Every video script is longer than it looks. You count the words, divide by a hundred and fifty, and get a tidy number that has almost nothing to do with the finished video. Then you record it, and the two-minute piece runs three and a half.
The gap is not your math. It is everything a word count cannot see: the pace you actually speak at, the pauses that make delivery sound human, the seconds where B-roll carries the story with no narration at all, the on-screen text a viewer needs time to read. Video script length is the sum of all of it, not just the talking.
A "two-minute video" almost never comes in at two minutes. It comes back from review with one more line, the intro grows, somebody asks for a title card, and at eleven at night you are cutting a scene you liked to make the slot. So treat this as a planning estimate, not a rule. Paste your script, or run it backwards from a target runtime. And when you want the real answer, rehearse it with the built-in teleprompter and time yourself.
A hard runtime is real when the slot is real: a broadcast ad, a conference reel, a paid spot you bought thirty seconds of. Almost everywhere else on the web, nobody is holding a stopwatch. Cutting a good story down to a "clean" thirty seconds because thirty sounds tidy is how you end up with something forgettable. Let the story decide the length, then tighten it. Use the number as a target when it serves the piece, and ignore it when it does not.
Never script to the exact runtime, especially if this is not your day job. It is easy to cut a strong piece down. It is miserable to stretch a thin one, and it shows. Give yourself at least thirty extra seconds of material, more if stakeholders are involved, because the moment somebody asks to remove a line you are suddenly under length with nothing to put back. Extra footage is cheap insurance. An empty timeline is not.
It starts with your pace. The National Center for Voice and Speech puts average conversational English at about 150 words per minute. Formal and documentary narration sits nearer 130, corporate and explainer narration around 140, and broadcast ad reads climb to 170 and beyond. Choosing the wrong pace is the single biggest error in a length estimate. The voiceover industry's own shorthand agrees: a sixty-second script is roughly 150 words.
Then it adds what the script does not contain. Real delivery includes breathing and pauses, which stretch a read by roughly eight percent. Silent B-roll, on-screen text and branding bumpers add seconds that no word count will ever show you.
The result is a range, not a single number, because delivery varies from person to person. If you want the honest answer rather than a good estimate, hit Rehearse and time it, read your script aloud with the teleprompter, and use your own measured pace. Nothing is recorded and nothing leaves your browser.
Write and record more than you need, then cut down to length. Aim for at least thirty seconds of extra material, and more when stakeholders are involved.
Cutting is easy and it makes a piece tighter. Filling a gap in a script that came up short is painful, and the padding is always visible. If someone asks to remove a line late in the process and you scripted to the exact second, you are now under length with nothing left to put back.
Only when the slot is real. A broadcast ad, a paid spot or a conference reel has a fixed duration you cannot argue with. Most web video does not.
Storytelling beats runtime almost every time. Forcing a good idea into a "clean" thirty seconds because broadcast works that way, when nothing on the web requires it, usually just damages the piece. Let the story set the length, then tighten it.
Around 150 words at a natural conversational pace, which is the voiceover industry's long-standing rule of thumb for a sixty-second script.
It shifts with delivery: roughly 130 words for a formal narration, about 140 for a corporate explainer, and up to 170 for a fast ad read. Subtract any seconds used by silent B-roll, on-screen text or an intro before you write.
About 3 minutes 35 seconds at a corporate explainer pace of 140 words per minute, once you allow for natural pauses. At a faster conversational pace it lands nearer 3 minutes 20 seconds, and a deliberate documentary read pushes past 4 minutes.
Between 130 and 170 words per minute for almost all video work. The National Center for Voice and Speech puts average conversational English at about 150 wpm, and the American Speech-Language-Hearing Association puts formal speech nearer 130.
In practice: documentary and formal narration around 130, corporate and explainer around 140, conversational talking heads around 150, and broadcast or social ad reads from 170 upward.
Only when it is silent. B-roll running underneath your voiceover costs you nothing, because the narration keeps going. What it buys you is room: you can slow the read down, let a line land, and give the piece somewhere to breathe.
B-roll with no narration over it is different. Those seconds are pure additions to your runtime, and they are the ones to count here.
Because a script only measures talking. Finished videos also contain pauses, breathing, silent B-roll, on-screen text that needs read time, and intro or outro bumpers. Together these routinely add 10 to 30 percent.
I am Gustavo Fonseca, Senior Content Lead. I have worked in video and content creation for more than twenty years, now based in Vancouver, Canada.
I hold a bachelor's degree in Cinema and Video from UFSCar (Federal University of São Carlos, Brazil), a postgraduate diploma in Strategic Business Management from Mackenzie Presbyterian University, a certificate in Digital Marketing Strategy from Harvard Business School, and a series of AI credentials from UBC, the most recent of them in AI Product Management.
Want to know more about my work? It is all at thegusta.com.
It is just your text scrolling and the timer. We do not ask for access to your camera or microphone, and nothing ever leaves your browser. Read aloud, as if you were already recording, and stop when you reach the end.