In short: As long as it stays interesting, and not a second longer. There is no optimal duration, and every "ads should be X seconds" rule you have read is someone generalising from their own account. What is true is that the audience drops away sharply in the opening seconds of almost every video ad ever made, so the beginning carries most of the weight. That is an argument for making the start extraordinary, not for making the whole thing short. A ninety second video that holds someone beats a ten second one that loses them at three.

Why the ideal length question is the wrong one

People ask this hoping for a number to build to. It does not exist, because length is not what decides whether a video ad works. Attention is.

Two videos of the same duration perform completely differently depending on what happens in their first moments. Two videos of wildly different durations can both work. Meta will happily show a fifteen second ad and a two minute one to the same person on the same day. The system is not measuring your runtime. It is measuring whether people keep watching and whether they act.

So the useful question is not how long it should be, but how long you can stay worth watching. Then stop there.

The first three seconds are most of the job

Nobody is looking for your ad. They are scrolling at speed with their thumb already moving. You get about a second of grudging attention before a decision gets made, and roughly three before it is final.

Which means the opening frame is the most important creative choice in the whole video. Not the offer, not the music, not the call to action. If the first moment does not earn the second one, nothing after it exists.

Four ways to buy that opening:

  • A reveal. Something wrapped, covered, closed or unfinished, resolved a beat later. The brain does not like an unclosed loop.
  • A zoom. Start impossibly close on texture, glaze, grain, stitch, then pull back to show what it is.
  • Motion that should not be there. The object rotating, tipping, dropped and caught, poured into.
  • A face and a sentence. Someone saying something specific and slightly unexpected, immediately, with no preamble.

What does not work: a logo, a title card, a slow fade in, or three seconds of establishing shot. Those are the visual language of an advert, and the moment someone recognises an advert, the decision is already made.

A structure that holds up

If you want a shape to build to, this one works across most products.

A shape that holds up

01

Hook

First 3 seconds

02

Reason to trust

As long as it takes

03

Why yours

As long as it takes

04

What to do now

One instruction

Only the first block has a time attached to it. That is the whole point: the rest take as long as they honestly take.

The hook is one of the four above, and it is where the disproportionate effort goes. The reason to trust is who you are, how long you have been doing this, that other people buy it. Why yours is the specific, concrete difference: not "quality craftsmanship", but what it is made of, how long it lasts, what it does that the cheap one does not. Then one instruction, not three.

When longer genuinely wins

Long videos have a reputation for failing, which comes from long videos that were boring, not from long ones as a category.

A longer edit tends to beat a short one when:

  • The product needs explaining. Anything where the value is not visible in a photograph. A process, a material, a technique, a service.
  • The price is high enough to need justifying. People will happily watch two minutes before spending several hundred dollars. They will not watch ten seconds of a hard sell.
  • You are the one talking. Someone speaking plainly about work they actually do holds attention far past where a produced advert loses it.
  • The story is genuinely good. If the making of the thing is interesting, let it be interesting.

The rule underneath all of those: the audience decides when it is too long, and it tells you by leaving. That is a number you can read after the fact rather than guess at beforehand.

Things that matter more than duration

  • Shoot vertical. Most of your audience is holding a phone upright, and a video that fills the screen wins attention before content enters into it.
  • Assume the sound is off. A large share of people watch muted. If your message only exists in the audio, it does not exist. Caption it.
  • Make more than one length. The same footage cut at fifteen seconds and at sixty gives the system two genuinely different options, and people differ in what holds them.
  • Check the current specs in Ads Manager, not in an article. Aspect ratios, file limits and placement requirements change often enough that any figure written here would be wrong before long.

So what should you actually do

Put a disproportionate amount of your effort into the first three seconds. Then keep going for exactly as long as you are still worth watching, and stop. Cut the same material into a short version and a long one, run both, and let the numbers tell you which your audience preferred. That answer is specific to your product, and it is worth more than any rule of thumb.


Related: how many ads you should be running, and which type of ad to make first.

If you would rather hand the whole thing over, I set up and run Instagram and Facebook ad campaigns for Melbourne businesses.