How to Write a Video Script

A video script pairs what the viewer hears with what the viewer sees, moment by moment, and says how long each moment lasts. That is the whole definition, and it is what separates a video script from a screenplay. A screenplay is written for a director and a cast: scene headings, action, dialogue, one column, no runtime. A video script is written for a producer, a client, a voice artist, and an editor, and it carries the pictures and the clock so that all four can work from the same page. If you are writing a short film or a feature, you want our guide to how to write a script instead. Everything below is for commercials, explainers, brand films, corporate and training video, and social.

Five elements do the work. The audio: voiceover, dialogue, music, and sound cues. The visual: what is on screen while the audio plays. The runtime: how long each scene lasts and how long the whole thing runs. The beats: the order the video moves through, from hook to call to action. And the one action the viewer should take at the end. In a two-column AV script the first two are the columns, the third is counted from the audio, and the beats are the rows. The steps below build the script in that order, and the order is most of the method.

Free-form is a single column of prose or dialogue, sometimes with headings for sections. It suits scripts where the words are the whole job and the pictures follow later: a talking-head piece to camera, an interview outline, a YouTube video you will record and cut yourself, a podcast.

Two-column AV is a table. Audio on one side, visual on the other, one row per scene, with a scene number and a running time for each row. AV stands for audio visual. It is the standard for commercials, explainers, brand films, corporate and training video, and animation, because four different people can work from the same page: the producer budgets from it, the client approves it, the voice artist records from the audio column, and the editor cuts to the visual column.

The rule of thumb: if anyone other than you will approve the script or shoot from it, write it in AV. The conventions are simple. One row per scene. The audio column word for word as it will be recorded, with music and sound cues in brackets. The visual column literal: the shot, the action, the on-screen text written out as a SUPER. Keep the columns in sync, so row four of the audio always plays over row four of the visuals. The video script template is the AV format as a working editor, free and with no signup, and the rest of this guide assumes you are writing in it.

There is a third format, the screenplay, and it is a different job. Scene headings, action lines, and centred dialogue on one column, written for a director and a cast. If you are writing a short film or a feature you want that guide instead of this one.

Now go back to row one and fill the visual column against the audio. Write it for the director and the editor, which means literal: the shot, the action, the on-screen text. A warm lifestyle montage is not a visual. A close-up of a kettle switch clicking on is a visual.

• One row, one moment. Usually one shot, sometimes a short run of cuts that reads as one idea. If a row needs a paragraph to describe, split the row.
• Show what the audio cannot say. If the visual repeats the words, the row is doing one job twice. Cut one of them or change the picture.
• Write supers exactly. On-screen text goes in the visual column as SUPER: with the precise wording, because the client will approve those words too.
• Give the mandatories their own rows. Product shot, logo, offer, legal. Placed deliberately, not stacked on the end card.
• An empty visual is a note, not a gap. If a row has no visual idea yet, write TBC and flag it. That is a question for the creative team, and papering over it with a generic shot only hides the question until the storyboard.

There is a practical reason to be this literal. Each row of an AV script becomes one storyboard frame in step eight, and the frame can only be as clear as the sentence it is drawn from.

On most client projects, writing the script takes an afternoon and approving it takes the rest of the week. Rounds go back and forth as attachments, comments arrive as an email that says paragraph three feels a bit long, and by round three nobody is sure which version has the change from round one. The way through is procedural rather than creative.

• Send one link, not an attachment. Everybody comments on the same document, and there is never a question about which file is current.
• Comments on the line. A note against row four of the audio is actionable. A note about the middle bit is not.
• A version per round. Each round of changes is its own version, so a change from round one cannot quietly disappear in round three, and you can show exactly what moved between them.
• Explicit approval on the final version. Looks good in an email is not sign-off. Ask for the approval to be recorded against the version, then lock it.

The reason to be strict is what happens next. The approved script is what gets recorded, storyboarded, and shot. Written sign-off is what stops a rewrite arriving on the day. This is what Scripts in Boords is built for: the client comments on the line through one link with no login, every round is a version, and the approval is on record. More on client review in Boords.