InteractiveResearch · Video length·11 min read

How long should a video ad be? The 2026 length playbook.

Six placements, six sweet spots, one rule: the right length is the one your audience actually finishes. Scrub the retention curves, then pick your sweet spot.

Reading mode
Start here

Why length is the wrong question.

Imagine you're at a party and someone starts telling a story. If they get to the point fast, you keep listening. If they ramble, you drift to the snack table. Video ads are exactly the same.

Most people don't watch a whole ad. They watch the first few seconds, decide if it's worth their time, and either stay or scroll. The trick isn't picking a magic number - it's matching the length to where the ad is shown and who's watching.

On TikTok, 30 seconds is fine. On a Meta Feed, that same 30 seconds is way too long. On YouTube before a video, 6 seconds can win awards. Same idea, different rooms, different lengths.

In one line: length isn't a creative decision - it's a placement decision. The placement tells you the budget, the goal tells you how to spend it.

0%

of total drop-off happens in the first 6 seconds

0s

median sweet spot across paid social placements

0x

completion-rate gap, native vertical vs reformatted

0

distinct optimal-length buckets across placement × goal

Retention scrubber

Scrub the timeline. Watch your audience leave.

Each curve is the share of viewers still watching at second t. Toggle placements on or off - the further the lines fall before second 15, the harder your first three frames have to work.

0%25%50%75%100%0s6s15s30s45s60s
6.0s

Meta Feed

In sweet spot
31% still watching

TikTok

Sweet: 21-34s
60% still watching

Shorts

Sweet: 15-30s
55% still watching
Sweet spot finder

Three picks, one length to ship at.

Pick a placement, a goal, and an audience temperature. The recommended duration range updates live - based on placement sweet spots and the modifiers that shift them.

Placement

Campaign goal

Audience temperature

Ship at

7–13s

Confidence

Medium

0s15s30s45s60s

Reels for direct response to cold prospecting audiences. Reels behaves like TikTok: hook by 1.5s, payoff by 7s, end before 15. Vertical-native shoots outperform reformatted feed ads.

Sweet spots, side by side

Every placement has its own attention budget.

The green band is where retention, completion, and CPM align - the red marker is the cliff past which performance falls fast.

Meta Feed (1:1 / 4:5)

615s · ceiling 30s

Sweet spot 6–15s. Andromeda penalizes low 3-second hold-rate, so the open does more work than the length.

Meta Reels (9:16)

715s · ceiling 30s

Reels behaves like TikTok: hook by 1.5s, payoff by 7s, end before 15. Vertical-native shoots outperform reformatted feed ads.

TikTok In-Feed

2134s · ceiling 60s

TikTok's own creative team publishes 21–34s as the strongest in-feed range. Long enough for narrative, short enough to loop.

YouTube Shorts

1530s · ceiling 45s

Loop bias means short Shorts (under 15s) re-watch heavily, but 15–30s carries the best balance of hook + payoff for ads.

YouTube In-Stream

1530s · ceiling 60s

Skippable in-stream: skip-block is at 5s, so frame 0–5 has to earn the next 10. 15s for DR, 30s for consideration, 6s bumpers for reach.

LinkedIn Feed

1530s · ceiling 90s

B2B intent allows longer dwell, but LinkedIn's own data shows view-through still concentrates under 30s. Use 60–90s only for retargeting + product demos.

Sweet spotPerformance ceiling0–60s axis
Pacing anatomy

Anatomy of a 15-second ad.

Five beats, each on its own clock. Drop a beat and the ad loses the viewer at the seam between the two segments that bridge it.

0s5s10s15s
1

0.0–1.5s

Hook

Visually decisive opening - pattern interrupt, identity cue, or problem name. No bumpers, no logo card.

2

1.5–4s

Tease

Land the promise. What is this and why should the next 11 seconds matter? Curiosity gap open.

3

4–9s

Demonstrate

Show the product, the proof, the transformation. This is the load-bearing segment for conversion.

4

9–13s

Confirm

Social proof, review, or stat. Validate that the demo wasn't cherry-picked.

5

13–15s

CTA

Single clear action. End-card or final caption. Brand reveal lives here, not at second 0.

Anti-patterns

Five length mistakes that kill performance

The same five failures show up across creative reviews, regardless of category or budget.

Cropping or trimming a brand film almost never yields a usable short. Bumpers are written backwards - one image, one line, one beat. Treat them as their own format, not a 'shorter version'.
How Shuttergen handles length

One cut in, six length-native cuts out.

When you save a winning ad to Shuttergen, the analysis pipeline identifies the hook, the demo, and the CTA segments - then the remix engine re-paces the same proven scaffolding into the native sweet spot for each placement. A 30-second YouTube cut becomes a 15s Meta Feed cut, a 9s Reels cut, and a 6s YouTube bumper, each ending on its own beat.

That's the length variation Andromeda and TikTok's ranking model actually reward: the same concept, paced for each placement's attention budget.

The checklist

Nine-step pre-ship audit

0/9

audit pass rate

Related Shuttergen reading

Where to go next

The connected pages that compound on this one.

Sources

What we read to build this

Ship the right length, every placement.

Save winning ads, let Shuttergen tag the structural beats, and remix into length-native cuts for Meta, TikTok, YouTube and LinkedIn - without re-shooting.

Get started free