How long should a YouTube video be? 20 minutes or more, in 3.5M videos

Make long-form YouTube videos 20 minutes or longer. Against the same channel's other videos, a 20 to 30 minute video gets 1.04x the channel's median views, and a 30 to 60 minute video gets 1.06x. A 10 to 15 minute video, the length most guides recommend, gets 0.94x, and a video under 4 minutes gets 0.81x or less. Past one hour the gain grows little, to 1.07x and 1.08x.

We compared each video with its own channel's other videos, so a few big channels do not decide the answer. The sample is 3,470,826 public long-form videos from 140,104 channels with 1,000 or more subscribers, posted from October 1, 2025 to October 21, 2026 and read by Virev's crawler from August 6 to October 24, 2026.

Most pages that rank for this question give a range from experience. Brafton says 6 to 8 minutes, Teleprompter.com and CapCut say 7 to 15 minutes, and Jason Scott Montoya found 30 to 40 minutes on his own channel. None of them compares lengths inside the same channel across a large sample. This page adds that comparison, the cut by subscriber count and category, a check on Shorts, and the lengths channels post.

How long should a YouTube video be for views

For each video we adjusted its views to a common age, then divided them by the median of the same channel's other videos. A video at 1.00x got a normal number of views for that channel. We then took the median of that ratio for each length band.

This table splits the result by subscriber count. "Share" is the share of all videos in the sample at that length. "All" and the five tier columns are the median views against the channel's own median. "Beats median" is the share of videos in the band that got more views than their channel's median.

• In the All column, every band under 15 minutes sits below the channel's median, and every band from 15 minutes up sits above it.
• Videos of 10 to 15 minutes get 0.94x in every tier from 10,000 subscribers up, and 0.96x for channels with 1,000 to 10,000.
• Short videos cost small channels more, because a 2 to 4 minute video gets 0.73x for channels with 1,000 to 10,000 subscribers and 0.94x for channels with over 1 million.
• The gain flattens after 30 minutes, from 1.06x at 30 to 60 minutes to 1.08x past 2 hours, except on channels with over 1 million subscribers, where videos past 2 hours reach 1.16x.
• 53.9% of the videos in the sample run 20 minutes or longer, so the lengths that beat the median are also the lengths most channels post.

YouTube lets monetized channels put ads in the middle of videos that are 8 minutes or longer. The views do not jump at that line, with 0.89x at 6 to 8 minutes and 0.91x at 8 to 10 minutes. Ad income per video is outside this study, because public pages do not show it.

Same channel, 10 to 15 minutes against other lengths

The table above compares each video with the channel's overall median. A stricter test takes channels that posted at least two videos in a band and at least two of 10 to 15 minutes, and compares the two medians for each channel. "Win" is the share of those channels whose median in the band beat their 10 to 15 minute median, and "ratio" is the typical channel's band median divided by its 10 to 15 minute median.

• 68.6% of the 35,171 channels that posted both got more views on their 20 to 30 minute videos than on their 10 to 15 minute videos, and the typical gain was 38%.
• For the 23,265 channels with both, a 30 to 60 minute video got a median 1.53x the views of the same channel's 10 to 15 minute videos.
• Only 29.5% of channels did better with 2 to 4 minute videos than with 10 to 15 minute ones, and the typical channel got 0.55x.
• Past 30 minutes the share of channels that win stays near 68%, so the extra length helps about as many channels as before, by more.

By category

The cut by the channel's main YouTube category shows the same pattern in all 8 of the largest categories. The columns are the median views against the channel's own median for six of the twelve length bands, and the CSV has all twelve. The last column is the number of channels.

• Music is the one category where 4 to 10 minute videos come close to the channel's median, at 0.98x to 1.00x, and 10 to 15 minute music videos dip to 0.95x.
• Education, News & Politics and Sports gain the most from videos past 2 hours, with 1.17x to 1.19x past 2 hours.
• Gaming flattens soonest, at 1.04x for 20 to 30 minutes and 1.03x past 2 hours.

Why other studies give a shorter answer

If you pool every video and read the median views at each length, the best length looks like 15 to 20 minutes. The pooled columns below use all videos 7 to 89 days old at our read, without the per-channel step.

• Pooled together, videos of 15 to 20 minutes get the most views, 19,434 at the median, and videos of 1 to 2 hours get 8,134.
• Against each channel's own median, the same 1 to 2 hour videos get 1.05x, above the 15 to 20 minute videos at 1.01x.
• The pooled drop for long videos comes from who posts them, since channels with 1,000 to 10,000 subscribers post a median video of 25.1 minutes and channels with over 1 million post 17.6 minutes.
• Divided by subscribers, pooled views stay near 30 per 100 subscribers for every band from 20 minutes up.

A study that compares lengths across channels mostly measures which channels post which length. To pick a length for your own channel, compare your own videos with each other.

Three channels where length moved views

We picked these from channels with at least 12 videos under 8 minutes and 12 videos of 20 minutes or more in the sample. We checked each video's page on YouTube on October 24, 2026.

SpongeBob's official channel posts compilations that run a median of 61 minutes, such as "2 Hours of SpongeBob's Best Moments from Season 12", which had 4.1 million views at our read and 4.8 million on October 24. Its 94 videos of 20 minutes or more got a median 523,805 views, 3.6 times its 22 videos under 8 minutes.

The 85 South Comedy Show posts full episodes of its comedy podcast, which run a median of 113 minutes, next to short clips. Its full episodes got a median 86,342 views, 4.8 times its 33 videos under 8 minutes. Its most-viewed video in the sample is a 79-minute episode with 1.2 million views at our read.

The ONE PIECE channel goes the other way. Its 65 videos under 8 minutes are mostly 30-second trailers and opening-theme clips, with a median of 147,559 views, and its 34 videos of 20 minutes or more got a median of 69,542, about half as many. One 30-second opening-theme clip had 17.9 million views at our read.

All three are official channels of a show, and their long and short videos are different formats, a full episode next to a clip. The same holds for many channels in the tables above, so part of the length gain may come from what goes into a long video.

YouTube Shorts length

Our crawler mostly reads the Videos tab of each channel, so it holds few Shorts. We had 7,560 Shorts with a length and a view count from 334 channels with 5 or more Shorts, and we ran the same comparison on them. "95% range" is the range the median moved in 1,000 resamples of the channels.

• Every Shorts length lands within 2% of the channel's median Short, and every 95% range includes 1.00x.
• In our sample, length does not change how a Short does against the channel's other Shorts, so pick the length the idea needs.
• The median Short in our sample runs 27 seconds, and 15.4% run past 60 seconds.

YouTube has allowed Shorts of up to 3 minutes for videos uploaded on or after October 15, 2024. Piktochart cites an Inflow Network study of 5,400 Shorts in which 50 to 60 second Shorts got about 22 times the views of Shorts under 10 seconds. That study pools Shorts across channels, which is the comparison the pooled table above warns about. With 334 channels, our Shorts sample is small, and a larger one may show a gap our sample cannot.

How this compares with our best-time study

Our best time to post on YouTube study measured length with a different scale on 173,516 videos from 6,212 channels. It reports the mean of each video's log views against the channel's average, as a percent. We ran that method on this sample with its length bands.

• Both samples give the same order, with videos under 8 minutes below the channel's average and videos past 15 minutes above it.
• From 4 to 60 minutes, the two studies differ by 2.4 points or less.
• The largest gaps are at 1 to 4 minutes, -22.5% against -27.7%, and at 60 minutes or more, +17.4% against +11.7%.

The best-time study used channels with 10 or more videos and a known time zone, and videos posted from June 2025 to September 2026. This study uses channels with 5 or more videos, videos posted from October 2025 to October 2026, and 20 times as many videos. The main measure on this page is a median ratio, which moves less than a mean of logs, so 1.06x here and +10.0% in the table describe the same 30 to 60 minute videos.

Our YouTube engagement rate benchmarks found that channels whose videos run 40 minutes or more get fewer likes and comments per view than channels at 10 to 40 minutes. Longer videos collect more views, and viewers of the longest videos like and comment on a smaller share of them.

How long can a YouTube video be

These are the limits YouTube states in its Help Center, read on October 24, 2026.

• The YouTube Help page on longer uploads says you can upload videos up to 15 minutes long by default, and that verified accounts can upload longer videos.
• The mid-roll help page says that on monetized videos of 8 minutes or longer, you can turn on ads during the video.
• The three-minute Shorts page says square or vertical videos of up to 3 minutes uploaded on or after October 15, 2024 count as Shorts.

What to do with this

For creators

• Plan long-form videos of 20 minutes or more when the material holds up, because 20 to 30 minute videos beat the same channel's 10 to 15 minute videos for 68.6% of channels.
• Do not cut a video down to 7 to 15 minutes to follow a rule of thumb, because 10 to 15 minute videos got 0.94x the channel's median in every tier.
• Post material under 4 minutes as a vertical Short or as an extra, because long-form videos that short got 0.81x the channel's median or less.
• To check your own channel, take your videos that are at least a week old, find your median, and see which length bands sit above it.
• For when to upload, see the best time to post on YouTube, where the weekend beat midweek by 7.0%, a smaller gap than length shows.

For brands

• If you sponsor a segment, prefer a creator's videos of 20 minutes or more, because those got 1.04x to 1.08x the channel's median views.
• If you ask a creator for a short dedicated video, expect fewer views than their median, since 2 to 4 minute videos got 0.81x.
• Compare a creator's views with their own median before you book, because SpongeBob's long videos got 3.6 times the views of its short ones.
• Virev creator search covers 180M+ creators on Instagram, TikTok and YouTube and shows each one's recent videos with their views, so you can see which lengths work for a channel before you book.

Start a 3-day trial to run your first search.

How we measured, and what can be wrong

• Channels are public YouTube channels with 1,000 or more subscribers that Virev's crawler read, minus 411 channels that Virev's profile label marks as a brand, business, media company, organization or government. Most channels have no label, so some company channels, such as TV networks and studios, remain.
• Videos are long-form uploads posted from October 1, 2025 to October 21, 2026, not marked as Shorts or live streams, with a length and at least 1 view at a reading taken 3 or more days after upload. Each channel needed 5 or more such videos.
• Views come from the latest crawler reading of each video, and the median video was 41.8 days old at that reading.
• Views keep growing after upload, so we estimated how views change with age inside each channel, in 14 age bands, and adjusted every reading to the same age.
• Each video is compared with the median of the same channel's other videos, leaving the video itself out.
• Subscriber tiers use the count at our read, with 41,603 channels at 1,000 to 10,000, 50,326 at 10,000 to 100,000, 29,656 at 100,000 to 500,000, 7,873 at 500,000 to 1 million and 10,646 over 1 million.
• The channel's category is the YouTube category most of its videos carry.
• The checks below all give the same order, and the CSV has each one by band.

• "No age step" uses raw views. "Same month" compares each video only with the same channel's videos from the same month, for months with 5 or more videos. "14-59 days old" keeps only readings taken 14 to 59 days after upload. "Median channel" takes each channel's median ratio in the band, then the median channel, so channels with many videos do not count more.
• Channels choose a length for each idea, and a long video is often a full episode while a short one is a clip or trailer, as the three channels above show, so part of the gain for long videos may come from the content.
• The sample holds only videos that were still public when we read them, so a channel that deleted weak videos of one length would make that length look better than it was.
• Some long uploads may be past live streams that our data does not mark as live.
• Our crawler collects channels for creator marketing, so the sample holds more active channels than YouTube as a whole.
• The Shorts sample comes from the 334 channels where our crawler read Shorts with a length and a view count, so it may not stand for Shorts on YouTube as a whole.

Download the data

The tables behind every chart are in one CSV: how-long-should-a-youtube-video-be-2026.csv. Each row is one length band in one table, with a format (long-form or shorts) and a group (all channels, a subscriber tier or a category). The tables are length_vs_own_median (videos, channels, share of videos, median views against the channel's own median, share beating the channel's median), length_vs_own_median_by_category, paired_vs_10_15_min, pooled_median_views, checks, checks_channel_weighted, best_time_method, length_posted and shorts_vs_long_form. Please cite "Virev, October 2026" and link this page.