Your audience is driving down a highway at 70 miles per hour. They're not browsing. They're not considering. They're moving, fast, on autopilot, thumb swiping with the same muscle memory they use to unlock their phone. Your hook has to make them see something compelling enough to stop, turn around, and come back the other way.
That metaphor comes from Kallaway, and it reframes what a hook actually needs to do. You're creating an interruption so strong that it overrides momentum, not crafting a careful introduction.
I've been studying hooks across two sources that together cover hundreds of examples: Kallaway's breakdown of irresistible hooks and heyDominik's analysis of 1,000 hooks. What emerged is a repeatable structure that works whether you're scripting short-form, long-form, or anything in between.
Sources: Kallaway, "How to Create Irresistible Hooks"; heyDominik, "I Studied 1,000 Hooks"
Your Viewer Is Mid-Scroll With Their Thumb Already Moving
Most creators write hooks as if the viewer is sitting in a chair, waiting, ready to be convinced. That's not the situation. The situation is a person mid-scroll, thumb already in motion, brain running a background process that filters out 95% of what passes by.
Kallaway describes it as the highway problem. At 70 mph, you don't stop for something mildly interesting. You stop for something that triggers an involuntary response, something your brain can't ignore even though your body is in motion.
heyDominik puts a finer point on it. People swipe on muscle memory. The thumb moves before the brain engages, and they keep swiping until something hits the specific brain receptors that produce a "hm, that is interesting" response. The swipe is the default. Stopping is the exception.
This means your hook has about 1.5 seconds to interrupt a physical habit. Writing it like an introduction, with preamble and context-building, is like putting up a small sign on the side of a highway and hoping someone reads it at speed.
Step 1: Context Lean-In, Common Ground, Benefit, Metaphor, or Mind-Blow
The first line of your hook has two jobs, and both are non-negotiable.
Job one: make it immediately clear what this video is about. Topic clarity is how viewers self-select. If someone can't tell within the first line whether this content is relevant to them, they won't stick around to find out. This sounds basic, but a surprising number of hooks fail here. They open with something clever or mysterious and the viewer has no idea what the video covers, so they leave.
Job two: create a lean-in. Once the viewer knows the topic, they need a reason to care about your specific take on it. Kallaway identifies four lean-in strategies:
- Common ground. Reference a shared experience or frustration that makes the viewer feel seen. "You've probably noticed that your best-performing videos have nothing in common with each other" immediately creates a sense of "yes, I've noticed that too."
- Benefit or pain point. Name the specific outcome the viewer wants or the specific problem they're experiencing. Be precise. "Get more views" is weak. "Stop spending 4 hours editing videos that get 200 views" is a pain point people physically feel.
- Metaphor. Take a complex idea and make it simple through comparison. Kallaway's own highway metaphor does this. The viewer didn't know they needed a new way to think about hooks until the metaphor made the problem vivid.
- Mind-blow. Tell them something that recalibrates what they thought they knew. A fact, a number, a counterintuitive claim. "The tech in the Vegas Sphere is insane. Biggest screen ever built, 20 times bigger than an IMAX." That opening from a video with 8 million views works because the number (20 times bigger than an IMAX) is so far outside expectation that the brain pauses.
The key insight is that Step 1 is doing two things simultaneously, not sequentially. You don't establish the topic and then create the lean-in. You do both in the same sentence or two. The Sphere example is a masterclass in this: the topic is clear (Sphere technology) and the lean-in is immediate (a number that blows your mind).
Step 2: The Scroll-Stop Interjection, One Word That Stuns
After the context lean-in, you drop a single contrasting line. Kallaway calls this the scroll-stop interjection, and it works like a stun gun.
The word is usually a conjunction like "but," "however," "yet," or "although," a single word that signals everything is about to change direction.
Why does this work? Because the first step set an expectation. The viewer's brain built a model of where this was going, and the interjection breaks that model. The brain had a prediction, now that prediction is wrong, and that creates an involuntary spike of attention.
Using the Sphere example: "The tech in the Vegas Sphere is insane. Biggest screen ever built, 20 times bigger than an IMAX." The viewer thinks they know what's coming, a breakdown of the specs, a tour of the screen. Then: "But..."
That single word does more work than the entire context lean-in. It signals that whatever comes next will be unexpected, and unexpected is the one thing the scrolling brain can't ignore.
The interjection doesn't need to be long. Think of it as a beat, a pause, a pivot, and then the setup for the real punch. If Step 1 gets their attention and Step 2 holds them still, Step 3 is what makes it impossible to leave.
Step 3: The Haymaker, The Curiosity Bomb
This is where you detonate the curiosity bomb. The line after the interjection has to make leaving feel like a loss.
Kallaway calls it the haymaker because it carries all the force. Everything before it was positioning. The context lean-in got the viewer's feet planted, the scroll-stop interjection froze them, and now the haymaker hits.
What makes a good haymaker? It opens a curiosity loop so deep that the viewer's brain genuinely needs the answer. The deeper the loop, the stronger the hook. "But the part nobody is talking about will change how you think about screens forever" is a curiosity bomb because it implies knowledge the viewer doesn't have and frames that knowledge as transformative.
Kallaway describes the ideal result as a domino effect: the viewer watches line 1 and feels so compelled they can't dream of doing anything else but watching line 2, and line 2 pushes them into line 3, and so on. Each line makes the next line feel inevitable.
The three steps together form a sequence that maps to how the brain processes interruptions. The context lean-in matches the viewer's existing mental model (recognition), the interjection breaks that model (surprise), and the haymaker offers a new, better model but withholds the full picture (curiosity). Recognition, surprise, curiosity. That's the sequence.
Autopilot Swiping Triggers a Specific Neurological Response
heyDominik's research adds a neurological layer to Kallaway's structural framework. After studying 1,000 hooks, his central finding is about autopilot behavior.
When someone opens a social feed, their brain shifts into a low-engagement scanning mode. The thumb swipes at a consistent rhythm, and the brain processes each piece of content just enough to decide "ignore" or "engage." Most content gets processed and dismissed in under a second.
The "hm, interesting" trigger is heyDominik's term for the moment when something gets past the filter. The brain recognizes that this particular piece of content is worth conscious attention. The trigger doesn't require agreement or excitement, it requires novelty, specifically something the brain's prediction engine didn't expect.
This explains why the three-step structure works at a biological level. The context lean-in passes the first filter (relevance), the interjection triggers the "hm" response (pattern break), and the haymaker converts that moment of attention into sustained engagement (unresolved curiosity).
It also explains why so many hooks fail. A hook that opens with "In this video, I'm going to show you..." doesn't trigger the "hm" response because there's no pattern break. The brain predicted generic content and received generic content, so the thumb keeps moving.
One practical finding from heyDominik: context-first hooks dramatically outperform hooks without context. Adding a text overlay at the top of the screen that tells viewers what the video is about fixes broken hooks almost instantly. People don't care about your clever opening if they don't understand the context. They need to self-select before they can lean in.
Open Loops: Why You Must Never Close All Your Questions
A hook creates a question in the viewer's mind. The entire body of the video exists to delay answering that question as long as possible while keeping the promise alive.
heyDominik's framework for this is straightforward: open loops and don't close them as long as you can. When you do close a loop, immediately open another one.
Think of it like a SaaS product's onboarding flow. Each step resolves one question but introduces the next. "Great, you've set up your account. Now let's configure your first workflow." The user never reaches a state where all their questions are answered and they feel done. There's always a next thing pulling them forward.
The same principle applies to video. If your hook opens a loop ("the part nobody is talking about") and you answer it at the 30-second mark, the viewer has no reason to keep watching. But if you answer it at the 30-second mark while simultaneously opening a new loop ("and this connects to something even bigger that I'll show you in a moment"), you've maintained the tension.
The practical implication for scripting is that every section of your script should be audited for loops. Where do loops open? Where do they close? Is there always at least one open loop pulling the viewer forward? If you find a moment where all loops are closed, you've found a drop-off point.
Swipe-Back Behavior Sends a Stronger Signal Than a Passive Stop
heyDominik identified a behavior pattern that probably matters more than most creators realize: the swipe-back.
A viewer swipes past your video. Their thumb has already moved. But their brain, which processes visual and textual information slightly slower than the thumb moves, catches something interesting a beat too late, so they swipe back.
This swipe-back behavior appears to signal much stronger viewer interest. It indicates that the content was compelling enough to override the autopilot swipe, which is a stronger signal than simply stopping on a video that happened to appear at the right moment.
The implication for hook writing is that your hook needs to work even when the viewer has already started to leave. The visual element, the first few words, the text overlay, all of these need to be processed quickly enough that the brain catches them before the thumb completes the next swipe. This is why bold, clear, context-first hooks outperform subtle, clever ones. Subtlety doesn't travel at swipe speed.
heyDominik illustrated the fine line with a memorable example: someone in a hardware store saying "I never pull out," which sounds shocking until you realize he's talking about screws. The point is that there's a narrow space between being genuinely attention-grabbing and being misleading or clickbait. The best hooks live in that narrow space, surprising enough to trigger the "hm" response but honest enough that the payoff matches the promise.
For teams producing multiple videos per week, all of this becomes a systems problem. You can't rely on one person having a flash of inspiration for every hook. The three-step formula (context lean-in, scroll-stop interjection, haymaker) needs to be embedded in your script template so that every writer on the team starts from the same structure. The difference between a team that produces consistently strong hooks and a team that produces occasional good ones surrounded by mediocre ones is almost always the template, not the talent.
At Clipflow, we build the three-step hook structure into the scripting phase of every video we produce. The writer fills in each step as a discrete field, which means the hook gets reviewed as a standalone unit before the rest of the script is even drafted. This catches weak hooks before they contaminate an otherwise good video.
The highway is real. Your audience is moving at 70 mph and your content is a sign on the shoulder. Three steps (context lean-in, scroll-stop, haymaker) give you a system for building signs that make people turn around. But the deeper principle is simpler than the framework. Every second your viewer spends with you is a second they chose not to spend somewhere else. The hook is your argument for why that choice is worth making, and you have about 1.5 seconds to make it.
