, a competitor wants those viewers, the bid moves to $6, then $8, then $10, and it climbs until the highest bidder takes the placement. That bidding war is the entire mechanism behind “high RPM niches.”

The bid ceiling is set by customer value. A tax-services firm can earn thousands from one client over decades, so it will pay heavily for a video reaching 30-year-olds about to buy a house. An advertiser selling to 10-year-olds has a bid ceiling of whatever a kid can ask their parents for. Same platform, same ad slots, wildly different auctions.

Three channels, three RPMs: the receipts

Eddie shows this side by side with his own analytics. A Minecraft roleplay channel he ran pulled a $555 video at a $2.26 RPM, because the audience, judging by the comments, is roughly 7 to 12 years old. A Daily-Dose-style compilation channel he voiced himself earned $10,000 on just over 2 million views, around $5 RPM, an audience maybe three to five years older. Then there’s Nick Invest, a five-month-old finance channel he highlights (not his own, and not verified as built with TubeGen): roughly 75,000 views per day at a $10-15 RPM, about $1,000 a day, covering taxes, mortgages, and car buying. Every part of it is AI-made, from the script to the voiceover to the images.

Read those three together and the pattern is hard to unsee. The Minecraft video took real production effort for $2.26 per thousand. The AI finance channel, with a middling voiceover Eddie openly calls not that high quality, earns five times the rate on a fraction of the views, because the viewer asking “how do I pay less in taxes” is worth more to advertisers than any number of kids. He also reports seeing analytics from ultra-wealthy-audience channels where RPM approaches $80 to $100. Those aren’t his channels, and the view counts are modest, but the auction logic scales that far.

The three levers that actually move RPM

Everything above compresses into three audience levers, and per Eddie, hitting even one makes money, while a niche that hits all three maximizes it.

Age: target viewers above roughly 25, the range where people hold real purchasing power. Region: aim for high-GDP-per-capita countries (the US, UK, Germany, Switzerland, Norway, Singapore and similar), which you can steer through topic choice: a US-news or NBA channel gets an American audience by definition. Wealth of topic: make content that people with money watch, like investing, property, taxes, or travel. Eddie’s travel example is telling: a channel called True Globe runs hour-long country-by-country videos, and one airline ad that converts a viewer into a miles-program member can be worth a lifetime of flights to that advertiser.

Video length is the one non-audience factor worth naming, and it works differently: longer videos mean more ads watched per view, which lifts revenue without changing what advertisers pay per ad. It compounds the audience levers rather than replacing them.

What this means for AI and faceless channels

The Nick Invest example cuts against the most common fear in the space. RPM has nothing to do with how a video is made; a fully AI-produced finance video out-earns hand-made kids’ content per view by 5x or more, because the auction only sees the audience. Production method doesn’t appear in the bidding.

One warning from Eddie’s channel-vetting: avoid the engagement-dead “AI slop” pattern, the mass-generated factory videos with zero comments. When you research a niche, English comments and real engagement are the signals that an actual monetizable audience is watching. That’s also the honest caveat on high-RPM niches: you still need proof that people watch the format. “Best mini yachts off the coast of Florida” targets a perfect audience, but only works if the view demand exists.

This is a research problem before it’s a production problem, and it’s what TubeGen’s Niche Finder, a channel database with revenue and engagement filters, is built for: checking a niche’s RPM signals and demand before you commit. From there the production stack is the standard faceless pipeline — the script writer drafts retention-structured scripts in a reference channel’s style, the voiceover narrates them, and the visuals generate the scenes — which suits high-RPM niches especially well, since finance and history content is narration-driven long-form. If you’re choosing a lane, the best niches breakdown ranks the options by exactly these economics, and the monetization requirements guide covers getting the channel eligible in the first place.

Audit your own channel in five minutes

Before chasing a new niche, run the audience test on what you have. Open your analytics and check three things. Viewer age: if the bulk of your audience is under 18, you’ve found your ceiling, and no upload schedule fixes it. Geography: a channel Eddie references was pulling 100,000-plus views per video at a $0.70 RPM because the audience was almost entirely from a low-GDP region; volume can’t outrun a low auction. Engagement language: scroll your comments the way an advertiser’s targeting system effectively does, since English comments from working-age viewers are the profile the high bids chase.

If all three levers point the wrong way, the move isn’t to grind harder in the same niche. It’s to pick a topic whose default audience is older, Western, and able to buy, then rebuild the content plan around it. The demand check comes first, the production second.

The short version

RPM is an auction over your audience. Advertisers bid up niches where viewers are older, Western, and able to buy, and they ignore how the video was produced. Pick the audience first, prove the demand, then let the production system handle volume. That’s how the same thousand views becomes $30 instead of $2.

Target the audience that pays. Validate the niche in TubeGen, then build the channel on it. Start now →