How to write a video script for a short documentary

Learning how to write a video script for a short documentary starts with one difference. In an advert or a training film, the script is written first and people perform it. In a documentary, the most important words come from real people, and you cannot write those in advance.

So a documentary script is written twice. Before filming, it is a plan: the story you expect to tell, the people and places that can tell it, and the questions you will ask. After filming, it becomes a paper edit built from what people actually said and what the camera actually captured. Treat the first version as a hypothesis and the second as the evidence.

Before any formatting, write the film's purpose in one sentence: who should watch it, and what should they think, feel or do afterwards. A script for an NGO video aimed at funders will choose different moments from one made for the communities a programme serves or for a ministry audience.

Then decide the length and the platform. A two-minute film for a conference screen and a sixty-second version for social media are different scripts, not one film cut shorter. In our own practice, we agree the main length and any cut-downs at this stage, because they change who we film and how many scenes we need.

Many short documentaries follow a simple arc, and it is a useful starting point even if the finished film departs from it. Write a short outline, a few lines per section, that you can share with colleagues and partners before anyone travels or books a location:

Next to each section, list the scenes you hope to film and the interview questions that might produce the lines you need. Write open questions, such as "What was a normal day like before?", never lines for people to repeat. If a contributor's words have to be scripted, the story is not yet ready to film.

• Opening: a moment, image or line that shows what is at stake.
• Person: someone whose experience carries the story, in their own words.
• Challenge: what made life harder, told specifically rather than in general terms.
• Change: what happened, who was involved and what is different now.
• Close: what it means beyond one person, and what you are asking the viewer to do.

The two-column script, sometimes called an audio-visual or AV script, is the most practical format for short documentaries. The left column describes what viewers see: shots, locations, archive, graphics and on-screen text. The right column holds what they hear: interview lines, narration, natural sound and music. Each row is one moment in the film, read from top to bottom.

The format makes gaps obvious. A row of narration with nothing to show, or a beautiful shot with no reason to be there, stands out at once. It also gives the editor, the producer and the person approving the film a shared reference, so feedback can point to a specific row rather than to "the bit near the middle".

Narration is heard once, at the speed of speech, so it should sound like a person talking. Use short sentences, one idea at a time and the active voice. The W3C advises avoiding or explaining jargon, acronyms and idioms, noting that an expression such as "raising the bar" can be read literally by some people with cognitive disabilities.¹ It also recommends speaking clearly, as slowly as appropriate and pausing between topics, which helps listeners and makes captioning and sign language easier.¹

Let the pictures and the interviews do most of the work. Narration should add what viewers cannot see or hear: context, a figure, a date, a link between scenes. Read every line aloud with a stopwatch before recording, and cut anything that repeats what is already on screen. If music runs under speech, the W3C recommends keeping background sound at least 20 decibels below the voice.¹

Accessibility is cheapest at the writing stage. The W3C asks teams to address accessibility as video is planned and produced, names script writers among those responsible, and advises developing any described version at the same time as the main video.² When writing the script, it says to make sure all relevant visual information is included, and to check before finalising that it is covered in the audio.³

The W3C's example is a speaker who states what a chart shows, with the figures, instead of saying "as you can see on this chart".¹ In our own practice, we also ask interviewees to say their name and role on camera and read out key on-screen text in the narration. Keep important information out of the bottom right corner if a sign language overlay may be added.¹ The finished script also gives you a head start on captions and a descriptive transcript, which people who are both Deaf and blind need to access video.⁴

After filming, transcribe every interview with timecodes. Highlight the strongest lines, then rebuild the two-column script using real quotes in the right-hand column and the footage you actually have on the left. This paper edit is where the story is truly written, and it saves days in the edit suite.

Check every quote against the recording, keep it in context and do not join phrases in ways that change their meaning. Share the paper edit with the people who will approve the film before editing starts, and agree that later changes go through the script. A clear, honest script is the best protection for your contributors, your organisation and your story.

• One-sentence purpose and a named audience
• Agreed length and any social media cut-downs
• Story outline with scenes and open interview questions
• Two-column script with pictures left and sound right
• Narration read aloud, timed and free of jargon
• Key visual information spoken in the audio
• Timecoded transcripts and a paper edit before the edit
• Every quote checked against the recording and kept in context