Watch the opening of a draft ad without sound. Read its caption on its own. Then listen without looking. Do those three readings make you expect the same immediate story?

For a UA manager reviewing a cut, this is a small check with a concrete output: three short descriptions of the promise, followed by the place the video delivers it.

Read the parts separately

Start with the picture and hide the caption. Write what appears to be happening, without adding information you already know from the brief. Then read only the caption and describe the question it creates. Finally, listen to the audio: what does it lead you to expect next?

Consider this illustrative example:

  • Picture: a hero before two doors. It sets up a choice between routes.

  • Caption: “Which door saves the hero?” It sets up the consequence of that choice.

  • Voiceover: “Build your kingdom.” It sets up a story about kingdom building.

The first two layers are connected. The voiceover introduces something else. If the next shot reveals what is behind a door, it continues the choice while leaving the kingdom building idea unresolved.

Align the question, not every word

If the ad is about choosing the door, I'd keep that as the opening's subject. “Only one way out” could add stakes in the voiceover. It says something different from the caption while remaining attached to the same situation.

Exact repetition is optional. One layer can establish the action, another can frame a question, and another can add urgency that the scene supports. The useful test is whether they fit together when the viewer sees the continuation.

Locate the payoff

Finish the note with one more line: where does the video deliver the opening's promise? Name the shot or timestamp. In the teaching example, that would be the reveal of the chosen route and its consequence.

For your own review, the note can be as short as: Picture promise / Caption promise / Audio promise / Payoff location / First mismatch.

This check makes the intended message inspectable. It does not predict retention or conversion, and it does not replace a full creative or technical review. Its value is that the next edit can address a named mismatch before the campaign test.

Try the three readings on one draft and mark the first place their expectations diverge.