Most "scroll-stopping" videos die in the first second, not because the idea was bad, but because the team fell in love with it in the edit suite. The room laughs at the clever transition. The brand manager nods at the elegant build-up. Everyone forgets that on the feed, nobody is in the room. They're on a sofa, half-watching, thumb cocked, deciding in less time than it takes to read this sentence whether you exist.

You get about three seconds. Realistically less. In that window the viewer isn't evaluating your production values or your tagline. They're answering one question: is this for me, right now? Everything that travels on the feed is built around that question. Everything that flops is built around something you noticed only because you watched it forty times.

Attention is a default to "no"

The feed is a rejection machine. Its job is to keep someone moving, and it's very good at it. So the honest starting assumption is that the viewer is actively trying to leave, and your video has to give them a reason not to before it gives them anything else.

This reframes the whole opening. You're not "setting up" a payoff. You're not "establishing" anything. Setup is a tax you charge the viewer before they've agreed to pay. On a paused, considered medium — a cinema, a landing page they chose to visit — setup earns interest. On the feed it just loses you people while you clear your throat.

So the rule is blunt: open on the thing you were saving for the middle. The conflict. The payoff. The unresolved question. The result, then the method. Whatever the most charged moment of your video is, that's frame one. You can rebuild context afterward, because now they've agreed to stay for it.

Why "clever" rarely travels

Clever is the most expensive trap in social, because it feels like quality. Clever is satisfying to make. It survives internal review. It wins the pitch. And it dies in the feed, because clever almost always depends on context the viewer doesn't have yet.

The pun lands only if you've read the line above it. The twist works only if you watched the setup. The visual metaphor is gorgeous only if you already know what it's a metaphor for. Clever is a reward for attention you have not yet been given. You're handing out the prize before anyone entered.

If your hook requires the viewer to have watched the hook, it isn't a hook. It's a punchline missing its joke.

What travels is rarely clever. It's clear, specific, and slightly uncomfortable. A flat statement that makes someone think "wait, is that true?" A visible stakes. A face mid-reaction. A claim a chunk of your audience will instinctively want to argue with. None of that needs a setup, because the tension is already in the frame.

Hook patterns that actually do the work

Frameworks beat inspiration here, because you'll write a hundred of these and inspiration runs out around the fourth. A few patterns that reliably survive the feed:

  • Result first, method later. Show the finished thing, the before/after, the number, the reaction — then explain how. Curiosity about "how" is durable. Curiosity about "what" expires in a second.
  • Name the viewer. "If you're opening a company in Bahrain and you've been quoted three different prices…" The specificity does the targeting. The wrong people leave, which is fine. The right people lean in, which is the entire point.
  • Open mid-conflict. Start inside the argument, the mistake, the awkward moment. No "so today I want to talk about." Drop them at the part they'd have stayed for anyway.
  • The contrarian flat claim. State something a sensible person would push back on. "Most of your ad budget is buying you nothing." Disagreement is engagement that hasn't decided yet.
  • The visible question. Pose the question on screen and in the first line, then make them wait for the answer. Not a vague tease — a real question they actually want answered.

One craft note that matters more than any pattern: the first frame is a frame. The feed often autoplays muted, so your opening line should be legible without sound, and your opening image should be doing work before the audio kicks in. If your hook only lands with the voiceover, you've built a radio ad and hung it in a window.

Test it the way the feed will. Watch your first three seconds with no sound, at arm's length, and ask whether a stranger would know what they're being offered. If you have to explain why it's good, it isn't — not here. Save the clever for the people who already stayed. Spend your three seconds buying the stay.