I chased the wrong number for a year
Let me start with the embarrassing part. For most of last year I judged every Short I posted by the view count on the screen. A Short hit 40,000 views and I celebrated. Another one stalled at 800 and I deleted it in frustration. Then I learned what YouTube actually counts as a view now, and I realised I had been reading the scoreboard upside down.
Here is the change that matters most. Since March 31, 2025, a Shorts view counts every time a Short starts to play or replays — no minimum watch time at all. Then, from August 24, 2026, YouTube extended that first-frame counting to every format, long videos and live streams included. Your public view count went up overnight, but nothing about your audience actually changed. The honest metric survived under a new name: engaged views. That is the count of people who chose to keep watching after the first moment, and it is the number YouTube still uses for Partner Program eligibility and Shorts ad revenue. If you only watch one number, watch that one.
The second change is even bigger for how I plan content now. On October 1, 2026, YouTube updated its systems to further prioritise original content and reduce the reach of videos re-uploaded from other creators without adding anything of your own. Clipping someone else's video and posting it as your Short used to be a growth hack. In 2026 it is a dead end. Original footage, or a clip with your own commentary and something added, is now the entry ticket to the feed.
And the scale is absurd. YouTube CEO Neal Mohan's January 2026 letter put average daily Shorts views at 200 billion. Two hundred billion, every single day.
How the Shorts feed actually decides
I used to imagine the algorithm as a judge handing out scores. It is not. It is a testing machine. When you publish a Short, YouTube shows it to a small seed audience of people who might enjoy it, then watches what they do. If they watch instead of swiping away, the Short goes wider. If they swipe, it stops. That loop is the whole game, and YouTube's own product lead for Shorts, Todd Sherman, has described it exactly that way.
The signals YouTube names for Shorts are refreshingly simple: whether people shown the Short choose to view it, ignore it, or tap "not interested"; the share of viewers who chose to view; average view duration; average percentage viewed; likes; and post-watch survey results. Your personal history matters too — the Shorts and channels a viewer enjoyed before, the topics they watch, the trending sounds they engage with.
Notice what is not in that list. Hashtags do not rank you. Follower count does not rank you. Posting time is not a ranking factor — Sherman himself has said it is not something to optimise for, except for news. There is no minimum posting cadence required for Shorts to perform well, and YouTube has confirmed that Shorts performance does not hurt your long-form recommendations, because Shorts and long-form now run on separate recommendation systems. All the folk wisdom I believed for years — post at 6pm, use 15 hashtags, never let a Short flop or your channel is finished — was wrong.
The metric I now check first in YouTube Studio is "Viewed vs. swiped away", the card labelled "How many chose to view". If most people shown my Short swipe past it, the hook failed. No amount of posting frequency fixes a hook problem. The commonly cited benchmarks I work to are keeping the first-two-second swipe-away rate under 40 percent and aiming for retention above 70 percent. Those are creator-community benchmarks, not YouTube's published thresholds — YouTube itself publishes no fixed numbers — but in my experience they are a useful bar.
The hook is the whole video
In long-form video, the thumbnail and title do the selling. In Shorts, nobody chooses from a thumbnail. The viewer is already swiping, and your first two seconds are the thumbnail, the title and the trailer all at once. I learned this the expensive way: my early Shorts opened with a logo animation, then my name, then a slow "hey guys, welcome back". Every single one of them died. The fix was embarrassingly simple. Start with the payoff.
Now I open with the result, the claim, or the question the video answers — in the first frame, with the words on screen and the thing itself visible. If the Short is a recipe, the finished dish is on screen at second zero. If it is a tutorial, the before-and-after is the first thing you see. If it is a story, the most surprising line comes first and the setup comes after. A viewer who has already decided to stay will happily watch the setup. A viewer who has not decided yet will never get there.
Looping helps too. A Short that ends where it begins gets watched twice, and a loop rate above 100 percent is one of the strongest secondary signals the system reads. I now write endings that flow back into the opening line, so a replay feels natural instead of jarring. It is a small craft trick, and it shows up directly in the retention graph.
One more thing about length. Shorts can now run up to three minutes — the cap moved from 60 seconds on October 15, 2024 — but a three-minute Short is judged by the same percentage-viewed math as a 30-second one, and a three-minute Short abandoned halfway reports 50 percent viewed. I keep most of mine between 25 and 45 seconds, and I only go longer when the idea genuinely needs the room.
What I post, and how often
Here is my honest weekly routine now, and it is much simpler than the posting schedules the growth gurus sell. I post three to five Shorts a week, all in one topic, all original. One topic matters more than volume: the feed learns who likes your stuff, and a channel that jumps between cooking, crypto and comedy confuses the very system it is trying to please. Niche channels grow by posting original Shorts in one topic and learning from the ones viewers chose to watch.
I also use Shorts as a testing lab for long-form. A Short that earns strong engaged views and subscribers is a proven idea; I expand the winners into full videos. A Short that flops is a cheap lesson instead of a wasted week of editing.
The money question, answered honestly
Growth is only half the story, so let me be straight about what Shorts pay, because the gap between the view count and the pay cheque shocks most new creators. Shorts ads run between Shorts in the feed, the revenue goes into a Creator Pool, and you keep 45 percent of your allocation. Creator-reported Shorts RPMs in 2026 sit roughly between one and seven cents per 1,000 views — some niche studies put the range wider, up to about $1.48 per 1,000 in the best niches — but the honest typical figure is cents, not dollars. A million Shorts views might pay tens of dollars, not thousands.
That is why I treat Shorts as a discovery engine, not an income engine. The money comes from what Shorts feed: subscribers who watch long-form, where creators keep 55 percent of ad revenue and typical RPMs run from about $1 to $30 per 1,000 views depending on niche. Shorts bring the audience; long-form, memberships, and brand deals pay the bills. Anyone selling you a "get rich from Shorts views" course is selling the wrong number. On eligibility: the Partner Program ad tier needs 1,000 subscribers plus 4,000 watch hours in 12 months or 10 million Shorts views in 90 days, and from February 1, 2027 YouTube will require 10 million qualified Shorts views per rolling 90 days to keep earning Shorts ad revenue — qualified views, not the inflated public number.
The mistakes I stopped making
Let me list the ones that cost me the most time, so you can skip them. First, re-uploading other people's clips with no added value — dead since October 2026, and it was never a business, just borrowed reach. Second, judging a Short after 48 hours. I now judge after 28 days in Studio, on impressions, CTR equivalents and retention, because the feed can pick a Short up late. Third, buying views or engagement. YouTube's fake engagement policy bans artificially inflated views, likes, comments and subscribers, and it is not worth the channel. Fourth, stuffing the description with hashtags and keywords as if it were 2019. The description barely matters for Shorts discovery; the first two seconds matter more than the first two paragraphs.
Fifth, and this one hurt: deleting Shorts that "flopped". A Short with 800 views in its first week can still be the one the feed tests again later, and a deleted video teaches you nothing. I keep everything up now and read the retention graph instead. The graph tells me exactly where viewers left, which is a better teacher than any course.
Conclusion
Shorts growth in 2026 is simpler than the industry makes it look, and harder than the gurus make it sound. The feed tests every Short on a small audience and expands the ones people choose to watch. Original content wins now, by design. Engaged views are the number that matters, not the public count that changed in March 2025. The hook is the first two seconds, the length sweet spot is roughly half a minute, and three to five original Shorts a week in one niche beats ten random ones every time.
My advice, after all the wasted experiments, is this: make something only you could make, open with the payoff, watch the "viewed vs. swiped away" card like it is your bank statement, and use Shorts to find the audience your long videos will keep. The feed is not a lottery. It is a mirror. Make something people choose to finish, and it will show it to more people who will do the same.
FAQ
How does the YouTube Shorts algorithm work in 2026?
YouTube shows each new Short to a small seed audience, then widens or narrows its reach based on whether those viewers watch, swipe away, or tap "not interested". It weighs the share who chose to view, average view duration and percentage viewed, likes, and survey responses. Shorts and long-form run on separate recommendation systems.
What is the difference between views and engaged views on Shorts?
Since March 31, 2025, a Shorts view counts every time playback starts or replays, with no minimum watch time. Engaged views are the older, stricter metric — viewers who chose to keep watching — and it is the number used for Partner Program eligibility and Shorts ad revenue.
How long should a YouTube Short be in 2026?
Shorts can run up to 3 minutes since October 2024, but the commonly cited sweet spot is roughly 25 to 40 seconds. Longer Shorts are judged on percentage viewed, so only go long when the idea needs the room.
How many Shorts should I post per week?
There is no minimum cadence required by YouTube. Most creators I follow land on 3 to 5 original Shorts per week in one niche, which is enough to learn from the data without burning out.
Do re-uploaded clips still work on Shorts?
Not anymore. Since October 1, 2026, YouTube prioritises original content and reduces the reach of content re-uploaded from other creators without adding anything of your own. Add your own commentary or footage.
How much do YouTube Shorts pay?
Creators keep 45 percent of their allocation from the Shorts Creator Pool. Creator-reported Shorts RPMs are roughly 1 to 7 cents per 1,000 views in most niches, far less than long-form. Treat Shorts as discovery, not income.
