Facebook Reels: what they actually pay and when not to make them

Updated 08.09.2026 · Kostas Obuchow

A breakdown of Reels under Content Monetization: why 3 million views paid out 14 dollars, how audience geography kills RPM harder than format does, what a qualified view is, what length works now and when images earn more than video.

Short answer

Reels do not pay for views. They pay for qualified views from an audience advertisers bid on — and the gap between those two numbers runs into tens of times. A fresh community measurement: 3 million views on a Ukraine — Poland audience paid out 14 dollars. That is $0.005 per thousand. The same volume on a US audience would pay two orders of magnitude more.

So “are Reels worth making” has no single answer. The real question is: for which audience and in which niche — and format ranks third, behind geography and retention.

Reels RPM: what the number actually measures

RPM on Reels is revenue per thousand qualified views — not per thousand plays. Two pages can report an identical view count and a different RPM purely because one collected payable watch time and the other collected swipes.

That is also why comparing your Reels RPM against someone’s in-stream figure means nothing: long video with mid-roll ads sells different inventory into a different auction. The same page can run a healthy in-stream RPM and a token RPM on Reels in the same month, and nothing is broken.

Three things move the number, in this order: audience geography, retention, and how much of your reach the ad system actually filled. Format sits below all three — how much of the gap geography alone explains is worked through in Facebook RPM by country.

Where to read it: the monetization dashboard reports Reels earnings separately from in-stream. Averaging the two into a single number hides whichever format is actually paying.

What they actually pay

The numbers authors of Ukrainian and Polish pages are reporting right now sit between $0.003 and $0.03 per thousand views on Reels. Against that, plain image posts on the same page bring comparable or better money for a fraction of the work — community members report that RPM on Reels and on images has levelled out.

Two forces pressing on the rate right now, neither of them yours to control:

• Summer is traditionally the weakest season for ad budgets. Rates sag for everyone, and it is not a penalty against your page.
• Ukrainian traffic carries an extra constraint. Since the start of the war a share of international advertisers limits or switches off delivery in Ukraine — the auction thins out and RPM drops with it.

Which means: before you rebuild your content, check whether your audience has shifted. A halving is explained by geography twice as often as by format. On rates in detail — see Facebook RPM.

Geography beats format

The order of magnitude reproduces reliably: the US and UK sit at the top of the rate card, Western Europe in the middle, Eastern Europe at the bottom, the rest of the world at token levels.

The practical consequence: a video with a million views “somewhere” can pay less than a video with a hundred thousand views in the US. So before optimising your edit, open the country breakdown of your audience — that is usually where the answer to “why did revenue fall while views did not” lives.

Authors running English-language pages confirm it from the other direction: when the US share of viewers rises, RPM rises with it; when cheap traffic dilutes it, RPM falls no matter what you do to the video.

The qualified view

Meta does not pay for the impression itself. A view has to run long enough to count as payable — which is why a video that gets swiped past in a second earns nothing, however many such seconds accumulate.

Hence the central mistake: chasing views instead of retention. A million swipes is worth less than a hundred thousand watched through.

The working benchmark monetized authors quote: retention of 40–45% or better on fully original material. Below that line a video drops into the cheap segment and stays there.

Length and retention

“Shorter is better” no longer holds. Fifteen-second clips, the norm two years ago, no longer accumulate enough payable watch time.

• The floor is around one minute. Anything shorter makes sense only as a test.
• Past the two-minute mark starts working in your favour: more watch time collected at the same reach.
• Length depends on the subject. One niche runs on ninety seconds, another on five minutes; importing someone else’s number into your topic is pointless.

A technical note people ask about constantly: uploading long videos from a phone fails where the same file goes through from a desktop. If a video will not upload, try another device before you start cutting it.

When images pay better

Reels are not the only way to earn from a page, and right now not always the best one.

An image post with a caption costs a fraction of the labour and, on weak geography, delivers a comparable RPM. Pages built on curated images comfortably pull millions of views without a single video — and none of the Reels-specific risks apply to them.

The sensible order is: images as the low-cost monetization base, Reels as the reach engine on top. Betting everything on video makes sense only when your audience sits in expensive geography.

Why Reels carry more risk

• Someone else’s music and footage. The fastest route to an unoriginal-content restriction runs through video, not images.
• Template AI videos. Material assembled from one template with no voice of its own gets recognised, and it pays token amounts: two million views on that kind of video can return less than a dollar.
• Cross-posting the same video. One video, one page. Duplicating it across a network leads to a flag, not to multiplied revenue.

Questions from the community