How to Write a Video Script That Sounds Natural
- Difficulty
- Moderate
- You need
- An approved brief, a stopwatch, a reader
- Works with
- Ads, explainers, brand films, narrated pieces
On this page
To write a video script, start with one sentence that states the message and the action you want, choose a structure, then write two columns: what the viewer hears and what they see. Write the audio for the ear, make every visual note something a crew can shoot, and time the draft aloud before anyone approves it. This guide is part of our video production guide.
You'll need an approved brief, a stopwatch and one person from your audience who can read the draft. If you're still choosing the kind of video to make, start with our overview of types of video production.
What a video script has to do
A video script does two jobs at once. For the client and approvers, it is the exact words and moments the viewer will get, in order. For the crew, it is the first instruction set: what has to be filmed, recorded, designed and cleared.
That's why the format matters. Prose scripts work for reading, but they hide the visuals inside paragraphs, and the crew has to guess. A script that separates picture from sound shows the gaps straight away. You can see a line with no shot under it, or a shot with nothing to hear. You can also see where a graphic is doing work the narration repeats.
A good script also reflects how people take in video. Viewers can't re-read a sentence they missed, so spoken lines have to land the first time.
The two-column A/V script format
The A/V (audio/visual) format splits the page in two. Video instructions go on the left: shots, action, graphics and on-screen text. Audio goes on the right: narration, dialogue, music and sound effects. Each row is one moment, so the crew reads across to see what plays together.
Here is the opening of a 30-second product spot written that way:
| # | Video | Audio |
|---|---|---|
| 1 | CU: hand sets bottle on wet dock post, sunrise | SFX: water lapping. Music in, soft |
| 2 | WS: empty boat deck, salt crust on rails | VO: "Salt gets into everything." |
| 3 | MS: owner sprays rail, wipes once, shine | VO: "This takes it off in one pass." |
| 4 | Super: product name + "One pass." | Music lifts. VO: "One pass. Done." |
| 5 | End card: logo, URL | Music out on beat |
The numbers in the first column become references for the storyboard and the shot list. Abbreviations such as CU (close-up), MS (medium shot), WS (wide shot), VO (voiceover) and SFX (sound effects) keep the video column short.
Steps
- Write the one-sentence message and the action.
Before any lines, write one sentence that says what the viewer should remember and one that says what they should do next. Put both at the top of the script. Every row you add later has to support one of them, or it gets cut. If approvers can't agree on these two sentences, stop here. The script won't fix a disagreement about the brief.
- Pick the structure before you write lines.
Choose the shape first. Problem, solution, proof and action suits most ads. Introduction, body and conclusion suits explainers and training. A story arc with a turning point suits brand films. List the beats as rows in the table, with a rough share of the running time for each, before writing a single word of dialogue.
- Earn the first few seconds.
The opening row has to give someone a reason to keep watching, often before the sound is even on. Open on the most visual moment, the problem or a surprising result, not the logo. Write the opening so it makes sense muted, using picture and on-screen text, and then again with sound.
- Draft the audio column for the ear.
Write the way people speak. Use short sentences with one idea each, familiar words, and round numbers such as "about 6,000" rather than "6,049". Cut anything the picture already shows. Read each line to yourself as you write it. If you trip over it, a narrator will too.
- Fill the video column with shootable instructions.
Each video cell should say who or what is in frame, the shot size, and what happens. "Owner sprays rail, wipes once, shine" can be filmed. "Show how easy it is" can't. Note practical needs in brackets, such as a product sample, a location or a second camera. These notes feed the pre-production checklist.
- Mark on-screen text, graphics and supers.
Write every word that will appear on screen in the video column, exactly as it will read, including name supers (lower thirds) for speakers. Keep on-screen text short enough to read in the time the row allows. Check that it reinforces the narration rather than competing with it.
- Read it aloud and time it.
Read the audio column aloud at a natural pace with a stopwatch running. Then add time for rows with no speech, such as music beats, product moments and the end card. If the total is over the deliverable length, cut whole rows before trimming words. A rushed narrator sounds worse than a missing line.
- Test it with a reader, then lock it.
Give the script to someone who matches the audience and ask them to read it and tell you the message back. Fix jargon, gaps and anything they misread. Then get one written approval from the named approver and lock the version number. After that, changes go through a change request, not a quiet edit.
Once the script is locked, the next job is to storyboard the script. If you'd rather have scripting, shooting and editing handled together, our in-house video production team starts every project with a treatment, shot list and clear scope.
Write for captions and audio description
Accessibility is cheaper to build into a script than to add after the edit. W3C's Web Content Accessibility Guidelines ask for captions on prerecorded video, and those captions cover more than dialogue. They identify who is speaking and include sounds that matter to the meaning. Writing speaker names and key sound effects into the audio column gives the captioner a head start.
The guidelines also ask for audio description of visual information. That means narration that describes actions, scene changes and on-screen text that the main soundtrack doesn't already convey. W3C notes that no extra description is needed when the audio already carries all the important visual information. So the best fix happens at the script stage. Have the narration name the product, read out key on-screen text and say what is happening in silent product shots.
Leave space in the frame for captions, too. W3C's caption guidance says captions shouldn't obscure important information, so avoid placing supers at the bottom of the frame where captions usually sit. Following these guidelines is good practice, not a legal guarantee of compliance.
Troubleshooting
The script runs long
Cause: the script is trying to say too many things, or the timing was estimated from word count instead of read aloud. Rows with no narration, such as product reveals, music beats and the end card, still take screen time and are easy to forget.
Fix: go back to the one-sentence message at the top. Mark each row as essential, helpful or nice to have, and cut every nice-to-have row first. Then merge rows that make the same point. Only after that, trim words within lines. Re-time the whole script aloud after each pass, including the silent rows, until it fits with a little room to spare for the edit.
The voiceover repeats what the picture already shows
Cause: the audio column was written first, as if it were an article, and the visuals were added afterwards to illustrate it. The result is narration that describes the picture ("Here you can see...") instead of adding what the picture can't show.
Fix: read across each row and ask what the viewer would miss with the sound off, and what they'd miss with the picture off. Let the picture carry what is visible and the voice carry reasons, results and feelings. Keep descriptive narration only where it does an accessibility job, such as naming the product or reading out key on-screen text.
Stakeholders keep rewriting after approval
Cause: there was no single named approver, or the approval was informal, so every new reviewer feels free to reopen the script. Late changes often come from people who weren't asked at the brief stage.
Fix: name one approver for the script in the brief, and list who else gets to comment and by when. Collect all comments in one round, resolve conflicts before sending a new draft, and lock the approved version with a number and date. After lock, log each change request with its effect on the shoot and the schedule, so the cost of a late rewrite is visible before anyone agrees to it.
Common questions
How long should a video script be?
As long as the video, read aloud at a natural pace, and no longer. Word counts are a rough guide only, because pauses, music beats and visual-only moments take time without adding words. The reliable method is to read the audio column aloud with a stopwatch, add the silent moments from the video column, and cut until the total fits the deliverable length.
Should I write the script before or after the storyboard?
Write the script first. The storyboard shows how each line and beat will look, so it needs locked words and timing to work from. A rough sketch can help while you write a visual-heavy piece, such as a product ad. But boarding a script that is still changing means redrawing panels every time a line moves.
Do interview and testimonial videos need a script?
They need a plan rather than a word-for-word script. Write the message the video must land, an outline of the story beats, and the questions that will draw each beat out in the speaker's own words. After the shoot, a paper edit built from the transcript becomes the working script. Scripted lines read by real customers usually sound stiff.
What is the difference between a script and a shot list?
A script is what the viewer will hear and see, in order, beat by beat. A shot list is the crew's version of the video column, broken into individual shots with size, angle, movement and priority, and reordered for the shoot day. The script is written for approval. The shot list is written for production and is built from the script and storyboard.