Narration for Video Manuals: Costs and How to Write the Script

Narration for a video manual should not be read like a corporate video

Training managers at manufacturers, consultants rolling out business systems, and production companies handling manual work sometimes tell me that commissioning narration the way they would for a corporate video did not quite fit.

The cause is neither the voice nor the technique. The viewer is in a completely different state.

Someone watching a corporate video is sitting down, looking at the screen. Someone watching a video manual is working with their hands and glancing at the screen. They may be holding a tool, or performing the actual operation in another window.

That difference determines everything about how it should be read.

What it costs

Pricing works the same way as a corporate video: it follows the length of the footage.

Pricing starts from a three-minute script, with roughly +¥10,000 for each additional minute.

Video manuals differ from corporate videos in two ways. They run long, because steps cannot be cut. And they come in batches — five videos for five functions, twelve for twelve processes.

For that reason, the per-video cost comes down when you commission them together. Once the tone and the pronunciation of your terminology are settled, every video after the first reuses that work. Tell me the total count when you ask for a quote.

Three things that change in the read

1. Flatten the emphasis

In a corporate video, you lift the words that matter to create an impression. Do that in a video manual and you misrepresent which step matters.

Read "always switch off the power first" with weight on "power," and the viewer takes "power" as the point. If the real point is the sequence — first, before anything else — the emphasis has landed in the wrong place.

For procedural work the baseline is an even tone where every word arrives equally. Emphasis is held back for the places where a mistake causes injury.

2. Wait for the work

Someone operating the screen cannot keep pace with continuous narration.

If "remove the four screws" is followed immediately by the next step, the viewer has to pause. Silence proportional to the actual task has to sit between steps.

That silence can be added in the edit, but it sounds more natural when the pause is taken during recording. If the script says "(work: approx. 10 seconds)," that is how it gets read.

3. Assume pausing and rewinding

A video manual is rarely watched straight through. Viewers jump in partway and rewind the same passage repeatedly.

Each step therefore has to make sense on its own. Phrases that lean on what came before — "in the previous step," "now take this" — lose their meaning when playback starts in the middle.

Writing the script

Always add readings for part numbers and terminology

This is where video manuals go wrong most often.

Is the "1" in a part number read as a digit or as a word? Is "M6" read as "M-six" or as separate characters? A reading that is obvious inside your company is not obvious from outside.

Without a stated reading, the narrator has three options: look it up, ask, or read it uncertainly. Removing one round of questions moves delivery up by half a day.

Roughly 300 characters per minute

This is the standard guideline, and it holds for video manuals too. A five-minute video runs to about 1,500 Japanese characters at a normal pace.

For a video manual, though, subtract the silence for the work itself. If a five-minute video contains a minute of waiting, the script should come in around 1,200 characters.

Decide whether step numbers are spoken

State whether "①" or "Step 1" should be read aloud. If the number is large on screen, leaving it out reads more cleanly; if the audio is sometimes used without the picture, reading it is kinder. Neither is wrong — the problem is leaving it undecided.

What the client gets back

These preparations are a courtesy to the narrator, and they return as practical benefits.

Revisions cost less. Video manuals get partially re-shot every time the specification changes. If the audio was delivered split by step, only the changed step needs replacing. Nothing else has to be recorded again.

Fewer rounds of questions. With readings and work timings written into the script, pre-recording queries all but disappear.

Frequently asked questions

Q. Can the audio be delivered split by chapter?
A. Yes. Tell me your file naming convention as well. If replacement is expected later, splitting finely from the start is the safer choice.

Q. If one step is added later, does the minimum fee apply again?
A. Replacements and additions are quoted individually. Work that continues an existing project is not treated as a new commission.

Q. It will be viewed where sound cannot be played, so subtitles do the work. Is narration still needed?
A. Even with subtitles carrying the load, audio speeds up comprehension. That said, one reasonable allocation is to prioritise subtitle accuracy and add voice only to the steps that carry risk. Tell me the setting and I will propose the range that is actually needed.

Q. How long does it take?
A. Within 24 to 48 hours of receiving the script. For large batches, chapters can be delivered as they are finished.

Q. Our terminology is not even consistent internally.
A. Come to me with it as it stands. The readings have to be settled before recording anyway, so I will organise the open questions and send them back. That often becomes the trigger for standardising the wording internally.

In short

Narration for a video manual is read fundamentally differently from a corporate video, because the viewer is listening while their hands are busy. What it needs is not emphasis but an even tempo that is not missed, and silence that waits for the work.

On the script side: readings for part numbers and terminology, stated work timings, and a decision on whether step numbers are spoken. With those three settled, the rounds of questions drop and delivery comes sooner.

And split the delivery by step. Video manuals always get updated. How you set it up now decides what a revision costs two years from now.