The First 3 Seconds: Writing Hooks That Make People Stop Scrolling
Written byNereida Enriquez • Communications Associate
Video - November 2025
Every video your brand posts is really two pieces of content. There is the video you spent a week planning, shooting, and editing. Then there is the three-second clip at the front of it that decides whether anyone ever sees the rest. Most brands pour their budget into the first one and improvise the second, which is roughly like catering a dinner party and forgetting to send the invitations.
The numbers back this up in a way that should make any marketing director uncomfortable. TikTok for Business has reported that 63% of its highest click-through videos hook viewers within the first three seconds, and roughly 71% of viewers decide within the first few seconds whether a video is worth continuing. The mechanism matters more than the statistic. When someone swipes away at second two, you have not merely lost one viewer. You have told the algorithm that your content failed its first test, and the algorithm responds by declining to show it to anyone else.
We see the same three openings constantly when we audit a brand's back catalog.
The first is the logo animation. Someone in the organization spent real money on a three-second brand sting, and now it plays at the front of every asset. On a website, that reads as polish. In a feed, it reads as an advertisement, and people have been trained since childhood to skip advertisements. Your logo has never once stopped a scroll.
The second is the warm-up. "Hey everyone, so today we wanted to talk about our new spring collection." By the time the sentence lands, the viewer is two videos away. Slow build-ups where the first three to five seconds go to pleasantries or context are among the most common retention killers.
The third is the vague tease. "This changed everything." "You have to see this." These feel like hooks because they mimic the rhythm of one, but they promise nothing specific, so they generate no genuine curiosity. A viewer has to be able to tell, instantly, what they will get by staying.
A hook has one job: create a gap between what the viewer knows and what they want to know, and make that gap feel worth three more seconds of their life. There are a handful of reliable ways to do it.
Lead with the payoff, then explain it. Instead of building toward your result, open on it. A skincare brand should not spend four seconds on the bottle before the reveal. Show the reveal, then earn it. This inverts how most people were taught to structure a story, and it is the single highest-leverage change most brands can make.
Make a specific claim you immediately prove. Broad claims read as noise. Specific ones create obligation. "This posting habit quietly killed my reach last year" works because it names a real consequence and implies you are about to identify the habit. Follow through fast, because a hook that overpromises and underdelivers tanks completion, and completion is the metric that decides whether the algorithm expands your reach.
Interrupt the visual pattern. The feed has a rhythm, and the brain scrolls on autopilot until something violates the pattern. Sudden movement, an unexpected cut, a bold text callout, or even an abrupt moment of silence jolts a viewer out of that autopilot. As feeds have gotten louder, the reverse has started working too. A slower, quieter opening in a chaotic feed can function as its own kind of interrupt.
Design for silence. A large share of your audience is watching with the sound off. Roughly half of silent viewers depend on captions to follow the content at all. If your hook lives entirely in the audio, it does not exist for those people. Put the promise on screen in text, in the first frame, in a font someone can read at arm's length.
Writing one great hook is a creative act. Writing a great hook for every video, every week, forever, is an operational problem, and that is where most in-house teams break down.
The most useful practice we have adopted is to separate hook production from video production entirely. Shoot the body of your video once. Then shoot three to five different openings for it. Same content, same edit, different first three seconds. Post them as separate variants, or run them as paid tests, and let the data tell you which framing your audience responds to.
This does two things. It removes the guesswork, because your instincts about what will hook your specific audience are probably wrong at least half the time, and it produces a hook library. After twenty tests you stop guessing entirely, because you know that your audience responds to blunt cost claims and ignores aspirational lifestyle framing, or whatever the pattern turns out to be for your category.
The second discipline is keeping a real log. Date, platform, hook text, first frame, and the outcome: three-second hold rate, average view duration, click-through. Reading the retention curve tells you where the problem lives. A dip in the first two seconds means the promise was weak. A cliff in the middle means the body has dead air. Most brands look at total views, which tells them almost nothing they can act on.
Two honest cautions, because the hook discourse online tends toward the evangelical.
A hook cannot rescue content that has nothing to say. If your video is a product feature in search of an audience, a clever opening will get more people to arrive at the moment they realize you have wasted their time, which trains them to skip you next time. The hook buys attention. The content has to justify it.
And chasing hooks can quietly deform your brand. If every video opens with a provocation or a shock cut because that is what tests well, you eventually become a brand made of provocations, which is exhausting and hard to trust. The strongest accounts we work with hold a consistent voice and let the hook vary within it, rather than the reverse.
Take your five most recent videos and watch only the first three seconds of each with the sound off. Ask whether a stranger would know, from those three seconds alone, what they are being offered and why they should care. If the answer is no on more than one, that is your entire content problem, and it is a fixable one.
Then take your best-performing video from the last quarter and cut three new openings for it. Repost it. Watch what happens. The video you already made is probably better than its performance suggests. It just never got past the door.
Nereida is a Communications Associate at Proach Media, covering platform strategy, short-form video, and the creative tooling shifting how content gets made.