Playbook · 12 July 2026
The first three seconds are a promise, not a trick
A hook that lies buys you a view and costs you a viewer. Here is what we found when we pulled apart the openings of our own clips — the ones people watched to the end, and the ones that spiked and died.
Everyone knows the rule: you have three seconds. What nobody says out loud is what those three seconds are for. Most people treat them as a toll — pay the algorithm, get the view — and so they open with whatever is loudest in the timeline, regardless of whether the rest of the clip has anything to do with it.
That works exactly once per viewer. The second time someone is pulled in by a bang and handed twenty seconds of something else, they stop trusting the account. They keep scrolling past you specifically, which is the most expensive outcome in short form and the one nobody measures.
What we actually looked at
We went back through our own library — a few hundred clips, all of them ours, all of them published with nothing riding on the result — and sorted them not by views but by what happened at second twelve. Views tell you the opening worked on the feed. Second twelve tells you the clip worked on a person.
Two shapes came out of it. The first: a steep opening spike, then a cliff — half the audience gone before the shot had resolved. The second: a slightly softer opening, a curve that flattens, and a tail that survives well past the halfway mark. The second shape almost always did better over a week, and it did it without a louder first frame.
The clips in the second group had one thing in common, and it was not editing rhythm or music or captions. It was that their first shot was a small, honest version of their last shot.
The promise, stated plainly
Think of the opening as a contract with one clause: stay, and you will get this. The viewer reads it in under a second and decides. If the clause is specific and the clip delivers it, you have bought their attention honestly and you can spend it. If the clause is vague — a bang, a face, a caption promising “wait for it” — you have borrowed attention at a very high interest rate.
The practical version of this is almost boring: show the subject in the first frame. Not the whole payoff, not the best moment — the subject. A pan on a flame. A spider on a leaf. A road with one car on it. The viewer immediately knows what kind of thirty seconds they are being offered, and the ones who want it stay for all of it.
A hook is not a trap. It is the shortest true sentence about what happens next.
Five openings that keep their word
Three that break it
“Wait for the end.” This asks for credit before you have earned any. It also tells the viewer that the first twenty seconds are filler, which they then correctly skip.
The borrowed bang. An unrelated loud moment cut onto the front. It reliably lifts the view count and reliably wrecks retention, and the second number is the one that decides whether you are shown again next week.
The logo. Three seconds of brand animation is three seconds of somebody else's scroll. Put the mark at the end, where a viewer who stayed will actually read it.
Write it before you shoot it
The reason hooks get bolted on in the edit is that nobody wrote one on the day. Our shot lists now carry the opening frame as a line item, described in words, before anything is packed: flame catching the edge of the pan, hard from the left, before the vegetables land. If we cannot write that sentence, we do not have a clip yet — we have footage.
It also saves money. Knowing the first frame tells you what to light, where to put the camera, and which take you actually need. Most of the expensive parts of a shoot are decisions that could have been made in a notebook.
Everything above came out of clips we made for ourselves, where a bad opening cost us nothing but our own afternoon. That is the cheapest place to learn this, and it is why the library exists at all.
Read next