How to Write a 40-Second Short People Actually Watch to the End
A five-beat script structure, the word counts that fit, and the opening habit that got my Shorts swiped away
One of the best Shorts on my channels lately is a 56-second story about the night hip hop was born in the Bronx in 1973. YouTube says the average viewer watched 39 seconds of it. That’s about 70% of the video, on a small channel, with no face and no trending sound.
In the same weeks, I was also running tutorial-style Shorts on a couple of other channels. Those opened with a setup: a bit of context, then the interesting part. Most viewers swiped them away in the first second or two. In the reports, “stayed to watch” sat between 16% and 29%.
Same few weeks, similar subjects, same person writing them. The biggest difference was the first line, and the way the rest of the script was built around it.
So here’s the structure I write every 40-second Short in now. It’s five beats, and each one has a job.
First, the math: 40 seconds is about 100 words
Most narration runs at roughly two and a half words per second. That puts a 40-second Short at around 100 spoken words. Not 150. Not “as many as I can fit”.
That number changes how you write. You don’t have room for a warm-up, a recap or a sign-off. Every sentence has to either raise a question or answer one.
Beat 1: the claim (0 to 3 seconds, about 8 words)
Open on the most surprising true thing in the story. Not the topic. The surprise.
- Weak: “Let’s talk about the history of hip hop.”
- Strong: “Hip hop started at a back-to-school party.”
The weak line tells people what the video is about. The strong one makes them want to know how. That’s the whole job of beat one: create a small question the viewer can’t leave unanswered.
Things that don’t belong in the first three seconds: “hey guys”, your channel name, “in this video”, a logo, or a slow establishing shot. Each one is a reason to swipe.
Beat 2: the stakes (3 to 8 seconds, about 12 words)
Now tell them why the claim matters, or make it stranger. One sentence.
“August 11, 1973. A rec room in the Bronx, a party thrown to raise money for school clothes.”
You’ve added a date, a place and a person. The viewer now has a picture in their head, and a picture is much harder to swipe away than an idea.
Beat 3: the turns (8 to 30 seconds, about 55 words)
This is the body of the Short, and it’s where most scripts lose people. The fix is to treat it as three small beats, not one long explanation.
- One fact per sentence. If a sentence has “and” in the middle, it’s probably two sentences.
- Each beat ends on a small open question. “But the crowd didn’t want the songs. They wanted one part of them.” Now they need the next line.
- Move forward in time or in stakes. Never sideways. A beat that repeats the last one in different words is where your retention graph dips.
For the hip hop story, the three turns are: the crowd only danced during the instrumental breaks; the DJ, Kool Herc, put two copies of the same record on two turntables and switched between them to stretch the break; people started dancing and rhyming over that stretched-out break.
Beat 4: the payoff (30 to 37 seconds, about 18 words)
Answer the question you opened with. Clearly, and in plain words. “That trick, at that party, is where most histories say hip hop starts.”
Get Kavithakanth’s stories in your inbox
Join Medium for free to get updates from this writer.
A lot of creators save a bonus fact for here. Don’t. If the viewer stayed for 30 seconds, they stayed for the answer. Give it to them.
Beat 5: the loop line (37 to 40 seconds, about 7 words)
The last line should lead back into the first one, so the replay feels natural. If your video ends on “follow for more”, the ending is a goodbye, and the graph drops off a cliff.
Try ending on a line that makes the opening land differently the second time: “All because of a back-to-school party.” Now the first line plays again and it means more.
When YouTube shows an average percentage viewed above 100% on a Short, this is usually why: people watched it more than once.
The whole script, put together
Here’s the hip hop Short as one script. Read it out loud and time it. It lands right around 40 seconds.
Hip hop started at a back-to-school party.
August 11, 1973. A rec room in the Bronx, a party thrown to raise money for school clothes.
The DJ was an 18-year-old called Kool Herc. And he noticed something.
The crowd didn’t care about the songs. They waited for one part: the break, when the singing stops and only the drums play.
So Herc put two copies of the same record on two turntables. When the break ended on one, he started it on the other.
The break never ended. People danced to it. Then someone started talking over it.
That trick, at that party, is where most histories say hip hop starts.
All because of a back-to-school party.
Notice what isn’t there: no intro, no “did you know”, no explanation of what a breakbeat is before you need it, and no request to subscribe. Every line either opens a question or closes one.
The five habits that quietly kill retention
- Repeating the title. The viewer already read it. Start one step past it.
- Explaining before showing. “To understand this, you need to know…” is a swipe.
- Numbers without a picture. “1973” is a fact. “A rec room in the Bronx in 1973” is a scene.
- A visual that doesn’t change. If the first image sits for five seconds, the video feels slow even when the words aren’t.
- Ending on a call to action. Put the ask in the description or a pinned comment. Let the video end on the story.
How I actually make them
I write the script first, in a plain doc, using the five beats. Then I read it out loud with a timer. If it runs past 45 seconds, something in beat three goes.
After that I make mine in Vidpal. I paste the script in as a Full Script, pick a visual style that suits the niche, and choose how the video opens. If my first line is strong enough, I set the hook intro to None so the video starts straight on it. If it needs help, I try a text reveal. The voice, visuals and captions come out of the same place, so I spend my time on the part that decides retention, which is the script.
Whatever you use to make the video, the structure is the part worth copying.
Test one thing at a time
When a Short underperforms, it’s tempting to change everything for the next one. Don’t. Change one beat.
- Low “stayed to watch”? Only rewrite beat one.
- Viewers stay but leave halfway? Look at the first steep drop in the retention graph. The sentence right before it is usually a sideways beat.
- Good retention but no replays? Rework the loop line.
Compare each Short against your own channel’s typical numbers, not against a screenshot from someone with ten times your audience. Small channels swing a lot from video to video. What you’re looking for is the pattern across ten Shorts.
The takeaway
A 40-second Short isn’t a short video. It’s a 100-word story with five jobs: make a claim, raise the stakes, turn three times, pay it off, and loop back.
Write those five beats before you think about visuals, music or captions. The script decides whether people stay. Everything else decides whether they enjoy staying.
