Facebook Video Ad Length: Limits by Placement and How to Test

As of 27 October 2026, Meta's Help Center lists Facebook Feed at up to 241 minutes and Facebook Stories at 1 to 120 seconds. Facebook in-stream video runs 5 seconds to 10 minutes on mobile but only 5 to 15 seconds on desktop. Instagram Stories, Feed and Explore accept up to 60 minutes.

Source: Meta Business Help Center, video length specifications across placements, as of 27 October 2026.

The 15-second cap you see quoted applies to desktop in-stream only. The same ad on mobile in-stream can run to 10 minutes. Meta's page also notes that shorter videos are usually more engaging on mobile.

Build a 5 to 15 second cut first. The highest minimum is 5 seconds (in-stream) and the lowest maximum is 15 seconds (desktop in-stream). This is our reading of the table above, not a Meta recommendation.

Fitting every placement is a reason to make one short cut, not a reason to make every ad 15 seconds. A product that needs a demonstration, a before-and-after or a testimonial will often need more time. Make the longer cut as a second version of the same idea, so the two can be compared.

Looping changes how you design the ending. Meta loops Facebook videos of 30 seconds or less for up to 90 seconds, sometimes up to 180. Its own examples are a 30-second video playing 3 times and a 16-second video playing 6 times, because 96 seconds is closer to 90 than 80 is. By the same rule a 15-second video plays 6 times, which is 90 seconds. Facebook videos over 30 seconds do not loop, and the viewer must tap play to replay. Instagram videos loop for as long as they are visible. Source: Meta Help Center, about video length and looping.

So design the last frame of a 30-second-or-shorter ad to work when the video restarts, and put the call to action where a viewer who has seen it once will still see it. On a longer ad, assume the viewer sees it once.

For vertical placements, read our guide to Reels ads versus Feed first. Reels and Feed can price the same ad differently, and that gap is easy to blame on length.

A short ad can show a higher hold rate because its ThruPlay needs fewer seconds of watching than a long ad's 15 seconds. Hold rate divides ThruPlays by three-second plays, so ads under 15 seconds are measured on completion and longer ads on 15 seconds of watching. The two numbers do not measure the same thing.

Meta defines a ThruPlay as a play of at least 15 seconds, or to 97% completion if the video is shorter than 15 seconds. You can also choose a 6-second goal, which counts a play of at least 6 seconds, or 97% completion if shorter. ThruPlay goals are available in the Awareness and Engagement objectives. If your ThruPlay goal is 6 seconds, the break point moves to 6 seconds; compare hold rate only between ads on the same side of the goal you chose. Source: Meta Help Center, about ThruPlay.

Applying the 15-second goal to four lengths gives the share of the video a viewer must watch. This is our arithmetic from Meta's rule: 97% of 10 seconds is 9.7 seconds, and 15 seconds divided by 20, 30 and 60 gives 75%, 50% and 25%.

A 10-second ad has to be watched almost to the end, and a 60-second ad only to the quarter mark. A high hold rate on a short ad tells you its viewers watched it through. It does not tell you the ad is better, and it says nothing about what each result cost.

Our definitions: hook rate is three-second plays divided by impressions, and hold rate is ThruPlays divided by three-second plays. Our guides cover each in detail: hook rate and hold rate.

Test it by running two cuts of one concept against the same audience and budget, then comparing cost per result once each ad has enough views. Export your last 60 to 90 days of video ads and fill in one row per length bucket.

Hook rate is 3-second plays divided by impressions. Hold rate is ThruPlays divided by 3-second plays. Cost per result is spend divided by results. Then run four checks.

Is the only difference between the two cuts their length?

Pass: the same hook, offer, audience, budget and placements, with only the running time changed. Fail: a different opening line, price or creator. If any of those changed, you are comparing concepts, not lengths, and you cannot read the result.

Does each ad have enough views to read?

Pass: at least about 669 three-second plays per ad before you read a hold rate, and at least about 1,250 impressions per ad before you read a hook rate. Fail: anything smaller, in which case you wait or pool similar ads. These two thresholds are our model, not Meta figures: a two-proportion test at 95% confidence and 80% power, with a hold rate falling from 22% to 16% and a hook rate near 25% differing by five points. They apply only to gaps that size, and smaller gaps need more. Our guides cover the background: hold rate and hook rate.

Cost per result, the metric that decides the winner, needs its own volume. Pass: one more or one fewer result on either cut would not change which cut is cheaper. Fail: it would, so keep the test running.

Are you comparing hold rate fairly?

Pass: hold rate is compared only between ads on the same side of your ThruPlay goal (15 seconds, or 6 seconds if you chose that goal), as the ThruPlay table above shows, and the verdict rests on cost per result. Fail: calling a 10-second ad the winner because its hold rate is higher than a 40-second ad's.

Does the winner hold in the placement that carries your spend?

Pass: when you break results out by placement, the longer cut wins or ties where most of your budget runs. Fail: it wins overall but loses in your biggest placement, because Reels, Feed and Stories do not price or play an ad the same way. If the two cuts tie, keep the short one, because it fits every placement.

Before launch, preview each active placement in Ads Manager. Then confirm the ad looks right where it will actually run.

Length is worth a slot when your ads all sit at one running time, so you have never seen another. If every ad in your account runs 30 to 45 seconds, the audit above has nothing to compare, and a 15-second cut of your best ad is a cheap test.

Length still competes with every other test you could run, including hooks, offers and formats. That ranking is the work we do. We supply a creative intelligence platform for consumer brands and agencies running Meta advertising. It combines ad-performance data, competitor activity, customer signals and past ad learnings into ranked weekly test plans and execution-ready creative briefs. The audit tells you what length did on your account. The plan puts that result next to the other things worth testing this week.

To check an opening before you spend on it, export its first frame as a still image and upload it to our free Attention Heatmap. It takes JPEG, PNG or WebP files, up to five images of 10 MB each, with no signup. It shows predicted attention on a frame, not your hook rate or hold rate.

If you want a ranked plan built from your own account, see Deepsolv pricing or book a demo with our founding team.