Video ad length has a practical answer that removes most of the debate: five to fifteen seconds clears every documented Meta placement requirement. A video inside that window can run anywhere the system chooses to put it.
That is an eligibility argument rather than a performance one, and eligibility is the constraint people usually discover late.
This is for you if you are commissioning video and want one specification that will not need redoing.
Why the five to fifteen second range works
Because it satisfies the documented requirements everywhere at once.
Meta’s placement requirements put in-stream at five to fifteen seconds and Facebook Reels at four to fifteen. A video sitting inside both windows qualifies for both, and for the placements with looser rules as well.
So one asset covers your whole delivery. That matters more than it sounds, because Meta decides where to run things, and an asset that is ineligible somewhere quietly narrows the pool it can choose from.
The alternative is producing several cuts for several placements, which multiplies production cost for an account that probably cannot supply enough angles as it is.
One video, correctly specified, running everywhere, is the efficient answer for most accounts.
What this range is not telling you
It is an eligibility rule, not a performance finding, and conflating the two is easy.
Nothing here says a twelve second video outperforms a six second one. That comparison has no published answer with a disclosed sample, in the same way the static against video comparison does not.
What the range says is where your advert is allowed to appear. Inside it, everywhere. Outside it, somewhere less, and which somewhere depends on how far outside you go.
So treat five to fifteen as the constraint you design within, then decide the exact duration from your own data rather than from a rule.
How to pick the length inside the range
Your own retention curve answers this, and most accounts already have the data.
Look at where people fall away on your existing videos. If retention collapses consistently at nine seconds, the material after nine seconds is doing nothing for the people who left and is not reaching the ones who stayed either.
Cutting to that point costs you nothing. It also improves eligibility margin and reduces production time.
Do that per angle rather than per account, because different arguments hold attention for different lengths. A simple offer resolves in four seconds. A comparison might genuinely need twelve.
The single number to avoid is one chosen because it felt right in the edit. That is the most common way videos end up at twenty-three seconds and ineligible in half their placements.
Why longer video ad length rarely earns its place
The argument for it is usually that the product needs explaining, and there is a better tool for that.
A landing page explains at the reader’s pace. They can re-read, skim, skip to the part that matters and return. A video explains at your pace, and rewinding is friction most people will not accept.
So the division of labour is worth stating. The advert earns the click by naming a problem somebody recognises. The page does the explaining.
Trying to do both in the advert produces a video that is too long to be delivered widely and too shallow to actually explain anything, which is the worst of both.
Where a genuinely long-form video is the right creative, it is usually a different campaign with a different objective rather than a prospecting advert.
The specification worth writing into briefs
Four lines, and they save a round of revisions every time.
Length between five and fifteen seconds, stated as a hard constraint rather than a preference.
Must work without sound, with captions burned in rather than relying on platform captioning.
The argument in the first two seconds, because the range assumes people decide early.
And a version that survives cropping to both square and vertical, since the placements that accept your video will not all show it the same shape.
Those four cover most of what goes wrong between brief and delivery, and none of them is a creative constraint in the sense that limits what the video can say.
Why specifications need dating
One discipline worth adopting on anything platform-related.
Requirements change. The range described here reflects documented requirements as of the source cited, and it may move.
So when you write a specification into a brief template or a process document, put the date on it and note where it came from. A specification with a date can be checked. One without reads as permanent and quietly rots.
That habit matters more in this topic than almost anywhere else, because platform specifications are exactly the sort of thing people copy into internal documents and never revisit.
Before a major production run, check the current documented requirements rather than trusting the brief template.
What video ad length does to cost
Not directly through the auction, and indirectly through where you can be delivered.
An asset ineligible in several placements restricts where your budget can go. The system has fewer options, so it competes harder in the placements that remain, which can raise what you pay.
That is a plausible mechanism rather than a documented one, and it argues for eligibility on its own terms regardless.
The clearer cost is production. Longer videos take longer to make, longer to revise and longer to replace when they stop working, which reduces how many angles your budget buys.
Given that angle count is usually the binding constraint, that production cost is the one worth taking seriously.
Where video ad length interacts with fatigue
One connection worth making, because it changes the replacement calculation.
A shorter video is cheaper to remake. When an angle is still working but the execution has worn out, a five second cut can be reshot or re-edited quickly, while a fifteen second one is closer to starting again.
So video ad length affects how easily you can refresh, not just how widely you can deliver. An account producing short assets can rotate more often for the same budget.
That compounds with everything else in this cluster. More rotation means more angles tested, and more angles tested means more chances at the roughly one in twenty that becomes a genuine winner.
None of which argues for the shortest possible video. It argues for treating length as a production decision with downstream consequences rather than as a purely creative one.
What to change this week
Three steps.
Check the length of every video currently running against the five to fifteen second window, and note which are ineligible somewhere.
Then look at your retention curves and find where people actually leave, so your next cut ends there rather than where the edit felt finished.
Then put the four-line specification into your brief template, with a date on it.
Fund the ad set and read what the algorithm chose is the position underneath, static vs video ads covers the format question this sits inside, briefing UGC creators covers getting the specification across without over-directing, and the weekly rhythm I run covers how often you will be commissioning these. What I do across paid, SEO and the website keeps the advert and the page working as one piece. To review what your current assets are eligible for, book a teardown.