Hooks & Openings · Data-Driven
The YouTube Hook Formula: A 3-Second Retention Predictor
I spent four months analyzing how 1,200 YouTube scripts open. The first 3 seconds predict 60% of whether someone stays or leaves. Not engagement, not topic, not production quality — just the opening. Here is what separates the hooks that work from the ones that lose people before the video starts.
The Formula Is Three Parts. Most Creators Use One.
[Pattern Interrupt] + [Promise] + [Curiosity Gap].
That is the formula. It is stupidly simple and somehow still rare. I see creators who nail the promise but forget to interrupt the scroll pattern. Or they create a curiosity gap but never actually promise anything. Either mistake tanks retention.
Here is how each piece works. The pattern interrupt is the moment you break whatever the viewer was expecting. It could be a number that does not make sense, a visual that contradicts the title, a statement that sounds wrong. The point is to make them stop — even for half a second — because autopilot scrolling is your real enemy.
The promise is what you are going to deliver. It should be specific and concrete. Not "today we are talking about hooks" but "here are the 3 hook types that outperform the rest by 40%." If your promise is vague, the viewer has no reason to stay.
The curiosity gap is the space between what the viewer knows now and what they will know after watching. You create it by hinting at something without revealing it, or by showing a result without explaining how. The human brain cannot leave a curiosity gap open — it has to close it. That compulsion is most of what keeps people watching past 10 seconds.
The Astryx Hook Framework: 5 Types Ranked by Data
We classified every hook in our 1,200-script dataset into one of 5 types and measured 30-second retention for each. The results are not what most creators expect.
| Hook Type | Retention at 30s |
|---|---|
| Contrarian — states something that contradicts common belief | 78% |
| Shock — reveals an unexpected fact or visual | 72% |
| Question — asks something the viewer cannot answer | 64% |
| Tease — promises a reveal or payoff later | 56% |
| Relatable — "if you have ever felt X..." | 41% |
Relatable hooks are the most common and the worst-performing. Creators gravitate toward them because they feel natural. But the data says people do not scroll YouTube to feel understood — they scroll to learn something they did not know, see something they did not expect, or have their assumptions challenged.
How to Build Each Hook Type
The formula — Pattern Interrupt, Promise, Curiosity Gap — works across all 5 types. But the execution changes. Here is how to build the top 3.
Contrarian Hook (78% retention)
Pattern interrupt: state the opposite of what everyone believes. Promise: I will prove it with data. Curiosity gap: why has nobody said this before?
Example: "Most people think YouTube growth is about posting more. They are wrong. In the next 3 minutes I will show you why the channels that grow fastest post the least."
Shock Hook (72% retention)
Pattern interrupt: a specific, surprising number or fact. Promise: here is what that means for you. Curiosity gap: what else does the data hide?
Example: "Your first 12 words on camera predict 47% of your retention. Not your thumbnail, not your topic. Twelve words. I pulled this from 1,200 scripts and it changes how you should open every video."
Question Hook (64% retention)
Pattern interrupt: ask a question the viewer has never considered. Promise: I have the surprising answer. Curiosity gap: how could the answer be different from what I assume?
Example: "Why do the best YouTube hooks actually perform worse than average ones? Let me explain — because this completely changed how I script my openings."
The Hook-Type-Niche Matrix
Not every hook type works in every niche. Our data shows clear patterns. Tech review channels do best with Shock hooks (specific spec reveals, benchmark numbers). Education channels thrive on Contrarian hooks (challenging common beliefs). Entertainment channels need Tease hooks (promising a payoff later). Finance channels cannot use Contrarian hooks effectively — audiences there trust consensus, not challenge.
The worst thing you can do is pick a hook type that clashes with your niche's expectations. A Relatable hook in a tutorial video works fine (2% drop from average). A Relatable hook in a financial analysis video does not (11% drop). The audience tunes into finance content for authority, not empathy.
The Mobile Factor
Here is something that gets overlooked constantly: over 70% of YouTube watch time is mobile. On mobile, the hook window shrinks from 5 seconds to about 2. The pattern interrupt has to land in those 2 seconds or the viewer scrolls.
This means your pattern interrupt cannot depend on audio. Many viewers scroll with sound off. If your hook is a sentence, they never hear it. The strongest hooks work visually first and verbally second. A before/after split screen that does not need explanation. A number on screen that contradicts the title. Text that appears before anyone speaks. Optimize for mute first.
Next Steps
Now that you know the formula, dig into what goes wrong and how intros work in practice:
Want to score your hook before recording?
Astryx analyzes your script against real retention data from 1,200+ scripts. It grades your hook, predicts retention at 30 seconds, and shows exactly what to fix. Free tier, no credit card.
Try Astryx Free →