The Shorts Tail Is Real. It's Just Much Smaller.

By Gary Mason · September 2026

I grouped every Short I've published on two channels by the week it went out, then read the median views for each group.

The batch that was eight days old had already caught up with the batch from two months earlier.

Not "was approaching." Had caught up. On my YouPush channel the newest group sits at 233 views a video and the group from nine weeks earlier sits at 262. On my acting channel, which is a completely different audience and nearly six years old, it's 725 against 882.

Nine extra weeks of sitting on the platform bought almost nothing. And it landed the same way on two channels that have nothing to do with each other.

How long YouTube Shorts keep getting views is a question about the shape of a tail: whether a Short quietly accumulates for months, or does its work in the first week and then mostly stops.

How long do YouTube Shorts keep getting views?

On my channels, the honest answer is roughly ten to twenty percent more than they had after their first week or so, and that's it.

That's a smaller number than almost anything you'll read, but it isn't zero, and I want to be careful about that, because "Shorts don't compound" is a snappier claim than my data actually supports.

Here's what I measured. Public Shorts only, grouped by the week they were published, median lifetime views per group, groups smaller than three videos dropped.

YouPush (channel about two months old at the start of the window)

Actor Myth Busters (nearly six years old, unrelated niche)

Look at those columns rather than at any single row. They bounce around a lot. What they don't do is climb with age.

The study that says I'm wrong

There's a much bigger piece of research that points the other way, and I'd rather put it in front of you than hope you don't find it.

Socialinsider analyzed 3 million Shorts across 56,000 accounts over 2025. Their finding is that Shorts gain momentum much later than people assume, with the biggest single bucket landing at day 25 to 30. In their numbers, the first five days account for only about eleven percent of a Short's first-month views.

That is not a small disagreement with what I just showed you. They have three million videos across fifty-six thousand accounts. I have two channels and about a hundred and twenty Shorts. On sample size it isn't a fair fight and I'm not going to pretend otherwise.

But their numbers and mine can both be right, and the reason why is worth more than either of us being the winner.

Averages and medians are measuring different things

Their per-bucket figures run like this: days 1 to 5, an average of 833 views. Days 5 to 10, 1,786. Days 10 to 15, 150. Days 15 to 20, 555. Days 20 to 25, 500. Days 25 to 30, 3,310.

Read that sequence again. Day 10 to 15 collects 150 views and day 25 to 30 collects 3,310.

That's not a curve. A curve doesn't collapse to 150 and then jump twenty-two-fold. That's the signature of a small number of Shorts going late-viral and dragging an arithmetic mean upward, in a set of three million where a handful of enormous outliers can move the average for everyone.

I'm reporting medians. The median Short is unaffected by the one in ten thousand that takes off in week four.

So I don't think the study is wrong. I think it's answering "what happens to the average Short across the whole platform," and I'm answering "what happens to my typical Short." Those are different questions and they have different answers, and almost nobody writing about this distinguishes between them.

The actual test

Here's the part that convinced me, and it's a falsifiable prediction rather than a vibe.

If the study's distribution described my channels, then a batch that's eight days old should be sitting at roughly a third of what it'll eventually have at thirty days. Which means my two-month-old batches ought to be somewhere around 2.7 times higher than my newest batch.

Observed: 1.12 times on YouPush. 1.22 times on the acting channel.

The tail is real. It's just worth ten or twenty percent on my channels, not a hundred and seventy. And my oldest groups are 64 and 78 days old, well past the thirty-day window, so if anything that comparison is generous to the study.

That's the claim I'll defend. Not "Shorts don't compound," which overstates it. Something more like: if you're planning around a back catalogue that keeps earning, plan around ten or twenty percent, not a multiple.

Three things that would make me wrong

I'd rather list these than have you find them.

My groups are small. Most of my YouPush cohorts are four to seven videos. A median over six Shorts is a noisy thing to build an argument on, and the acting channel's May 25 group, sitting at 67 views against a channel norm in the hundreds, shows exactly how much a single bad week can move a row.

Both channels are small. Socialinsider's sample covers accounts of every size. Late spikes may well be something that happens to big channels with deep search demand, and I wouldn't be able to see it from where I'm standing.

My newer Shorts might just be better. This is the one that genuinely cuts against me. I've been iterating hard, so if the recent ones are simply stronger, some real late accrual could be hiding underneath a rising quality gradient. My newest batch matching my oldest could be a good new batch cancelling out a real tail. I can't separate those with the data I pulled.

What I'd need to settle it properly is daily analytics data per video rather than lifetime totals read once, and that's a different API scope I haven't pulled yet.

So what do you actually do with this

Stop waiting on old Shorts. Whatever the true number is, nobody's data supports treating a Short like an asset that keeps paying for months. Each one gets its run and then you're mostly done.

That's freeing rather than depressing. It kills the habit of checking a three-week-old Short hoping it'll finally take off.

And it means volume isn't the lever you think it is. If every batch lands in roughly the same band no matter how many you publish, more of the same thing gets you more of the same band. I dug into that separately when I looked at whether scheduling Shorts affects views, and the same lesson turned up: the thing that moves is which Short, not how many.

Then go and measure your own. Export your Shorts, group them by publish week, take the median, and read the column. It's an afternoon and it beats taking anyone's word for it, including mine. If yours climb with age, brilliant, you're in the study's number and you should plan differently to me.

I build a tool called YouPush that handles the packaging and scheduling side of this, and I'll be honest that it does nothing to make old Shorts earn more, because nothing does. Worth saying, given I'm the one selling something in this conversation. If you want the workflow side, drip scheduling Shorts is the more useful read.

Frequently asked questions

How long do YouTube Shorts keep getting views?

On two channels I measured, groups two to three months old sat only 12 to 22 percent above groups eight days old. A large platform-wide study reports a bigger late tail, but it uses averages that a few viral outliers can distort. Expect a modest tail, not a multiple.

Do YouTube Shorts stop getting views after 30 days?

There's no cliff at 30 days. My 64 and 78 day old groups are not meaningfully behind or ahead of newer ones. Claims about a specific 28 to 30 day drop trace back to one analyst's stated working theory, which YouTube has never confirmed.

Should I delete old YouTube Shorts that didn't perform?

No. They cost nothing to leave up and occasionally one finds an audience through search. Just don't build a strategy around the back catalogue earning meaningfully more over time, because on my numbers it doesn't.

Is it better to measure Shorts performance with the median or the average?

The median, for your own channel. Averages get pulled upward by rare viral outliers, so a platform-wide average can describe a curve that almost no individual creator experiences. The median tells you what your typical Short does.

How do I check whether my own Shorts compound?

Export your Shorts with their publish dates and view counts, group them by publish week, and take the median views for each week. Then read the column oldest to newest. If it doesn't climb with age, your Shorts aren't compounding either.