Explainer Video Words per Minute: Word Budgets for 2 to 10 Minute Scripts, Why Longer Scripts Drift, and How to Time One Properly

Searches for 'explainer video words per minute' and 'how long is a 3 minute script' usually come from someone with a half-written tutorial and a deadline. This is the reference post for that moment: speaking-rate bands, word budgets for 2 to 10 minutes at three paces, why longer scripts drift from any single rate, and a budgeting method that survives a screen recording. It is written around the free iPhone app Teleprompter: Camera Overlay, but the arithmetic works anywhere.

Scope: 2 to 10 minute explainers, tutorials and course lessons. Short clips follow different rules; the ideal script length for a 60-second video post covers 15 to 90-second budgets and is not repeated here.

Speaking-rate bands: which words per minute to budget for

There is no correct speaking rate, only a correct one for the job. These are rules of thumb, not measurements from a study.

• Slow, roughly 110 to 130 wpm. Tutorials where the viewer follows along on their own screen: software walkthroughs, coding lessons, recipe steps. The viewer is doing something while listening, so every sentence needs air after it.
• Conversational, roughly 140 to 160 wpm. Explainers, course lessons and commentary where the viewer is only watching. What most people mean by 'normal talking speed'.
• Brisk, roughly 170 wpm and up. Ads, trailers, anything cut hard to music. Tiring past a minute or two, so rare in a 10-minute explainer.

If you have never measured yourself, assume 150 wpm for a straight explainer and 130 wpm for a tutorial with on-screen steps.

Word budgets for 2, 3, 5, 8 and 10 minute scripts

Minutes multiplied by words per minute, at the three paces above.

• 2 minutes: 260 words at 130 wpm, 300 at 150 wpm, 340 at 170 wpm.
• 3 minutes: 390 words at 130 wpm, 450 at 150 wpm, 510 at 170 wpm. So a 3 minute script is about 400 to 450 words for most people, nearer 390 if viewers are following along.
• 5 minutes: 650 words at 130 wpm, 750 at 150 wpm, 850 at 170 wpm.
• 8 minutes: 1,040 words at 130 wpm, 1,200 at 150 wpm, 1,360 at 170 wpm.
• 10 minutes: 1,300 words at 130 wpm, 1,500 at 150 wpm, 1,700 at 170 wpm.

Treat these as ceilings, not targets. The table assumes constant talking for the whole running time, and nobody does that; a 1,300-word script almost always runs past 10 minutes once the demos slow it down.

Why longer scripts drift away from the number

A 60-second clip is one pace from start to finish. A 7-minute tutorial is four or five paces stitched together, and a single figure averages them away.

• Demo and screen-recording sections run slow. You say 'click Export', then actually click Export, wait for the dialog and point at the thing. Thirty words can take thirty seconds.
• Lists run fast. 'Three reasons: cost, speed, and support' is six words in three seconds. Recaps race ahead of the average, which is why the ending of a long video often feels rushed even when the count was right.
• Explanations sit in the middle. Narrated parts land near your measured pace and are the only parts the figure predicts well.
• Pauses add up. More section breaks and transitions than a short clip, each buying no words.

Net effect: a script that is 40 percent demo runs long, one that is 40 percent bullet points runs short. Same word count, different videos.

The section-budget method

Budget per section, not per script. Five minutes before you write saves an afternoon of re-recording.

• Outline the video as sections with a target time each: intro 30 seconds, concept 90 seconds, demo 3 minutes, pitfalls 90 seconds, recap 30 seconds.
• Give each section its own pace: narration at your measured rate, demos at roughly half of it (half the time is showing, not saying), lists and recaps at the brisk rate.
• Multiply, and write the budget into the heading: 'Demo (3 min, about 220 words)'. When the demo section hits 300 words, that is the signal to cut, not to talk faster.
• Add the sections back up. If the sum is over the whole-video ceiling, cut from the section furthest over its own budget, usually the concept explanation.

The teleprompter for online courses and tutorials post covers laying out a lesson so each section can be re-recorded on its own; a section 20 percent over is then a two-minute retake, not a new afternoon.

How to time a script (and why a script timer is usually wrong)

Online 'video script timer' tools divide the word count by a single rate, so they reproduce the drift above: accurate for a narrated explainer, optimistic for a tutorial with demos, pessimistic for a list-heavy recap. Fine for a first sanity check, poor for a final one.

The reliable method is older. Read the script aloud with a stopwatch, section by section, at the pace you will use on camera, including the moments where you would be clicking or pointing. Write the real time next to each heading; anything more than 15 percent over gets cut before you record. Reading aloud also catches sentences that do not fit in a breath.

A total 25 to 30 percent over is a normal first draft. The how to edit a script down by 30 percent post is the cutting routine for exactly that.

What a voice-following teleprompter does and does not fix

Classic auto-scroll makes long scripts harder: a speed set for the narration is too fast for the demo and too slow for the recap. Teleprompter: Camera Overlay has that mode with a speed slider, and it suits a steady read. For a tutorial, voice-driven scrolling fits better: the script floats over the camera preview next to the front lens and moves at the speed you speak, using Apple's on-device speech recognition. It detects the script's language automatically, works in any language iOS supports for on-device recognition, and finds your place again if you stumble or stop to click something. The script never appears in the saved 4K portrait video, and it all runs offline.

That removes the timing problem during recording: pause for the demo, speed through the recap, and the text follows. It does not remove the budgeting problem. If the script is 1,700 words for a 10-minute slot, voice-follow will cheerfully let you read all 1,700 in 13 minutes. A prompter cannot shorten a script; cutting still happens on the page. The takes library helps: record the long version, watch it back with the stopwatch, cut, record again. Scripts and takes are unlimited; the app is free with occasional ads, a small one-time purchase to remove them, no subscription, no watermark, iOS-only.

When not to budget at all

Live tutorials with real-time typing are the exception. If the lesson is you writing code or building a spreadsheet as the viewer watches, the pace is set by your hands, not your script, and a word budget fights the format. Do not script those; outline the steps, narrate as you go, and cut the dead air in the edit. Budget the planned sections and leave the live ones alone.

Next in the series

Once the script fits, the next thing that goes wrong on a long recording is the light changing halfway through. The next post covers the best teleprompter settings for different lighting conditions, from a bright window to a single ring light in a dark room.