Guide
Twenty to thirty seconds, and here is how to argue yours
Twenty to thirty seconds for most direct response. Anything outside fifteen to sixty is a decision you should be able to defend out loud. The number sits where two curves cross. Viewers leave continuously, and a viewer has to understand the offer before they can act. This page covers where your own two curves meet, the four questions that set the number, and how to test a pair. By the end you will have a length for your account rather than for somebody else's.
- The default to argue away from
- 20-30s
- What thirty seconds actually holds
- 75 words
- Of most scripts is the third example
- 30%
Decay and sufficiency, and the crossing point is yours alone
Decay is the first force. Viewers leave continuously. Every extra second costs a slice of the remaining audience whatever the ad is doing.
Sufficiency is the second. A viewer has to understand what you sell, why it is different and what to do next. Short of that they cannot convert, however long they stayed.
Both curves are yours specifically. That is why borrowing somebody else's number fails. Your decay depends on your audience and your opening. Your sufficiency depends on your price, your category and how much explaining the product needs.
A cheap, obvious, visually self-explanatory product crosses early. Fifteen seconds is plenty. A considered purchase that needs a mechanism explained crosses much later. Cutting that to fifteen seconds does not make it perform. It makes it incomprehensible.
That is why length advice contradicts itself everywhere you read it. Nobody is lying. They are describing where their own two curves crossed.
What each length buys you, in spoken words
At two and a half words a second. Count your script before you argue about duration.
- 15 seconds38 words
- 30 seconds75 words
- 60 seconds150 words
A three second ad that sells nothing has a perfect completion rate
The argument for short is that completion rates are higher and the message is denser. Both are true. Neither is the goal.
Completion rate is a metric you can maximise by making a three second ad that sells nothing. What matters is how many people arrive at the ask having been persuaded.
A sixty second ad that keeps a fifth of its viewers and convinces them can beat a fifteen second ad that convinces nobody. The second one looks better on every retention chart in the account. That is how it survives so long.
Longer ads also self-select. Somebody who watched fifty seconds has shown more interest than a hundred people who watched five. That effect shows up in cost per landing page view rather than in completion rate.
None of that is an argument for length as a strategy. Most long ads do not fail because they are long. They fail on the third example, the caveat, and the sentence that restates the sentence before it.
There is one more reason the short camp keeps winning arguments, and it has nothing to do with viewers. Short ads are cheaper to produce. Teams that favour them make more of them, and volume flatters whichever format produced it.
Separate those two effects before concluding anything about duration. Otherwise you are measuring your own production constraints and filing it under audience behaviour.
This is the whole editor
Highlight a phrase and a clip lands on those exact words. No timeline, no keyframes, no layers.
Four questions, answered in writing before you script anything
Every one of them can be answered before you write a word, and the answers decide the number without a meeting.
Answer them in writing rather than in your head. A length chosen out loud always drifts towards whatever the last ad was.
- How much does it cost? A higher price means more objections to clear, which means longer.
- Is it obvious on sight? A product whose value is visible in one shot needs less time than one whose value is a claim.
- Is the audience cold or warm? People who already know you need the reminder and the offer, not the education.
- How many claims are you making? Each claim needs evidence. Two claims properly supported beat five asserted, and they fit in less time.
Four answers, then a number
Each answer pushes the number one way. Write them down and the length argument stops being a matter of taste.
Price
Higher means longer
Obvious on sight?
Visible value means shorter
Cold or warm?
Cold needs framing
How many claims?
One, supported
Cut from the middle, and treat the short version as a separate ad
Length is the easiest variable to test, because it needs no new material. Record one take. Produce a short version and a long version of the same argument. Run them against each other.
When you cut down, cut from the middle. The opening buys attention and the close asks for action, so both survive. What goes is the third example, the caveat, the restatement. Almost every ad has thirty percent of that in it.
Treat the short version as a separate creative rather than a compressed one. The best short ad is built around a single idea lifted out of the long one. Compression without selection feels rushed and says less.
Watch the pace as well as the duration. A forty second ad with a cut every two seconds feels faster than a twenty second ad held on one static shot.
In Cutroom pace is a lever rather than an edit. Change it and the whole cut rebuilds. A fast version and a slow version of the same take cost minutes instead of an evening.
The source ceiling here is three minutes, which is built for short-form direct response. That is roughly 450 spoken words, far more than a thirty second ad needs.
Run the pair for a fortnight. Log which won and by how much. That result beats any length guidance on the internet, this page included.
Log it in one line. Two lengths, one winner, the margin, the volume. Six of those entries settle the length question for your account permanently.
Rerun it after any change of offer. Sufficiency moves when the price does. A length that was right at forty dollars is rarely right at four hundred.
Three seconds in every ten are the caveat
Roughly thirty percent of a written script is restatement. Find it before you argue about duration.
Questions people ask
- Is there a minimum length for an ad to work?
- Below about eight seconds you have room for a hook and an ask and nothing else, which limits you to reminding warm audiences of something they already know. That is a legitimate job. It is not the job most advertisers are trying to do.
- Do longer ads get worse distribution?
- Not directly for paid placements, where delivery follows performance rather than duration. Indirectly, longer ads have more opportunities to lose people, so weak long ads look much worse than weak short ones. Length punishes sloppiness rather than being penalised in itself.
- Should I cut my long ad down into a short one?
- Yes, and treat it as a separate creative rather than a compressed version. The best short ad is usually built around one idea from the long one, not the long one with the gaps removed.
- What length works for a cold audience?
- Longer than you expect. A cold audience needs the problem framed before the product means anything. Twenty to forty seconds is a reasonable working range. Going extremely short with strangers sells to people who were already convinced.
- Is it right for a four minute explainer?
- No. Build that on a timeline. Then record separate short takes for the ads that sell it, which is the job this does in minutes rather than evenings.
Long enough to make the case, short enough that the case is still being watched. Test the pair. In Cutroom a second length is a pace lever and a 10 credit export rather than a second edit.