Keep a Reel under three minutes if you want it recommended to people who do not follow you. Inside that ceiling, 15 to 30 seconds wins on completion rate, and 45 seconds to three minutes wins on accumulated watch time. The technical maximum is far higher and mostly irrelevant, because length past three minutes costs you discovery (socialpilot.co, 2026).
So the real question is not “what length performs best”. It is which of two different signals you are trying to earn, and that depends on what the video is for.
This article covers the ceiling that matters, the two length strategies and when each wins, and how to pick from the job rather than from a benchmark.
Key Takeaways
- Three minutes is the discovery ceiling. Reels over it are not recommended to new audiences (socialpilot.co, 2026).
- Uploaded footage can run much longer, but length past the ceiling trades reach for depth.
- 15 to 30 seconds maximises completion, which suits a single idea with a punchline.
- 45 seconds to 3 minutes maximises accumulated watch time, which suits explanation and personality.
- Pick from the job. A demo, an explainer and a story have different natural lengths and fighting that costs more than it gains.
The ceiling that actually matters
Reels up to three minutes are eligible for discovery. Past that, Instagram does not recommend the video to people who do not already follow you, even though the upload limit is considerably higher and some creation flows go further still (socialpilot.co, 2026).
That makes three minutes a hard planning constraint rather than a guideline. A four-minute Reel is not a slightly worse three-minute Reel. It is a video shown mostly to people who already know you.
There is one legitimate reason to cross it: content made for your existing audience, where reach was never the point. Everything else should come in under.
The two strategies
Short and long both work, for different signals.
| 15 to 30 seconds | 45 seconds to 3 minutes | |
|---|---|---|
| Optimises | Completion rate | Accumulated watch time |
| Suits | One idea, visual payoff, a demo | Explanation, opinion, story |
| Fails when | The idea needs context | The middle sags |
| Risk | Too thin to be worth sending | Loses people before the point |
| Rewatch potential | High | Low |
Completion rate is easiest to win when there is barely time to leave. Watch time is a larger number to accumulate when the video is longer and holds. Neither is better in the abstract, and the ranking signals Instagram has confirmed, watch time, likes and sends, do not pick a side. Instagram ranking signals covers the full on-the-record list.
Picking length from the job
Five common jobs and their natural lengths.
A visual demo. 15 to 25 seconds. The value is in seeing it, and a demo stretched to a minute is a demo with filler in it.
A single strong opinion. 30 to 60 seconds. Long enough to state it and give one reason, short enough that it stays an opinion rather than an essay.
An explainer. 60 seconds to 3 minutes. The job needs context, and cutting context to hit a shorter target is how explainers become confusing.
Personality-led talking to camera. 45 seconds to 3 minutes. The format accumulates watch time rather than winning completion, which is exactly what it is good at. Yapping covers why.
A story. As long as the story needs, under the ceiling. Stories hold through the middle naturally because people want the end.
The failure in all five is the same: picking a target length first and then cutting or padding to reach it.
What to check when length is not the problem
Most “my Reels get no reach” problems are not length problems.
Check whether the opening states its subject, whether the middle sags, whether the file carries another app’s watermark, and whether the content is original. Each of those costs more than thirty seconds of runtime ever will. Why your Reels get no views works through the usual causes in order, and unscripted video that converts covers the structure that keeps a longer video from drifting.
Length and the handoff
A longer video earns more attention and then has the same four seconds to spend it as a short one.
Whatever you mentioned at the end needs to be the first thing visible when someone taps through from your profile. A three-minute video that held somebody to the end has earned more than a fifteen-second one, and losing that person to a page where the thing you named is sixth in a list wastes the difference.
A More.You page can auto-pin the link tied to a scheduled post when it goes live, for a configurable number of hours, so the thing you just talked about sits at the top while the video is circulating. Create your More.You page and stop spending the extra attention at the last step.
What to do with footage longer than three minutes
Three options, in order of how well they usually work.
Split it into parts. Each part gets its own opening and its own point, which is more work than cutting one video in half and considerably more effective. Instagram has been testing a Series feature that groups related Reels into a hub on your profile with a next-episode path, which suits exactly this (techcrunch.com, June 2026). Instagram Series covers how to build one that holds.
Cut it to the single best three minutes. Most long footage has one strong section and a lot of setup. The setup is almost never the part people needed.
Post it long anyway, for the people who already follow you. A legitimate choice when reach was not the goal, as long as you know that is the trade you are making.
What rarely works is speeding the whole thing up to squeeze under the ceiling. The runtime fits and the pacing stops matching how people actually listen, which costs the watch time the length was supposed to earn.
FAQ
What is the best length for an Instagram Reel?
Under three minutes, because past that Reels are not recommended to people who do not follow you (socialpilot.co, 2026). Within that, 15 to 30 seconds favours completion rate and 45 seconds to three minutes favours accumulated watch time.
How long can an Instagram Reel actually be?
Longer than the discovery ceiling. Uploaded footage can run well beyond three minutes and some creation flows go further, but anything past three minutes trades recommendation to new audiences for runtime.
Do shorter Reels get more reach?
Shorter Reels get higher completion rates, which is not the same as more reach. Watch time is the confirmed signal, and a longer video that holds attention accumulates more of it than a short one that is merely finished.
Should every Reel be under 30 seconds?
No. A video cut to 30 seconds when its job needed 90 loses the context that made it worth watching. Pick length from the job and keep it under the three-minute ceiling.
Does Reel length affect whether it shows in search?
Search ranking draws on signals from the account, the content including captions and recency, and engagement such as saves, shares and watch time. Length matters through watch time rather than directly.
Sources
- The three-minute discovery ceiling, reach behaviour past it, and completion-rate guidance for shorter Reels (socialpilot.co, 2026)
- Reels length limits across upload and creation flows (metricool.com, 2026)
- Meta testing Series for episodic Reels on Instagram and Facebook (techcrunch.com, June 2026)
- Watch time, likes and sends as the confirmed ranking signals, as stated by Adam Mosseri, January 2025
- More.You product brief, docs/moreyou/01-product-brief.md, sections 2 and 3 (auto-pin, blocks, analytics)
