How to Write a Youtube Script (+Template)

Most of the YouTube script feedback I give comes down to the same handful of problems.

"The intro runs long."

"The hook describes the video instead of selling it."

"The middle wanders."

"The call to action tries to do three things at once."

And nobody can tell whether they're stuck on the script or on the delivery, so the next video gets re-shot when it should have been re-written.

I've been writing scripts for the Synthesia channel for almost a year now and I have 10+ years of YouTube experience with 2 Silver Play buttons.

The longer I do it the more convinced I am that the script does more work than anything else in the production.

Cameras are important. Lighting is important. The editing software you use also matters. But the script (and especially the first 200 words) matter more than all of them combined.

This post is the process I actually use, broken into 8 steps. Some of the steps sound obvious. A couple of them will probably sound wrong on first read.

But they work. So stay with me.

How to write a YouTube video script in 8 steps

Turning a script into a video is the easy part nowadays. What’s challenging is writing an awesome script.

The sequence matters more than people realise. Each step is meant to set up the next one. The single most common reason scripts come out flat: people skip ahead to writing dialogue before they've worked out what the video is for.

Step 1: Start with an idea you'd actually click on

I need to be able to fill in the blank with one thing. If I cannot do that then the video does not have a point yet. No matter how clever my writing is later it will not save the video.

Next I come up with a working title. This is not the final title but something close enough to give my script a target. The hardest part for teams is deciding what to leave out. Brand teams want a video to do a lot of things:

• explain the product
• generate demand
• address an objection
• welcome viewers

I can only pick one. The others will get their videos. If I try to do too much my script will not be effective.

For example I had an idea to make a video about translating any video into any language with lip sync. This idea was already popular with our customers so it had a built-in audience. I supported it with a statistic that showed how big the problem was:

70% of business content online is in English but about 19% of people speak it. This one sentence did a lot of work.

What I wish someone had told me years ago is that a good idea can do some of the persuasion before the thumbnail, title or hook even get a chance. The hook is an amplifier. If the idea is weak no amount of amplification will help.

Step 2: Write your hook before anything else

Next I write my hook. The first 15 seconds of the video decide whether the rest of the script matters. YouTube even has its metric for this called Intro retention, and the official YouTube Creator Playbook names the first 15 seconds as the hook window.

A good example of a hook is the Synthesia "How to make AI videos in 3 minutes" video.

In the first 15 seconds the viewer gets three things:

• a clear outcome
• an objection removed
• a credibility signal

The intro itself was generated with an AI avatar. This means the hook is also a product demo. I write the hook before anything. Before the outline before I decide what the body will cover.

There are two formats that work for me:

• the promise hook, which tells the viewer what they are about to get and why now
• the in-media-res hook, which starts in the middle of the action

Whichever format I use the hook has to leave something. A question without an answer or a claim that needs proof. If my hook fully describes the video it is a headline.

Step 3: Set up the gap between what people expect and what you will deliver

Once you have written the hook the next thing to do is to add a piece of psychology that keeps the viewer watching: the gap between what they expect and what you are going to deliver.

People keep watching when what actually happens is better than what they expected.

They stop watching when what actually happens is what they thought it would be. So the script needs to set up an expectation and then beat it.

In practice this means that in the 30 seconds after your hook you should say what the common sense answer is and then promise something different.

For example you could say

‍"Most teams try to fix retention by editing. That does not work. The fix is actually in the script."

Now the viewer knows you are going to say something different from what they expected and they will keep watching at least until you prove it.

This is also how you avoid having a point that is not very interesting. If your first real argument is something the viewer has heard before they will stop watching. The first point of the body has to be either new or sharper than other people's version or framed in a way the viewer has not seen before.

Step 4: Outline in bullet points then order them so the second best point goes first

Once the hook and the setup are in place make a list of every section of the video as a bullet point.

Just the bullet points, not sentences yet. For example if we were making a video about the ai video tutorial maker the outline might look like this:

• Hook
• Start with the Assistant
• Customise your AI avatar
• Edit the script. Turn the video interactive
• Add screen recordings. Let the team collaborate
• Translate the video into 140+ languages with lip sync
• Call to action and outro

Now most guides do not tell you that the order of your bullet points matters a lot.

The rule that surprises people is: lead with your second-best point, not your strongest. Then your best point. Then your third best point. And so on.

The instinct most creators have is to lead with the strongest argument because they are worried viewers will leave before they hear it. What actually happens is the opposite. If you start with the best material the next point will feel weaker by comparison and viewers will feel the drop and stop watching.

If you start with a strong-but-not-strongest point and then deliver something noticeably better you create a pattern of "this keeps getting better" that keeps people watching to see what is next.

If the outline already feels boring on paper the finished video will be boring on camera. Move things around until you would watch it yourself. It is easy to move a bullet point. It is hard to move a paragraph or reshoot a section. So edit ruthlessly while it is still easy.

Step 5: Build each bullet point as a value loop

For each point in the body give the viewer three things in this order:

1. What it is. Explain the point as simply as you can.
2. How to do it. Show the application: an example, a tactic, a real situation.
3. Why it matters. Connect it back to the outcome the viewer came for.

Most YouTube videos do the first step and stop which is why they feel thin. The "how" and the "why" are what move a point from interesting to useful and useful is what gets people to keep watching.

One more pattern to add here is: every body section should zoom out and zoom in. Big picture, tactical, then back to big picture. The rhythm of going wide, then narrow, then wide again is what makes a 10 minute video feel like four.

Step 6: Mark pattern interrupts every 30 to 45 seconds

A pattern interrupt is anything that breaks the visual or audio rhythm of the video. I mark them inline in the script in italics.

For example "cut to b-roll of customer office," "graphic showing the retention curve overlay," "aside to camera saying 'and yes I tried this myself'."

For a 10 minute video that is 13 to 20 interrupts. It sounds like a lot until you actually watch a high-retention channel back with the markers in mind.

If you are working with an editor they will cut exactly what is in the script, no more. If you are your own editor (using something like Synthesia's AI video editor) you need to leave yourself something to work with.

Step 7: End with a call to action that doesn’t feel pushy

One ask. Not three. For example "Subscribe," "Watch the video," or "drop a comment with your worst script story." ‍

Pick one. Make it specific.

A call to action works well when it feels like a next step from what the video has been talking about. If the body has been about how hard it's to write good scripts at scale, a soft mention of a template or tool that solves that exact problem reads as helpful.

Keep the outro short. A sign-off, a thank-you. That's it. 

Step 8: Read it, time it, trim it

Read the script out loud at a natural pace. Wherever you stumble, simplify. Wherever you get bored, cut. Then time the read.

At a pace one minute of finished video runs to roughly 120 to 150 words. If you wrote a 1,700-1800 word script for a five-minute video you are going to talk fast or run long.

Trim before you record. Trimming in post is twice the work and post-shoot you are emotionally attached to the footage. My honest experience is that the first drafts I write are always too long.

I try to write a little more than I need to 110% of what I'm aiming for. This way I can go back. Cut out the extra parts when I read it out loud.

The more I cut the better the video gets. The main thing I learned is to read my work out loud. I time myself to see how long it takes. If I stumble on any part I cut it out.

The final version should be shorter and better than the version I wrote.

YouTube script example and template

As I discussed earlier in this guide, start with a hook. One sentence on what viewers will get and a teaser to keep them watching.

Add a concise setup after your hook. In just a few lines, call out the common sense approach that fails most teams, then pivot to a sharper angle that promises a better way forward.

Divide the body into value loops: define the concept, demonstrate the application, and anchor it to the goal. Mark visual or audio shifts in italics at least twice a minute, leading with your second-best material before hitting the strongest point two-thirds in.

Wrap up with a single, specific ask that solves a pain point mentioned earlier, followed by a ten-second goodbye.

Here is a practical template:

"How we translated 50 marketing videos into 12 languages in a week."

Goal: marketing teams understand how AI dubbing can scale a brand YouTube channel across markets without re-shoots.

The hook is: "Last month our marketing team had 50 product videos in English. We got demand from 12 markets. We didn't have the budget to re-shoot or hire 12 voice actors. So we did something that took less than a week."

Cut to a montage of the product video playing in Spanish, German, Japanese and Portuguese.

The setup is: "Most teams localize by subtitling or re-shooting.. Subtitles get ignored.. Re-shoots blow the budget. We did neither."

The first section is: "We used AI dubbing to translate the speaker's voice into each language. We also matched their lip movements. That alone lifted our retention in Spanish-speaking markets by 40%."

Aside from the camera: "No it doesn't sound robotic. You can try it on one of your videos before you decide."

The second section is: "The unlock was running the workflow across every video in our library overnight. We did twelve languages and fifty videos. No re-recording was needed. The cost per language was lower than a voice-over session."

The third section is: "We also tweaked each language version slightly. We added idioms. We used the product names per region. We used market- examples. Small changes. They made a big lift on engagement."

The call to action is: "If you want to do the same without a hassle you can just use our free AI video translator and localise your videos in hundreds of languages."

The outro is: "That's the playbook. See you next week."

A few things to notice about this script:

• The hook leads with a claim (50 videos, 12 languages, one week). It pays it off immediately with a visual.
• The setup names two answers (subtitles, re-shoot) and rejects both. This opens an expectation gap.
• Each body section follows the value loop: what the move was, how it worked and why it mattered.
• The CTA references a resource the article already discussed. So it lands as a continuation, rather than a sales pitch.
• The outro is one line.

How long should your YouTube video script be?

It depends on the type of video you are making.

• If you are making an explainer video or an educational video it should be around 5 to 20 minutes long (possibly even longer). It really depends on the amount of information you have to share in this case.
• If you are making a brand or product video it should be around 3 to 7 minutes long.
• If you are making a Short it should be under 60 seconds.

The key is to make sure every second of your video is useful.

Once you know how long your video should be you can work backwards to figure out how words your script should be. A good rule of thumb is to assume a speaking pace of around 120 to 150 words per minute. This takes into account pauses and other things that will happen during your video.

• 2-minute video: around 240 to 300 words
• 5-minute video: around 600 to 750 words
• 10-minute video: around 1,200 to 1,500 words
• 15-minute video: around 1,800 to 2,250 words

These are upper limits. My rule of thumb is to write a script that's 110% of the target length and then cut it back by around 10% when I read it out loud. I have found that cutting my script always makes it stronger.

How do you write a script for a YouTube Shorts?

A Shorts script is not a version of a long-form script. It is a different format altogether. If you take a long-form script and just shorten it it will not work well.

There are three rules to keep in mind when writing a Shorts script:

• The first rule is to hook the viewer in one to three seconds. Not 15 seconds. Viewers on vertical screens will swipe away instantly if you do not grab their attention right away.
• The second rule is to not have an introduction. No "hey everyone," no name card, no explanation of what the Short's about. The Short is the value so you should start with it.
• The third rule is to loop the ending. You should end with a sentence that makes the first sentence make sense so the algorithm will replay it. Replays count as watch time and watch time is what the Shorts algorithm rewards.

Here is a template for a Shorts script:

• Open with one sentence that creates an information gap.
• Deliver the payoff in two or three beats over the next 20 seconds. Use on-screen text to reinforce the phrases.
• Close with a line that loops back to the opener.

Your script should be around 50 to 75 words long. If it is any longer than that, you'll have to rush your narration.

YouTube script formats you can use

Outline plus dialogue: 

This is a simple list of sections with the spoken lines underneath. It is fast to write, easy to share with stakeholders and easy to edit. It is best for explainers, talking-head videos, founder messages and product demos.

AV Format:

• Two columns, one for audio and one for visuals on screen.
• Good for interviews, product reviews and videos where the audio track and the visual track are doing different things

VAG Format:

• Like AV. With a third column for graphics.
• Good for training video explainers with charts and on-screen text.

Sequential Flow:

• A running document with action lines, dialogue and direction.
• Good for high-production work, short films and narrative videos.

Voiceover Script:

• Dialogue written as narration with timestamps and visual cues.
• Good for B-roll- videos where the speaker isn't on camera.

Interview Script:

• A list of questions with cue lines.
• Good for podcasts, founder interviews and customer story videos.

If you're shooting alone or with a team and unsure where to start, use a bulleted outline with dialogue. It's fast to write and easy to refine. You can transfer it to formats later.

Mistakes that can ruin a YouTube script

• The hook doesn't grab attention.
• The intro is too long.
• You haven't read the script aloud.
• Sentences are too long.
• There are no pattern interrupts marked.
• The call to action is confusing.
• The outro is too long.
• The word count is off.

Let’s see those in more detail:

1. A good hook should be a claim, a question or a moment not a summary. The viewer already knows what the video is about from the title.
2. If your hook is ten seconds long your intro shouldn’t  be more than 30. A small paragraph is enough.
3. Always read your script aloud before recording. This ensures your script sounds natural. You'll notice things that don't sound right.
4. Long sentences are hard to follow. Break them up into two or more sentences if they take more than five seconds to say.
5. Mark pattern interrupts in your script. These are places to add visuals or music to make the video more interesting.
6. Your call to action should be simple and clear. Pick one thing. Make it specific.
7. Your outro should be short. A simple sign-off line is enough.
8. The word count of your script is important. Aim for 120 to 150 words per minute.

Practical tip: Using a checklist ask a colleague to read the first 200 words aloud and tell you what they think the video is about. If their answer is different, from your goal rewrite the script. Alternatively record yourself reading the hook and first body point and listen to it in the morning. Taking a break can help you see problems you didn't notice before.

Ready to write your script?

Open the template, write your hook first and time yourself reading it out loud.

If you want to skip recording you can use Synthesia's AI video generator to produce a video with an AI avatar in over 140 languages.

Whatever you do next, write the hook first. If you only take one thing from this post, take that.