Short-form video statistics: what people actually watch
Ofcom's panel measured 51 minutes a day of video-sharing platforms. Metricool measured the average Reel getting 8.5 seconds. Here is what each study counted, and which famous numbers count nothing.
8 minute read
The rule this page follows
Search for short-form video statistics and you get pages of numbers with no owner: no publisher, no sample, no date. So here is the rule this page follows. Every figure below names who measured it, how much they measured and when, and anything that failed that test was cut. That knocked out most of the usual list, including the line about human attention spans falling to eight seconds, which gets attached to short-form video constantly and whose trail runs back to a 2015 marketing report and stops there. What survives splits into two kinds: panel measurement of what people watched, and platform data on what got posted and how far it travelled. Those answer different questions, and quietly mixing them is where most of these pages go wrong.
How much of the day short-form actually takes
Ofcom's Media Nations 2026 report, published on 29 July 2026, is the closest thing to audited measurement in this space. It draws on Barb as-viewed panel data covering devices on the home's wifi network, so it is UK-only and in-home only. In 2025, individuals aged four and over watched 4 hours 23 minutes of video a day at home across all devices. Video-sharing platforms took 51 minutes of that, up from 43 minutes in 2022. YouTube alone accounted for 80% of platform viewing at 41 minutes a day, which is 15% of all in-home video. TikTok came in at 4%. That last number understates badly, because viewing over mobile networks is not captured at all, and TikTok is a mobile-network habit.
Short-form moved to the television
The more interesting movement in Ofcom's data is which screen. Average daily YouTube viewing on a TV set went from 9 minutes per person in 2022 to 19 minutes in 2025, a rise of 111%, and among 16 to 34-year-olds it went from 5 minutes to 21, up 320%. By 2025, 46% of in-home YouTube viewing happened on a television, against 27% in 2022. Ofcom flags something else worth pausing on: Barb counts YouTube reach at three or more consecutive minutes rather than the fifteen used for broadcast, precisely because most of what people watch there is short form. YouTube's Neal Mohan said at Cannes Lions in June 2025 that viewers watch over a billion hours of YouTube on televisions every day.
The biggest number in short-form, and why it moved
The headline figure is Shorts. Mohan said in that same June 2025 keynote that Shorts averages more than 200 billion daily views, against the 70 billion YouTube reported in March 2024. Treat the jump carefully, because the definition changed in between. On 31 March 2025 YouTube changed what counts as a view on Shorts: a view now registers the moment the Short starts playing or replays, whether or not anybody watched it. The old measure was not deleted, it was renamed engaged views, and that is still what YouTube uses for Partner Programme eligibility and Shorts revenue sharing. So the 200 billion counts scrolls and the 70 billion did not. In your own Studio analytics, engaged views is the number that maps to what you were tracking a year ago.
How long anyone actually watches
For watch time, Metricool's 2026 Instagram Study is the largest recent look: 24,364,803 posts from 375,118 accounts, comparing early 2026 against the same window a year earlier. Average watch time per Reel came out at 8.5 seconds, up 118% year on year, with total view time up 139%. Reels drew more than four times the interactions of single-image posts. Two caveats matter. That 8.5 seconds is an average across every Reel served, including the vast number flicked past in under a second, so it does not mean people watch 8.5 seconds of your video — it means the typical Reel gets under nine seconds of anyone's day. And in the same dataset, Instagram Reels pulled roughly 30% fewer views than TikTok videos.
Clip length: the data disagrees with the advice
On length, the most useful public dataset is Buffer's analysis of 1.1 million TikTok videos, published in March 2025. It confirms the obvious first: 86% of TikToks run under 60 seconds, with the 10 to 30-second bucket most common at 33.7% and 30 to 60 seconds next at 27.3%. Then it contradicts the standard advice. Median reach rose steadily with length, from 194 views for 5 to 10-second clips to 432.5 for videos over a minute, and median watch time went from 3.1 seconds to 11.3. Videos past 60 seconds took 43.2% more reach and 63.8% more watch time than 30 to 60-second ones. That is correlation across one tool's users, not proof that padding helps, but it is a decent argument against treating 60 seconds as a ceiling.
The numbers that did not make it
It is worth naming what got cut, because those are the figures you will meet everywhere else. There is no credible published completion-rate benchmark for Reels, Shorts or TikTok. No platform releases that data, so every claim that the average completion rate is some specific percentage comes either from one tool vendor's customers or from nowhere at all. The daily minute counts quoted for TikTok swing between roughly 50 and 95 depending which page you land on, with no consistent measurement behind them. The eight-second attention span belongs in the same bin. If a short-form statistic does not tell you who counted, how many, and over what period, assume it was copied from a page that did not know either.
What to do with the numbers that survive
What is left is directional, and it is enough. A typical short clip gets under nine seconds of attention, most of it decided in the first two. Short-form viewing is still growing but slowly, and increasingly on a television rather than a phone. Longer clips are not punished the way the folklore says, so 60 seconds is not a rule. Measure what you can actually see: engaged views rather than views on Shorts, retention curves rather than view counts, and your own median rather than an industry average. For most people the real constraint is not picking the right length, it is producing enough clips to find out, which is the part FrameOS takes off your hands. If you want to test that against your own footage: 300 credits for 3 days · no card.
FAQ
How much short-form video do people watch per day?
The best measured figure comes from Ofcom's Media Nations 2026 report, using Barb panel data: UK individuals watched 51 minutes of video-sharing platforms per day in the home during 2025, up from 43 minutes in 2022. YouTube was 41 minutes of that. It excludes viewing over mobile networks, so real totals are higher, particularly for TikTok.
What is the average watch time on Instagram Reels?
Metricool's 2026 Instagram Study, covering 24,364,803 posts from 375,118 accounts, put average watch time per Reel at 8.5 seconds, up 118% year on year. That average includes every Reel served, including the ones scrolled past instantly, so read it as what a typical Reel earns rather than how long an engaged viewer stays.
How long should a short-form video be in 2026?
Buffer's analysis of 1.1 million TikTok videos found median reach rising with length: 194 views for 5 to 10-second clips against 432.5 for videos over a minute, with videos past 60 seconds getting 43.2% more reach than 30 to 60-second ones. Length is not the lever people think it is. Cut to where the idea ends.
Why did YouTube Shorts views suddenly jump?
On 31 March 2025 YouTube changed the definition. A view now counts the moment a Short starts playing or replays, regardless of whether anyone watched. The previous measure was renamed engaged views and still governs Partner Programme eligibility and revenue sharing. Compare engaged views, not views, if you want continuity with older numbers.
Sources
- Media Nations 2026 · Ofcom
- Neal Mohan at Cannes Lions 2025: What 20 years of YouTube reveals about creativity's future · YouTube
- A Change to How We Count Views on Shorts · YouTube Help
- Metricool's 2026 Instagram Study Reveals Reels and Carousels Are Reshaping Platform Strategy · Metricool
- Longer TikToks Get More Views — Data Shows Best TikTok Video Length · Buffer
Related pages
Keep reading
More Creator Growth →The best time to post on Instagram, according to four studies
Buffer says Thursday 9am. Later says 5am. Sprout says early afternoon. They analysed billions of posts and still disagree — here is why, and what to do about it.
8 minute readThe best time to post on TikTok, according to five studies
Buffer analysed 7.1 million posts and rates Saturday the best day. Sprout Social analysed 2 billion engagements and says avoid the weekend entirely. Both are right, for different accounts.
8 minute readThe best time to post on LinkedIn, according to five studies
Buffer's 4.8 million posts point to Wednesday at 4pm and rank Tuesday among the worst days. Sprout Social's 2 billion engagements call Tuesday the best day of the week. Here is why.
8 minute readTurn one long video into a week of posts
300 credits for 3 days · no card.
Start free