Hook, proof, promise: a three-beat structure for the first six seconds

The first six seconds of a UA video or playable carry more weight than the other twenty-four combined, because that is the window before a viewer decides to keep watching or swipe past. Most openers spend it badly: a logo, a slow pan, a line of copy nobody reads. One structure holds up across genres and formats better than the alternatives, and it goes hook, proof, promise, in that order, roughly two seconds a beat.

The hook is a claim, not a scene

A hook states something specific enough that a viewer either believes it or doubts it on the spot. "This board has never been solved by anyone watching" is a hook. A wide shot of a castle is not, however nice it looks, because it asks nothing of the viewer and offers nothing to disagree with. Mechanical claims work best: a board state that looks unsolvable, a resource count that looks impossible to hit, a timer that looks too short. That holds for match-3 and merge creative, and for strategy too. The claim does not need to be true for the whole game. It needs to be true for the following four seconds.

Proof closes the gap the hook opened

Once the hook has made a claim, the viewer's next question is whether it holds up, and proof answers that with action rather than assertion. If the hook implied an impossible board, proof shows a move that starts solving it. If the hook implied a resource shortfall, proof shows the player finding an unexpected source. This beat should carry the first real interaction in a playable, or the first real gameplay footage in a video, because proof that gets narrated rather than shown reads as a promise, and mixing up the two beats is the most common failure in openers that otherwise have a strong hook.

Promise is the door left open

The final beat tells the viewer what continuing gets them. Make it one concrete thing rather than a list. "Ten more boards like this one" is a promise. "Endless fun and exciting levels" is not, because it describes a feeling instead of an outcome. A promise that names a specific, repeatable unit of gameplay, a level, a boss, a build, gives the viewer a reason to tap that survives past the install screen, which matters because a hook that overpromises and a promise that underdelivers both show up later as day-one churn rather than only as a weak playable.

Testing the structure honestly

Cutting an opener to this beat structure does not guarantee performance, and hook, proof, promise is a diagnostic rather than a formula that replaces testing. What it does is give a creative team a shared vocabulary for picking apart a weak opener before it goes into a test slate: if click-through is low, ask which beat is missing rather than reshooting the whole first six seconds. Most underperforming openers are missing proof. Proof is the beat that requires actual gameplay to exist before the edit starts, and it is the one most often cut for time.

Building the first six seconds around three beats rather than one continuous scene also makes iteration cheaper. A team that has isolated the hook, then the proof, then the promise as separate shots can swap one beat against a new concept without recutting the whole opener, which turns creative testing from a full-remake exercise into a beat-level one.

Related archive reading

These articles provide related context and remain subject to their stated review status.

• Video length: what 6, 15 and 30 seconds each do
• The first two seconds are a thumbnail problem