Key takeaways
- YouTube does not pay a rate per view. It pays you a share of what advertisers spent against your videos, and the number in your dashboard is the result of that division, not a price list.
- The split on long-form advertising is 55% to you, 45% to YouTube. Shorts run on an entirely different mechanism — a pooled fund, allocated by your share of engaged views, of which you keep 45%.
- CPM is what advertisers pay; RPM is what you earn. RPM is divided across every view, including the ones that never served an ad, which is why it lands well below CPM on every channel.
- Every niche-by-niche pay chart online is reconstructed guesswork. Two of them, published the same year, put US finance CPM at $12–$45 and above $60 — because nobody outside your own Studio has the data.
- Between Studio and your bank sit four deductions: invalid-traffic adjustments, the $100 AdSense threshold, the monthly payment window, and US tax withholding that reaches 24% if you never filed a form.
- RPM is mostly decided by who watches you and what you make. Views are the term you can move this week — which puts the money back on the packaging.
The question is always asked as a rate. How much does YouTube pay per 1,000 views — as though there were a published figure, the way there is a published minimum wage. There is not, and the reason is not secrecy. It is that YouTube never agreed to pay you for views in the first place. It agreed to run an auction against your video and hand you a fixed share of whatever that auction cleared.
Everything that makes the answer feel slippery follows from that one structural fact. Two channels with identical view counts can earn amounts that differ by a factor of twenty, and neither of them is doing anything wrong. The advertiser demand behind an American personal-finance viewer is not the advertiser demand behind an entertainment viewer in a market where nobody is bidding, and the split takes its percentage of whichever one you have.
So this is the question worked out as arithmetic rather than answered with a number: what YouTube actually shares with you and in what proportion, what the two metrics in your dashboard genuinely measure, the full chain from an advertiser's bid to money in your account, and the four things that quietly take a cut on the way. Where a figure exists, it is here with its source. Where one does not exist, this post says so rather than inventing a table.
Why there is no pay-per-view rate, and never was
Advertising on YouTube is sold by auction. An advertiser bids to reach a kind of person, that auction resolves in the moment your video loads, and the clearing price depends on how many advertisers want that particular viewer at that particular moment. A viewer in a wealthy market being served a video about mortgages in the second week of December is expensive. The same person watching a gaming clip on the eighth of January is not.
Your earnings are that auction result multiplied by your share. Nothing in the system contains a rate card, which is why the honest answer to "how much per 1,000 views" is a distribution rather than a figure, and why every site publishing a confident niche table is reconstructing it from creator screenshots and survey responses rather than reading it out of YouTube's books.
The scale of the disagreement between those tables is the tell. Search for 2026 CPM figures and you can find US personal finance quoted at $12 to $45 in one widely circulated chart and above $60 in another, with entertainment somewhere between $1 and $6 depending on who you ask. Those are not refinements of each other; they differ by five times on the same niche in the same year. This is the same structural problem as the invented click-through-rate benchmarks covered in what counts as a good CTR: the underlying numbers never leave individual Studio accounts, so anything presented as an industry average was assembled from whatever creators volunteered.
What can be checked is the size of the pot. Alphabet reports YouTube advertising as a line item, and its second-quarter 2026 results put it at $11.06 billion for the quarter, up around 13% year on year, after $9.88 billion in the first quarter. Neal Mohan's 2026 letter to creators put the amount paid out to creators, artists and media companies at over $100 billion across the previous four years. Those are real numbers about a real revenue share. They tell you the mechanism works; they tell you nothing about your channel.
What YouTube shares, and in what proportion
There are several distinct revenue mechanisms behind the single "estimated revenue" line in Studio, and they do not share a formula. It is worth knowing which of your money comes from which, because the levers are different.
| Source | Your share | How it is calculated |
|---|---|---|
| Long-form advertising | 55% | Share of the ad revenue earned against your videos, after YouTube's 45%. |
| Shorts | 45% | Of your allocation from a pooled fund, not of ads on your Short. Explained below. |
| YouTube Premium | 55% | Subscription money pooled and distributed by Premium watch time, not by impressions. |
| Memberships, Super Thanks, Super Chat | ~70% | Direct payments, roughly 70/30 — and a purchase made inside an app can carry the app store's own cut first. |
| Shopping affiliate commissions | Varies | Set by the merchant's programme, not by YouTube's split. |
| Sponsorship sold directly | 100% | YouTube takes nothing. It is also the only one where you do the selling. |
The Shorts mechanism is the one most often described wrongly, and YouTube's Shorts monetisation documentation is unusually specific about it. Revenue from ads between Shorts, and the Shorts portion of Premium, goes into a Creator Pool. Music licensing is paid out of that pool first: a Short using one track sends 50% of the revenue associated with its views to rights holders, and a Short using two sends 66%. What remains is allocated to creators by their share of total engaged views on monetising creators' Shorts, per country — if your Shorts account for 5% of eligible engaged views in a market, you are allocated 5% of that market's pool. Of your allocation, you keep 45%.
The detail worth holding onto is that the music deduction happens to the pool, not to you. Your allocation is calculated on 100% of your engaged views whether you used a track or not, and your 45% rate does not change. Using music does not cost you individually; it costs the pool everyone draws from.
The mechanism explains the earnings gap
Creators repeatedly find that Shorts earn a small fraction of what long-form earns per thousand views, and treat it as a policy decision against Shorts. It is closer to an arithmetic consequence. Long-form runs an auction against your specific video and gives you 55% of what it clears. Shorts hand you a proportional slice of a communal fund at 45%, after music licensing has been taken out of it. Two different machines, and only one of them lets a high-value audience bid up your particular video.
CPM and RPM: which one is your pay
Studio shows you both, they are quoted in the same units, and they answer different questions. Getting them the wrong way round is the single most common reason a creator's expectations are out by a factor of three.
CPM is the advertiser's number: the cost per thousand ad impressions, measured before YouTube's revenue share. It describes the market for your audience. Playback-based CPM narrows that to the playbacks where an ad actually ran — the example in YouTube's ad revenue documentation works it as $7 of revenue across 1,500 monetised playbacks, giving $4.67.
RPM is your number: total revenue per thousand views, after YouTube's share, and across every view you got rather than only the monetised ones. It also includes revenue that has nothing to do with advertising — memberships, Premium, Super Chat and Super Stickers all land in it.
Two things follow. First, RPM is always meaningfully lower than CPM, and that is not a sign anything is wrong. It has had the 45% removed and then been divided across a larger denominator. Second, RPM is the only one of the two that answers "how much do I earn per thousand views", because it is definitionally that question.
The gap between the denominators has a name: monetised playbacks, the views that served at least one ad. Views miss out for ordinary reasons — the video was not eligible for the full range of advertising under the advertiser-friendly guidelines, or no advertiser bid for that viewer, or the ad simply did not fill. No channel monetises anything close to all of its views, and the proportion that does is one of the largest hidden variables between two channels with the same CPM.
The full chain, from a bid to your bank
Here is the whole path as one calculation. The inputs below are illustrative — they are chosen to be round, not typical, and the point is the structure rather than the values.
| Step | Working | Result |
|---|---|---|
| Impressions served | The algorithm's decision, not yours | 200,000 |
| × click-through rate | 200,000 × 5% | 10,000 views |
| × monetised playback rate | 10,000 × 60% | 6,000 monetised playbacks |
| × playback-based CPM ÷ 1,000 | 6,000 × $8 ÷ 1,000 | $48 gross ad revenue |
| × your share | $48 × 55% | $26.40 to you |
| ÷ all views × 1,000 | $26.40 ÷ 10,000 × 1,000 | RPM of $2.64 |
An $8 CPM became a $2.64 RPM, and nothing unusual happened. The 45% share took its portion, and then the result was spread across four thousand views that never carried an ad. That ratio — RPM landing at roughly a third of CPM — is unremarkable, and it is why creators who plan their income off a CPM figure they read somewhere are consistently disappointed.
Read the chain from the top, though, and the more useful fact appears. Four of the six lines are outside your control on any given day: impressions are the algorithm's decision, the monetised playback rate is set by ad supply and eligibility, CPM is set by the auction, and the 55% is fixed. The line you can act on this week is the second one. Click-through rate is the only term in the calculation that responds directly to a decision you make in an afternoon, and it multiplies everything below it.
The four things that actually set your RPM
Geography. The single largest factor, and the one creators have least influence over. An audience concentrated in markets with heavy advertiser competition earns multiples of an audience of the same size elsewhere. This is not a judgement about viewers; it is the price of reaching them.
Subject. Advertiser demand varies enormously by what the video is about, because the value of a converted customer varies. A viewer who might open a brokerage account is worth more to bid on than a viewer who might not buy anything. This is the axis all those disagreeing niche tables are trying to describe, and the direction they agree on — finance and business high, entertainment and gaming low — is sound even where the magnitudes are not.
Format. Long-form and Shorts run the two different machines described above. Watch time on Shorts is not a smaller version of watch time on long-form for earnings purposes; it is a claim on a different pot.
Season. Advertising budgets are annual, and they are spent unevenly. The fourth quarter — with retail, gift-buying and year-end budget flushing all landing at once — is consistently the strongest period of the year for ad rates, and the first weeks of January are consistently the weakest as budgets reset. The magnitude quoted for that swing varies wildly between trackers, so treat any specific percentage you see with suspicion, but the shape is not in dispute. It is worth knowing simply so that a January revenue graph does not read as a channel problem when it is a calendar one.
Notice what is missing from that list: subscriber count. Subscribers do not carry a rate, and the split does not improve as a channel grows. A million-subscriber channel and a ten-thousand-subscriber channel in the same niche and the same market are on identical terms.
The levers on the video itself
Three decisions in your control genuinely change ad revenue, as distinct from changing views.
The first is the eight-minute line. Mid-roll ad breaks require a video of at least eight minutes; below that, a video can carry pre-roll and post-roll but no mid-rolls. That is a hard threshold, and it is the reason so much of YouTube's long-form catalogue sits just past eight minutes. It is also the most abused lever on the platform, because padding a six-minute idea to eight costs retention, and retention is what earns the impressions that produce views in the first place. Stretching a video to clear the line is a trade, not a free gain, and it is only worth making when the extra minutes hold up.
The second is ad placement. YouTube changed how mid-rolls are positioned in May 2025, moving them towards natural pauses and transitions and away from mid-sentence interruptions, and it published the reasoning alongside an experiment run in July 2024: channels running automatic mid-roll slots in addition to manual ones saw an average of over 5% more ad revenue than channels using manual placement alone. That is a modest, credible, first-party number — and one of very few in this entire subject area.
The third is eligibility. Videos that fall foul of the advertiser-friendly guidelines can be limited to a reduced set of advertisers, which shows up as a normal-looking view count attached to a disappointing revenue line. If a specific video's RPM collapses while the rest of the channel holds steady, check its monetisation status before rewriting your content strategy.
What the August 2026 view-count change does to the number
From 24 August 2026, YouTube counts a view the moment playback begins across every format, aligning long-form and live with the way Shorts have always been counted. The older, stricter definition survives in Analytics as engaged views.
YouTube has been explicit that this is a measurement change rather than a monetisation one: earnings and Partner Programme eligibility are unaffected, and monetisation continues to run on engaged views and engaged watch hours in advanced mode. Nothing about your revenue changes because of it.
What can change is a ratio you use for planning. RPM is revenue divided by views, and if the count in the denominator becomes more generous while the revenue in the numerator does not, the displayed rate falls without anything having gone wrong. Whether that applies to the specific figure in your own dashboard is worth confirming against your own data across the boundary rather than assuming in either direction — and it is a good reason to be careful comparing an RPM from July 2026 with one from September. The 2026 platform changes explained in full cover what else moved this year.
The dashboard number is an estimate until it is not
Studio's revenue line is labelled estimated for a reason. The previous month's earnings are finalised in AdSense for YouTube around the 7th to the 12th, after adjustments for invalid traffic, Content ID disputes and certain campaign types. Small movements between the estimate and the final figure are routine. A large drop is a signal worth investigating, because it usually means a chunk of your traffic was judged invalid.
Between Studio and your bank
The last stretch is administrative, and it is where a first payment gets delayed by months for reasons nobody warned the creator about.
| Stage | What happens | Why it delays people |
|---|---|---|
| Finalisation | Last month's earnings confirmed in AdSense, roughly the 7th–12th | Invalid-traffic removals can shave the estimate |
| Threshold | $100 minimum balance before any payment is issued | Small channels accumulate for months first |
| Address verification | A PIN posted to your physical address, entered in AdSense | Ordinary post; weeks, and it can go astray |
| Payment window | Issued between roughly the 21st and 26th of the month | Missing the threshold by a dollar costs a full month |
| US tax withholding | Tax information required from creators outside the US | The expensive one — see below |
The tax step deserves its own paragraph because it is the one that costs real money through inaction. Creators outside the United States are required to submit US tax information through AdSense. With a valid form on file, withholding applies only to earnings from US viewers, at the rate your country's treaty with the US provides — commonly between zero and 15%. Without one, reporting on the policy has consistently described withholding of up to 24% applied to total earnings rather than only the US-sourced portion, and creators in countries with no treaty face 30% on their US-sourced revenue. YouTube's guidance on US tax requirements is the place to check your own position. Filling in a form correctly once is the highest hourly rate available anywhere in this article.
Running your own numbers
Because RPM is the honest unit, the useful exercise is not to look up someone else's rate but to multiply your own by the traffic you actually have. Find your RPM in Studio under Revenue, then read across:
| Views | RPM $1 | RPM $3 | RPM $6 | RPM $12 |
|---|---|---|---|---|
| 10,000 | $10 | $30 | $60 | $120 |
| 100,000 | $100 | $300 | $600 | $1,200 |
| 500,000 | $500 | $1,500 | $3,000 | $6,000 |
| 1,000,000 | $1,000 | $3,000 | $6,000 | $12,000 |
That table is multiplication, not a forecast — the columns are not claims about what any niche earns. Its purpose is to make the shape of the thing visible: at a plausible mid-range RPM, a hundred thousand views a month is a few hundred dollars, and the first meaningful income arrives somewhere around the point a channel is doing that every week. Which is also roughly the point at which advertising stops being the main line. Sponsorship, memberships, products and affiliate commissions have no 45% deduction and no auction, and for most channels earning properly they overtake ad revenue well before the ad revenue becomes a living.
None of it starts before the Partner Programme, and that door moves on 1 February 2027, when the watch-hour requirement doubles. The new monetisation thresholds and the arithmetic of reaching them are worked through separately.
The term you can actually move
Go back to the chain. Your share is fixed at 55%. Your CPM is set by an auction over an audience you assembled over years. Your monetised playback rate is ad supply. Your geography is your geography. Four of the multipliers are effectively constants on any timescale shorter than a year of deliberate repositioning.
Views are not. Views are impressions multiplied by click-through rate, and while impressions belong to the recommendation system, click-through rate belongs to the two rectangles you publish alongside the video. Doubling a click-through rate from 4% to 8% doubles the views, doubles the ad revenue, doubles everything downstream of it — with no change to RPM at all, because RPM was never the term that was moving. This is why so much of the earnings advice online is aimed at the wrong end: chasing a higher-paying niche is a multi-year identity change, while fixing the packaging on the video you publish on Thursday is an afternoon.
The other half is holding the click once you have it, because the recommendation system decides tomorrow's impressions from today's watch time — the compounding described in what the first thirty seconds decide. Impressions earned by retention are the only kind that keep arriving, which makes retention the slow lever and packaging the fast one. You need both; they just operate on different clocks.
If the practical obstacle is that producing thumbnails takes long enough that you ship one and never test a second, that is a production problem rather than a design one, and it is the problem Thumblore was built for — three variants in the time it takes to open an editor, so the test is cheap enough to actually run. Before publishing either way, it is worth checking how the artwork reads at the size it will really be seen, which the thumbnail preview tool does in the browser. And if you want the tactical layer rather than the theory, the thirty-six tactics that move views is the companion piece to this one.
How much does YouTube pay per thousand views? However much the auction cleared, times 55%, divided across every view including the ones nobody bid on. That is the entire formula. It is not a rate you can look up, but every part of it is a number you can read in your own dashboard tonight — which is more than can be said for any of the charts claiming to know it on your behalf.