YouTube Shorts now pays for retention: what changes for clippers

YouTube Shorts now pays for retention: what changes for clippers

In March 2026, without much noise, YouTube changed the way it splits Shorts revenue. Before, the payout pool was divided in a way that came pretty close to "per head": a view landed, it went into the tally. A view that watched the whole clip and one that swiped away at second two were worth roughly the same come payday. That stopped being true.

Now revenue is weighted by watch time per impression. It still matters how many people the algorithm handed your clip, only now it also matters how much of each one the clip managed to hold. Retention, which used to mainly tell the algorithm whether to distribute you, has entered the revenue math itself.

This isn't cause for panic and it isn't some hidden trick. Honestly, it's a change that rewards anyone already clipping with care — and a well-chosen clip is, at its core, retention cut out and framed. In this piece I want to explain how the new ruler works today and, above all, how to build and review the clip, in practice, so it works for you.

How the math works now, no mystery

The simplest way to get it is to drop the idea of "a thousand views = X dollars." Two thousand views identical on paper can be worth different things, depending on what happened while they rolled. The weight moved from how many impressions to what happened inside each impression.

The main signal is completion rate: the slice of people who opened the clip and got close to the end. The numbers going around talk about something like 65% average retention as a healthy benchmark for clips under 30 seconds, and around 50% for longer ones. Don't treat that as law — it swings by niche, by format, by source — but the direction is consistent: the more people finish, the better the view pays. It's the same principle that always sat behind a strong hook in the first seconds and a caption that holds retention. The difference is that before, this care only improved reach; now it also improves how much the same view earns.

Why clipping fits this so well

It's worth understanding why, because it changes how you pick your material. A good clip is born from a moment that already gripped someone once — the line that closes an argument, the genuine reaction, the turn in a story. When you dig through a stream or a podcast for that peak, you're pulling out exactly the stretch that proved it could hold attention in its original context.

Content made from scratch bets it'll grip. A well-chosen clip starts from a stretch that already gripped. In a model that pays for watch time, that difference stopped being just "quality" and started showing up in RPM. It's no magic: you're publishing pieces with retention already baked in, instead of hoping it happens.

The right length is the one that maximizes time, not percentage

There's a hasty conclusion worth dodging: "if retention is everything, I'll make 8-second clips everyone finishes." The problem is that watch time is retention multiplied by length. Eight seconds watched by 100% of people delivers 8 seconds per person. Forty seconds watched by 60% delivers 24 seconds per person — triple the time, with a completion rate that's still good.

That's why the 30-to-45-second range showed up as a benchmark in 2026: long enough to stack up time, short enough for most to finish. It's not a fixed rule. Dense material — a good podcast stretch, an explanation — holds more; a shallow moment asks for less. I've written before about this tension between the 60-to-90-second rule and the paradox of stretching a clip without killing retention. What the change adds is that nailing the length now has an extra effect, beyond reach.

Three habits that started carrying more weight

None of this is brand new. These are old practices that gained a concrete weight they didn't have before.

Burned-in captions. A large share of the Shorts audience watches on mute, and the caption is what keeps that person following along. In the 2026 reports, captioned Shorts showed up with wider distribution, and it makes sense: the caption helps both reach and retention at the same time. If your flow still publishes clips without captions, that's the highest-return fix to make first.

The opening. The first second always decided whether the algorithm kept testing your clip. With revenue tied to watch time, the opening now also decides whether the impression it handed you earns anything. A clip that starts on the warm-up — ten seconds of "so, let me tell you" — loses people before it stacks up any time. Start right on the part that matters.

Cadence. The tests going around linked a daily streak of around 30 days to an RPM above the channel's baseline. It's not a guaranteed number, it's an observed pattern — but it's where the clipping pipeline has a natural edge: whoever turns one stream into several clips can post regularly without forcing shallow content just to keep the streak alive.

How to do this in practice, in Cut.Pro

Up to here it's analysis. The useful part is the process, and the goal of the flow is to spend your time where it moves retention — picking the moment and nailing the opening — and take the grunt work that doesn't move it off your plate. That's what we built the pipeline for:

1. Drop the stream or the VOD into Cut.Pro. A YouTube, Twitch, or Kick link works. You don't have to download or cut anything by hand.
2. Let the AI clipping sweep the material. It looks for the peaks — in audio, in reaction, in chat movement — and suggests the stretches that look most like clips. This is where the "baked-in" retention I mentioned above comes in: the system aims at moments that already held attention.
3. Check the vertical reframe. The framing follows face and action in 9:16, so the viewer is never left staring at an empty corner of the screen — one of the quiet reasons retention drops in the middle.
4. Set the length with watch time in mind. Use the generator's limits to keep clips in the range that retains, instead of letting them come out too short. If a stretch has legs, let it breathe; if it's just a flash, don't force the stretch.
5. Review the caption and the opening. The caption comes out synced and burned in; your job is to read the first seconds and make sure the clip already starts on the high point. This is the step that moves RPM the most, and it's the only one the machine doesn't decide for you.

The rest — transcription, cutting, reframe, captions — the pipeline handles, and that's exactly what used to make a daily cadence impossible. What's left for you is the part that decides the result.

A better way to measure from here on

If you take just one thing from this piece, take this: stop judging a clip by the views it piled up and start looking at the retention curve it held. YouTube Studio itself shows you which second the audience swipes away. That drop-off point stopped being only where you lose reach — it became where revenue leaks too.

In practice, it's the same care that always separated the well-made clip from the rushed one, and it's worth revisiting the mistakes that tend to bury a Short with that lens. The new ruler just made this care easier to justify: doing the clip right and earning more from it stopped being two separate goals.