• YouTube Scripts
  • Retention
  • Content Strategy

How to Write a YouTube Script (Template + Example)

A practical YouTube script template: hook, promise, chapters, and retention techniques — with a full example you can copy.

Kallaway — a script strategist whose channel has crossed a billion combined views — opens his own breakdown of script writing with a claim worth sitting with: he has watched hundreds of top creators across every niche, and they all run “the exact same script writing process.” Not the same personality, not the same editing. The same structural decisions, made in the same order, every single time. This guide is that structure — how to write a YouTube script that survives the first 15 seconds, earns a chapter or two of trust, and pays off what it promised: a hook formula, a way to pace chapters against your own retention graph, an open-loop technique, and a complete template you can paste into a doc and fill in today. If you want the general craft underneath all of this — three-act structure and the words-per-minute math that sets your runtime — that lives in the companion guide; this one stays specific to what YouTube itself rewards and measures.

How to write a YouTube script that survives the first 15 seconds

Start with what YouTube itself actually measures, because the popular version of this advice is slightly wrong. Scroll through YouTube-strategy accounts right now and you’ll see “the 15-second rule” stated as fact — lose them before second 15 and the algorithm buries you. That number isn’t in YouTube’s own documentation anywhere. What YouTube actually defines and reports, in Studio’s audience retention tab, is “Intro performance” — the percentage of viewers still watching after the first 30 seconds. “15 seconds” is a third-party convention, useful as a writing discipline, just not a YouTube-published threshold. Say what you mean: write for the first 15 seconds if that’s the habit that works for you, but don’t tell yourself YouTube is grading you on it specifically.

What YouTube does tell you, in the same help article, is how to read the four shapes your retention graph will take: a flat line means viewers are “watching that part of your video from start to finish,” a gradual decline means they’re “losing interest over time” (expected — “videos on YouTube generally taper off during the playback period”), spikes mean people are “watching, rewatching, or sharing” that moment, and dips mean viewers are “abandoning or skipping” it. That vocabulary — flat, decline, spike, dip — is the diagnostic language the rest of this guide uses.

What a healthy retention graph actually looks like

30 sec

YouTube's own "Intro" retention window — not 15

60%+

Retention past 30s that vidIQ calls healthy

65%+

vidIQ's bar for a genuinely strong hook

~30%

Realistic average-percentage-viewed for a 6–9 min video

Sources: YouTube Help — Measure key moments for audience retention · vidIQ — How to Increase Audience Retention on YouTube

The 15-vs-30-second gap, in plain terms

Treat “15 seconds” as a writing discipline you impose on yourself, not a number YouTube is checking. The metric that shows up in your own Studio dashboard, and the one worth optimizing against with real data, is the 30-second Intro figure. vidIQ’s own worked example shows a typical opening falling from 95% to roughly 65% by the 30-second mark before it flattens out — that drop is normal. The question a script has to answer is whether it flattens at 65% or at 35%.

The promise: what comes right after the hook

A hook earns attention. A promise is what you do with it — the sentence that tells a viewer specifically what they get if they stay, and by when. Kallaway calls the mechanism behind this click confirmation: your first few lines have to confirm, and ideally exceed, whatever expectation the title and thumbnail already set. His example is a Mr. Beast video titled “I Spent 100 Hours Inside the Pyramids” — showing the pyramids in the first line only meets the expectation. Adding unprecedented access to chambers that have never been explored is what beats it. That gap — reality landing slightly ahead of what the title promised — is the whole mechanism. Every line after the hook is either building that gap or wasting it.

On-screen graphic reading 'Expectation VS Reality' over a YouTube creator's video, illustrating the psychology behind script hooks
Kallaway's framing for why hooks work: a script's job is to make reality land ahead of whatever expectation the title already set.Kallaway — How To Write A Killer Script That Keeps Viewers Hooked

The 5-part hook formula

Kallaway’s channel — 435,000 subscribers, built almost entirely on breaking down what makes videos retain — packages the opening of a script into five moves, always in this order. It’s content-agnostic: the same five parts work whether you’re scripting a 90-second short or a 25-minute documentary.

1. Context

Confirm the click, immediately

State plainly what the video is about. This is where you satisfy the promise the title already made — no cold open, no “before we get started.”

Best for: The first 2–3 sentences, before anything clever

2. Common belief

Say what they already think

Restate the conventional wisdom on your subject. It signals “I understand your situation” and sets up the turn in the next beat.

Best for: Viewers who already have an opinion on your topic

3. Contrarian take

Break the belief you just stated

Contradict the common belief directly. The contrast between steps 2 and 3 is what creates the curiosity gap — you can’t build it without stating the belief first.

Best for: The line that actually earns the rest of the video

4. Proof

Why they should believe you

One concrete reason your contrarian take is trustworthy — experience, data, a result. Generic authority (“I’ve done this for years”) is the weakest version of this beat.

Best for: Credibility that's specific, not just a subscriber count

5. Plan

What they're about to get, in order

A short, ordered preview of what's coming — this is the beat that becomes your chapter list. Proof and plan can trade order, but both have to land before you move to body content.

Best for: The line that turns a hook into a promise

The tell for whether it’s working: does the first line of your script’s body — the sentence right after this five-part open — deliver something the viewer hasn’t already heard? Kallaway is blunt about the failure mode here: if that first point “repeats something they’ve already heard or is just not that compelling,” the gap between expectation and reality flips negative, and that’s the moment a viewer scrolls to the description to check the timestamps instead of watching.

How To Write A Killer Script That Keeps Viewers Hooked — Kallaway — the full breakdown of the five-part hook and the click-confirmation mechanic above.

Chapter pacing against your retention graph

Once the hook and promise are locked, a script’s remaining job is to keep paying off that opening plan without letting the retention line find a clean place to fall. Chapters are the tool for this on anything past 6–8 minutes, and the math that makes long-form worth the extra structure is exactly why: a 20-minute video has far more places for a script to lose someone than a 60-second one, and no amount of editing rescues a chapter that was never paced against attention in the first place.

The practical move is to write each chapter as a miniature version of the whole script — its own small promise at the top, its own payoff before the next chapter starts — rather than one long undifferentiated ramble with markers dropped in after the fact. Watch your own retention graph (or, before you have one, a competitor’s in the same niche) for dips at chapter boundaries specifically: a dip right where a chapter starts usually means the previous chapter didn’t close its loop before opening a new one, and the viewer took the transition as a natural exit. A spike mid-chapter, by contrast, is data — it’s the exact sentence or beat worth rewriting your next hook around. Before you have your own graph to study, borrowing one is legitimate research: pulling the transcript of a competitor's highest-retention video and mapping its chapter breaks against its own retention curve tells you what worked in your niche before you've published anything.

Reading spikes and dips like YouTube defines them

Straight from YouTube’s own help documentation: spikes “appear when more viewers are watching, rewatching, or sharing those parts of your video,” and dips “mean that viewers are abandoning or skipping at that specific part.” Segments with almost no drop-off at all are labeled Top moments — and that report only populates once a video has at least 100 views and runs 60 seconds or longer.

Open loops

An open loop is a question or tension you raise on purpose and deliberately leave unresolved, because an unresolved thing is cognitively harder to walk away from than a resolved one. The underlying idea has a name outside of YouTube entirely — the Zeigarnik effect, the psychological finding that incomplete tasks stick in memory longer than finished ones. Copyblogger’s classic explainer puts it in terms anyone who’s worked a restaurant will recognize: waiters “can easily remember the orders of each of the tables they’re serving — until the food comes out, at which point retention and recall diminishes greatly.” An unclosed order stays loaded in memory. A closed one evaporates.

Kallaway calls the same mechanic a curiosity loop and applies it at the title level, before the script even starts: “does the title alone get the viewer curious enough about the pain or problem that they cannot resist clicking?” A script extends that same trick scene by scene — open a loop in chapter one (“there’s one detail about this that changes everything, and I’ll get to it”), let two chapters of real content go by, then close it. The mistake is opening more loops than you close. A script with five open threads and one payoff reads as scattered, not suspenseful — every loop you open is a debt the ending has to pay off in full, or the whole device reads as clickbait in retrospect.

How to write a YouTube script: the complete copyable template

Everything above compresses into one shape. Copy the block below into a doc and write directly into the brackets — it holds for anything from a 4-minute explainer to a 25-minute long-form video; only the number of chapters changes.

youtube-script-template.txtcopy this
[HOOK — 0:00][Context: what this video is about, one plain sentence]
[HOOK — cont.][Common belief: what people already assume about this]
[HOOK — cont.][Contrarian take: the specific way that belief is wrong or incomplete]
[HOOK — cont.][Proof: one concrete, specific reason to trust this take]
[PROMISE — ~0:20][Plan: 3–5 things you're about to show them, in the order you'll show them]
— — — — — — — — — — — — — — — — — — — — — — — — — — — — —
[CHAPTER 1][Open with the strongest point you have — not a recap of the promise, the first real payoff]
[ → open loop][Optional: tease a detail you'll resolve 2+ chapters from now]
[ → close][One sentence that resolves this chapter before the next begins]
[CHAPTER 2][Repeat: payoff → optional loop → close]
[CHAPTER N][Repeat for as many chapters as the topic earns]
— — — — — — — — — — — — — — — — — — — — — — — — — — — — —
[LOOP CLOSE][Resolve every open loop from earlier chapters — none left dangling]
[RECAP][One line per chapter, in the order delivered]
[OUTRO / CTA][What you want them to do next — one ask, not three]

Before you decide whether to write this by hand or lean on AI-assisted drafting, it’s worth knowing what actually separates a usable youtube script generator from one that produces generic filler — the checklist there covers what to check before trusting a tool with a 20-minute video. And once a script like the one above is locked, the words still have to become a video: Longform Studio’s workspace takes a finished script through storyboard, consistent AI stills, and ElevenLabs narration with measured timing — but as covered in how script-to-video AI actually handles a finished script, no rendering pipeline fixes a script that was already losing people at chapter two.

What changes between a weak opening and a click-confirmed one
BeatWeak versionClick-confirmed version
First line“Hey guys, welcome back to the channel, today we're talking about...”“Today we're talking about [topic] — the reason [specific outcome] keeps happening.”
Credibility“I've been doing this for years.”“I've analyzed [specific number] of [specific thing] and they all do this one thing.”
The planLeft implicit — viewer has to trust it'll be worth itNamed explicitly as 3–5 steps, so the chapter list is already set

A filled-in YouTube script example

Here’s the template above with real words in it — an original excerpt for a long-form explainer on why container ships keep getting stuck, written the way you’d script it for narration rather than for reading on camera.

youtube-script-template.txtcopy this
[HOOK — 0:00]Today we’re talking about why massive cargo ships keep running aground in narrow channels.
[HOOK — cont.]The common explanation is always pilot error — someone miscalculated a turn.
[HOOK — cont.]But in most of the incident reports, the ship’s own hull was working against the pilot before anyone touched the controls.
[HOOK — cont.]I read through six official grounding investigations to find the pattern — and it’s the same physics term in every one.
[PROMISE — ~0:20]By the end: the physics term, the one channel design flaw that makes it worse, and the fix ports are quietly building around it right now.

Notice what the promise did: it converted straight into a three-chapter outline (the physics term, the design flaw, the fix) without any extra planning step. That’s the payoff of writing the plan beat honestly — chapters stop being a separate decision.

The script is the hard part. We handle what comes after it.

Longform Studio takes a finished script through storyboard, consistent AI characters, per-scene regeneration, and ElevenLabs narration with measured timing — synced and rendered without you touching an editing timeline.

Try Longform Studio

$49/mo plus transparent per-dollar production costs. No markup on provider usage.

FAQ

How long should a YouTube script be?

Length should follow the topic and your own speaking pace, not a fixed target — the words-per-minute math for converting a target runtime into a word count lives in the companion guide, how to write a video script.

What's the difference between a hook and a promise?

The hook earns the click and confirms it — context, common belief, contrarian take, and proof. The promise is the next beat: a specific, ordered list of what the viewer gets if they stay, which usually becomes the chapter list.

How many open loops should one script have?

Fewer than you're tempted to open. One or two unresolved threads across a long-form video is enough to pull someone forward; five open loops with one payoff reads as scattered rather than suspenseful, and every loop you open has to close by the end.

Do YouTube chapters affect retention, or just navigation?

Chapters themselves are a navigation feature, not a ranking input — but writing a script in chapter-sized units, each with its own mini hook and payoff, is what keeps the retention line from finding a clean exit at a chapter boundary.

Should I write a full script or work from bullet points?

Bullet points make the hook and promise beats weaker, because the exact wording of the first 30 seconds is doing real work — click confirmation depends on specific phrasing, not just topic coverage. Bullets are fine for the body of a chapter once the opening is scripted word for word.