How long should a Reel or TikTok be? Length studies count views, not who stays to the end

KEY TAKEAWAYS

1. Three minutes is the one hard line on Instagram. Below it, length differences in engagement are small.
2. Longer TikToks earn more views per post, but the median viewer watches seconds, not minutes.
3. Ad placements and ad studies point shorter than organic data does.
4. Decide where the point sits before deciding how long the video runs. Then test cut-downs of the same footage.

Instagram refreshed its guide to making Reels in late September. It sorts length by content type: 3 to 15 seconds for trends and one-liners, 15 to 60 seconds for tutorials and most formats, and 60 seconds to 3 minutes for longer walkthroughs. Social Media Today, reporting it on 28 September 2026, noted the guide adds nothing new. That is a fair reading. It is also a useful prompt, because the length data published this year seems to say something different, and paid evidence says something different again.

Why do the length studies disagree?

Because they measure different things, and none of them measures the thing you control.

Most length studies report one of three numbers. Views per post. Engagement per post, divided by followers or by views. Or, less often, median watch time. Each answers a separate question. A video can win on views and lose on watch time in the same dataset.

There is also a fourth question the studies cannot answer: what happens when the same footage is cut to a different length. Every published dataset compares different videos from different accounts. Long videos are made by different people, about different things, than short ones.

So the useful move is to split the length decision into separate checks and see what the evidence supports for each.

What is the hard limit on Instagram?

Three minutes, for reach.

Instagram's creators blog said on 21 January 2025 that Reels up to 3 minutes can now reach non-followers in the Reels tab and Explore. Before that, only Reels up to 90 seconds were eligible for recommendation. Social Media Today noted at the time that this reversed Instagram's May 2024 advice that longer Reels see less engagement.

What happens above 3 minutes is less clearly documented. Socialinsider states that longer Reels are not placed in the Reels tab or recommended to new viewers. We could not open an Instagram help page that says so directly. The data is consistent with it, though.

Socialinsider's analysis of 6 million brand Reels from January to June 2026 gives these median views by length:

Two readings matter. Above 3 minutes, median views fall to 43% of the 45 to 60 second level (our calculation). Below 3 minutes, the spread is modest. Engagement ranges from 0.28% to 0.35% across every bucket under the line.

One caution on the source. Socialinsider's August 2026 benchmark page shows the same Instagram figures, labelled as averages and attributed to 111,000 videos. Its July blog post labels them medians from 6 million Reels. Treat the table as directional, not precise.

Do longer TikToks really get more views?

In the published data, yes, by a wide margin.

Socialinsider's TikTok analysis covers more than 6 million brand videos from the same six months. Median views run 2,200 for 30 to 60 seconds, 7,200 for 60 to 90 seconds and 11,136 for 120 to 180 seconds. The jump across the one-minute mark is about 3.3 times (our calculation). Buffer's earlier study of 1.1 million TikToks found the same direction: median reach of 302 for 30 to 60 second videos and 432.5 for videos of 1 to 10 minutes.

Two things qualify this.

First, the incentive. Since March 2024, TikTok's Creator Rewards Program has paid only for videos over a minute long. Its formula rewards play duration, which covers watch time and finish rate. A population of long TikToks therefore includes many creators who are paid to hold attention and are practised at it. That is a selection effect, not proof that adding length adds views.

Second, the watch time. Buffer reports median watch time of 11.3 seconds for videos of 1 to 10 minutes, against 6.9 seconds for 30 to 60 seconds. Longer videos hold people somewhat longer. But 11.3 seconds is under a fifth of even a one-minute video (our calculation). The median viewer of a long TikTok never sees most of it. Buffer's own data scientist cautioned that engagement varies wildly and depends on virality.

So longer TikToks collect more views per post. Whether your next video should be longer depends on whether you have enough to say past the point where most people leave.

What does the paid evidence say?

It points shorter, for a different reason.

Organic data counts views per post. In paid media you buy each impression, so what matters is the response per impression. On that question the best evidence we found favours short cuts.

A peer-reviewed study by Fossen, Kim and Chae in the Journal of Marketing (online June 2025) used TV advertising, web traffic and sales data. It found very short ads drove more immediate web traffic than longer ads, with similar direct effects on online sales. A field experiment with social media ads pointed the same way on traffic and engagement. A Science Says summary puts the short category at 10 seconds or less.

Not everyone agrees on sales. MNTN, a vendor analysing more than 1,800 TV commercials, reports 15-second ads did 46% better than 30-second ads on visit rate, while 30-second ads drove higher conversion. Both sources agree that short wins on traffic. They part on conversion, and both are based on TV.

Platform guidance is thin. TikTok's creative tips document recommends 21 to 34 seconds for In-Feed ads, based on research from May to July 2021, and the same document's quality checklist says 9 to 15 seconds. Its current help page allows non-Spark In-Feed videos up to 10 minutes and gives no recommendation. On Meta, ads on Facebook Reels and Facebook in-stream video cap at 15 seconds, so a single video meant to run in every placement has to be 5 to 15 seconds, as Jon Loomer's reading of Meta's creative guide shows.

inBeat, writing on 9 October 2026, found no published paid UGC study comparing length buckets across Meta, TikTok and YouTube. We found none either.

Where should the point sit?

Earlier than you think. This matters more than total length.

A hypothetical example. A brand posts a 75-second tutorial that gets 10,000 views. Assume a retention curve of this shape, which is illustrative and not from any dataset:

If the product step sits at second 60, about 1,100 people see it. Move it to second 10 and about 3,500 do, from the same views. The extra length can stay. It now serves the viewers who chose to stay, rather than holding the point hostage.

The shape of this curve is consistent with Buffer's median watch times, though your own will differ. Read your own retention or average watch time before choosing a length. The first seconds are a separate problem, covered in the 3-second hook rule.

How should you set length, by job?

Whether a video or a carousel suits the job at all is a prior question. We covered it in carousel or Reel.

A seven-step cut-down test

1. Pick one piece of footage with a clear point and enough material for a long version.
2. Edit three cuts: a short one under 15 seconds, a middle one of 30 to 60 seconds, and a long one under 3 minutes for Instagram or over a minute for TikTok.
3. Put the point in the first 10 seconds of every cut. Only the material after it should differ.
4. For organic posts, post the cuts as separate videos over several weeks and record views, average watch time and the share reaching the point. Accept that organic results will be noisy.
5. For ads, run the cuts as variants in one ad set with the same hook, offer and landing page. Judge on cost per result, not completion rate, which favours the shortest cut by design.
6. Size the test honestly. Three cuts are three variants competing for the same budget. See how many ad concepts your budget can test.
7. Keep the winning length per job, not per account. A tutorial result does not transfer to a trend post.

What the evidence supports

Three minutes is the one firm line on Instagram. TikTok rewards longer videos with more views, in data that cannot separate length from who makes long videos. Paid evidence favours short cuts for traffic and is divided on sales. None of it tells you the right length for your next video. Where the point sits does more work than the running time, and a cut-down test of the same footage is the only clean comparison available.

The same discipline applies to how a brand shows up in search and AI answers. If you want to see where yours stands, the waitlist is open.

Questions

What is the best length for an Instagram Reel?

There is no single best length. Instagram's 2026 guide suggests 15 to 60 seconds for most formats and up to 3 minutes for walkthroughs. Socialinsider's 2026 data on 6 million brand Reels shows median views peaking at 45 to 60 seconds and falling sharply above 3 minutes.

Do longer TikToks get more views?

In published data, yes. Socialinsider and Buffer both find median views rise with length up to two or three minutes. Neither controls for who makes long videos, and TikTok's Creator Rewards Program only pays for videos over a minute. Buffer's median watch time on 1 to 10 minute videos was 11.3 seconds.

How long should a TikTok or Reels ad be?

Shorter than organic posts, as a starting point. Ads on Facebook Reels and in-stream cap at 15 seconds. A Journal of Marketing study found very short ads drove more immediate web traffic, with similar direct effects on sales. TikTok allows up to 10 minutes for non-Spark In-Feed ads but recommends no length.

Are Reels longer than 3 minutes recommended to non-followers?

Instagram's own announcement extended recommendation eligibility from 90 seconds to 3 minutes in January 2025. Socialinsider says Reels above 3 minutes are not recommended to new viewers. Its data shows median views for 3 minute plus Reels at less than half the 45 to 60 second level.

Sources

1. Social Media Today, Instagram shares Reels creation tips in new guide socialmediatoday.com
2. Instagram Creators, Finding success on reels in 2025 creators.instagram.com
3. Social Media Today, Instagram will recommend longer Reels socialmediatoday.com
4. Social Media Today, Instagram officially expands Reels length to 3 minutes socialmediatoday.com
5. Socialinsider, What is the ideal Instagram Reels length socialinsider.io
6. Socialinsider, How long should TikTok videos be socialinsider.io
7. Socialinsider, 2026 Social Media Video Performance Statistics socialinsider.io
8. Buffer, Longer TikToks get more views buffer.com
9. TikTok Newsroom, Introducing the new Creator Rewards Program newsroom.tiktok.com
10. TikTok for Business, 7 Top Creative Tips ads.tiktok.com
11. TikTok Business Help Center, TikTok auction In-Feed ads ads.tiktok.com
12. Jon Loomer, Meta video ad length requirements jonloomer.com
13. Journal of Marketing, The Impact of Ad Length on Ad Effectiveness: Do Micro Ads Work? journals.sagepub.com
14. Science Says, Keep video ads under 10 seconds app.sciencesays.com
15. MNTN Research, Shorter TV ads drive more site visits, but longer ads fuel conversions research.mountain.com
16. inBeat, UGC video length benchmark 2026 inbeat.agency

NEXT STEP

See where your brand stands.

Your Digital Authority Score, who AI recommends for your target questions, and the placements that close each gap. Your first report is free.

Get your free report →