How Long Should a Video Ad Be? Lengths That Work by Platform and Goal

Short answer: start at 15 seconds for most social feeds, 6 seconds for reminder ads, and 30 to 60 seconds only when viewers need to see proof before they act. From there, keep cutting until the ad stops working, then go back one step.

That's the method. The rest of this article turns it into lengths you can use on each platform, because "it depends on the platform" is true but doesn't help you brief a video.

Allowed length and useful length are different numbers

Every ad platform publishes a maximum length. That number tells you what the upload form will accept. It tells you nothing about what people will watch.

A few examples from the official specs:

• TikTok auction in-feed ads can run up to 10 minutes, and Spark Ads have no duration limit at all, according to TikTok's in-feed ad specs.
• LinkedIn sponsored video can run from 3 seconds to 30 minutes, depending on the campaign objective (LinkedIn video ads).
• YouTube in-feed video ads have no maximum length (YouTube ad formats).

Nobody should read that and plan a 10-minute TikTok ad. Platforms set high limits so they can host many kinds of video, including long walkthroughs that someone chose to watch. A cold-audience ad that interrupts someone's scroll is a different situation.

It helps to treat duration as a budget. Every second costs some attention, and you spend it on four things: a hook, a promise, proof, and a next step. Once all four are done, the ad is finished, even if the platform would let you keep going.

The rule

The right length is the shortest cut that delivers one promise, enough proof to believe it, and one next step.

Meta's video advertising guidance makes a similar point. Length should depend on placement and funnel stage, shorter videos tend to hold mobile viewers better, and the product or brand message should show up early with one clear message.

The part that changes from ad to ad is "enough proof." For a $15 phone case, showing it in someone's hand may be all the proof needed. For B2B software, proof might be 20 seconds of the dashboard doing the thing you promised. That's why the same rule gives one brand a 6-second ad and another a 45-second ad.

Starting lengths by platform

The second column comes from official specs. The third column is an editorial starting point for testing, not a platform rule.

How long should a Facebook ad video be?

For Facebook and Instagram feed placements, start at 15 seconds. The same answer works if you're asking how long a Meta video ad should be in general, with one exception: in-stream. Meta's in-stream ad specs say ads of 15 seconds or less play in full, while longer ads may get cut off after 15 seconds, and desktop in-stream accepts only 5 to 15 seconds. Whatever mobile allows, 15 seconds is the practical ceiling for in-stream.

Reels are stricter in a different way. Your ad sits between organic videos, so anything that looks like a TV spot gets swiped past. Aim for 10 to 20 seconds, film or generate it vertically, and open the way a creator would open a post.

How long should a TikTok ad video be?

Start between 15 and 30 seconds. The 10-minute limit is real, but nobody needs it for a cold ad. Spark Ads are a reasonable exception. Because they boost an existing organic post, you can check that post's retention before you spend money on it. If the organic version holds people for 45 seconds, run it at 45.

YouTube

YouTube has the clearest constraints. Bumpers are 6 seconds and can't be skipped, so they hold exactly one idea. Non-skippable in-stream ads get full attention, which is a reason to keep them tight. Skippable in-stream ads can run longer, but viewers can skip after 5 seconds, so those first 5 seconds have to deliver the brand and the promise even if nobody watches second 6. In-feed is the one YouTube placement where 2 or 3 minutes is normal, because the viewer chose to click.

LinkedIn

LinkedIn video autoplays muted, and LinkedIn strongly recommends captions. Shorter videos catch attention in the feed, and longer ones leave room for a deeper story. For cold audiences, 15 to 30 seconds with on-screen text is a solid start. Go to 60 to 90 seconds when the viewer already knows the problem and is comparing options.

CTV

Connected TV works differently. Many buys are sold in fixed slot lengths, so ask the publisher or platform which lengths they accept before you edit anything.

How the goal changes the length

The platform sets the range. The ad's job decides where in that range you land.

Cold and retargeting audiences need different ads, even from the same campaign. A cold viewer needs the problem set up before the product means anything. A retargeted viewer has already been to your site, so the setup wastes their time. Give them the offer, the price, or the answer to whatever stopped them from buying. That's often a 6-second cut. For how these stages connect, see this full-funnel marketing strategy breakdown, and for direct response specifically, this guide to performance marketing video.

A time budget to plan against

This is a planning template, not a benchmark. It's a way to see where your seconds go before anyone starts producing.

15-second cut:

30-second cut:

The hook carries the most weight. Meta says the first 3 to 5 seconds matter most, and in a skippable or scrolling placement they decide whether anyone sees the rest. If you need opening ideas, start with these video hooks that stop the scroll.

Check every block with the sound off. If the promise only exists in the voiceover, a muted feed never hears it. Put it on screen.

Make a cutdown ladder, not one master

The most common mistake is making one 30-second video and then either stretching it or cramming it into every placement. Plan a set of cuts from one concept instead.

A few rules make the ladder work:

• Write each cut as its own script. A good 15-second cut is rarely just the first 15 seconds of the 30.
• Script the 30 first, since it holds the full argument, then compress it into the 15 and the 6.
• Keep the promise the same across all cuts. If the message changes too, a test can't tell you whether length made the difference.
• Change the opening for each placement. Use a vertical, person-first open for TikTok and Reels, put the product and brand on screen within 5 seconds for YouTube skippable, and add captions everywhere.

If you're briefing an AI video creator on Viralix, put the full ladder in the brief so the package, timeline, revisions, and rights cover every cut.

Where length goes wrong

The 15-second ad with six claims

Fast price, free shipping, three colors, eco materials, 4.8 stars, and a discount code, all in 15 seconds. Viewers don't remember any of it. Keep one claim and move the others into other cuts or other ads.

The 30-second demo that earns its length

Imagine a booking app whose whole pitch is "three taps to schedule." The viewer needs to watch those three taps happen. Cut it to 15 seconds and you're left with a logo and a promise nobody believes. Here the extra time pays for proof, and proof is what sells.

The skippable ad that saves the brand for the end

The ad builds a mood for 20 seconds, then reveals the product. On YouTube skippable placements, most people never get there. If the first 5 seconds don't name the product and the reason to care, assume the ad has already failed.

The voiceover-only ad in a muted feed

LinkedIn feed video ads autoplay muted, and you can't count on sound in any social feed. If an ad only makes sense with audio, a muted viewer sees a silent clip of b-roll. Every social-feed cut should still communicate its promise and its action without sound, so put both in on-screen text.

"We're allowed 10 minutes"

You are, on TikTok and on mobile Facebook in-stream for some objectives. But the limit is there for content people choose to watch. A cold ad that runs long usually means nobody made the hard editing decisions.

When longer is justified, and when it's weak editing

Longer is justified when:

• The product has to be seen working before anyone believes it.
• The purchase is expensive or considered, and you need to answer objections in the ad.
• The audience is warm and already knows the problem.
• The viewer chose to watch, as with YouTube in-feed or a LinkedIn video they clicked to expand.

Longer is usually weak editing when you see:

• A logo intro before anything happens.
• Setup that repeats what the hook already said.
• Three proof points that prove the same thing.
• Slow atmosphere shots, which is common in AI-generated ads, where a nice-looking clip gets kept because it took effort to make.
• The call to action said twice.

Here's a quick test. Remove each shot one at a time. If the promise, the proof, and the next step all still come through, that shot goes.

Read the results honestly

Shorter cuts tend to have higher completion rates simply because there's less to finish. That doesn't make them better ads. Compare cuts on the result you're paying for, whether that's cost per purchase, cost per lead, or click-through rate. Use completion rate to find where people drop off, not to pick a winner. This guide to video completion rate covers how to read it.

What to do next

1. Write your next ad's job as one sentence: the promise, the proof, and the next step. If it takes two sentences, you have two ads.
2. Script a 30-second version, then write separate 15- and 6-second scripts with the same promise and a different opening for each placement.
3. Run the cuts against each other in the same campaign and choose the winner on cost per result, not completion rate.