ON-CAMERA DELIVERY BLOG / 8 MIN READ

How fast should I talk in a video?

There is a real answer to this, and it is not "just be natural." For most talking-head video the sweet spot is roughly 150 to 160 words a minute, with deliberate slowdowns on the lines that matter. Here is why, and how to check yours.

150–160words per minute
~130slower, for tutorials
3sto earn the view
0–100craft score

By Thomas, founder of CutScore · Updated June 2026

DELIVERY CHECK · talking_head.mp4
A presenter mid-sentence in front of a camera, gesturing as they speak, the moment where talking pace either holds a viewer or loses one.
CRAFT SCORE
FIXES ADVISED
how your pace reads to a viewer
Average pace in range · 154 wpm
Rushed stretch · 198 wpm, no pause00:41
Flat monotone section · little variation02:17
The 30-second answer For most talking-head video, aim for about 150 to 160 words per minute. That is roughly natural conversation, brisk enough to hold attention, slow enough to follow. Drop toward 130 for tutorials and anything technical, where people need a beat to absorb each step. Push faster, up past 180, only for high-energy shorts, ads, or comedy. The number matters less than the movement: vary your pace, slow down on the important lines, and leave real pauses. A steady 155 that never changes still puts people to sleep. If checking your own pace by ear sounds unreliable, that is exactly what CutScore measures.
WHY THIS IS HARD TO JUDGE

You are a terrible judge of your own talking speed, and I say that with love, because I am too. The first scripted videos I shot came back from a friend with one note: "slow down, I missed half of it." I did not feel fast. In my head I was practically dawdling. The camera does something strange to the nervous system, and the result is almost always a rush.

Here is the mechanism. The lens reads as mild social pressure, your heart rate ticks up, and the brain decides the fastest way out of an uncomfortable moment is to finish talking. So you sprint. Most people land 20 to 30 words a minute faster on camera than in real conversation, and they have no idea. Then the edit makes it worse: you trim every breath and gap to "tighten" it, and now there is no air left for anyone to catch up.

The flip side exists too. Read a script you wrote, with no energy, and you flatten into a slow monotone that feels safe to you and feels like a hostage video to everyone else. So "how fast should I talk" is really two questions: what is the right average, and how much should it move. Both have answers. Both are easy to miss from inside your own head.

THE PACE TARGETS

How fast should you talk, by type of video?

There is no single magic number, but there is a sensible range for each kind of video. Find the row that matches what you make, then aim for the band, not the exact figure.

Type of videoTarget paceWhy it sits there
Conversational talking head150–160 wpmThe comfortable default: bright, natural, easy to follow without effort.
Tutorial / how-to120–140 wpmPeople are following steps, often while doing them. Give them room.
Explainer / educational140–150 wpmNew ideas need a beat to land. Slightly slower than chat, never droning.
Vlog / storytelling150–170 wpmEnergy carries it. A little faster reads as enthusiasm, not panic.
High-energy short / ad170–200 wpmShort runtime, punchy hook, you want urgency. Keep the pauses surgical.
Audiobook / calm narration~150 wpmListeners settle in. Steady and warm beats fast every time here.
For referenceNormal everyday conversation runs around 120 to 150 words a minute, and a slow, careful presenter can drop under 110. Above 200 you start losing people who are not already locked in. These are speech research norms, not platform rules, so treat them as a comfortable band rather than a hard line.
STOP GUESSING AT YOUR PACE

CutScore reads your speaking rate straight from the audio, flags the rushed and flat stretches, and gives you the timestamps. No stopwatch, no counting words by hand.

Join the waitlist
THE PART THAT ACTUALLY MATTERS

Why pace variation beats one perfect speed.

If you take one thing from this, take this. Hitting 155 words a minute and holding it like a metronome is its own kind of failure. The human ear treats a constant pace as background noise, the same way you stop hearing a fridge hum. Attention lives in change. A delivery that speeds up through the setup and slows down hard on the punchline keeps people leaning in, even if the average is identical to a flat read. The number is a starting point; the contrast is the craft.

Slow down on the lines that carry weight

Your most important sentence should be your slowest. The price, the punchline, the one statistic, the "here is the thing nobody tells you" line: drop the pace, leave a pause before it, leave a pause after. That silence is not dead air, it is emphasis, and it tells the viewer this part counts. Most beginners do the opposite. They rush the payoff because they are nervous it will land flat, and the rush is precisely what makes it land flat.

A microphone and headphones set up in a podcast studio, the kind of close, controlled space where small changes in talking pace become very easy to hear.
In a quiet, close mic setup every change of pace is audible. Use it on the lines that matter. Photo: Jakub Żerdzicki / Pexels.

Pauses are pace too, and you keep cutting them

A pause is not a gap to be deleted. It is part of the rhythm. In my early edits I treated every silence as a mistake and snipped it out, which left a breathless block of speech that no one could follow. The fix was counterintuitive: I added air back. Real talking has small pauses at the end of thoughts, and your viewer uses them to process what you just said. Cut too tight and you get a perfectly trimmed video that somehow feels exhausting. Tight is good. Suffocating is not.

Fast talking and filler words are the same problem wearing two coats

When you rush, you also reach for filler. "Um," "like," "so basically," "you know" multiply when your mouth is moving faster than your brain has decided what to say. Counting your filler words per minute is often a better read on your pacing problem than the raw word count, because the fillers are the audible symptom of going too fast. Slow your default down by ten or fifteen words a minute and a surprising number of "ums" just evaporate, because now you have time to think before you speak.

RATHER SEE IT THAN READ IT?

Here is a real CutScore coaching report for an everyday talking-head video: pace, filler words, hook and the rest, scored, with timestamps and the exact fixes.

See a sample report
FIX IT THIS WEEK

If you only change three things.

Most of the jump from "this person is nervous" to "this person knows what they are doing" comes from these three. Fix them first.

1
2-MIN FIXDELIVERY
Slow your opening 15 seconds on purpose
The rush is always worst at the start, when nerves peak. Deliberately begin slower than feels right, breathe before your first line, and you set a calmer pace for the whole video. The opening is also where most drop-off happens, so it is the worst place to sound panicked.
How Mark your first sentence "SLOW" in the script, and take one real breath before you hit record on the line. More on the opening hook.
2
EDITRHYTHM
Leave the pauses around your key lines
When you tighten the edit, resist trimming the silence before and after your most important sentence. That little gap is what makes the line land. Cut the dead filler, keep the deliberate pauses. Tight is good, breathless is not.
How Find your single best line, then protect a short pause on each side of it before you trim anything else.
3
QUICKSPEECH
Measure your words per minute, once
You cannot fix a pace you have never measured. Run the number on one real video and you will instantly know whether you are a sprinter or a droner. Most people are shocked. After that, you have a target instead of a feeling.
How Count the transcript words, divide by minutes of talking, compare to the table above. Or let a tool read it for you.
THREE WAYS TO CHECK YOUR PACE

By ear, by stopwatch, or in one pass.

OPTION 01

By ear, a day later

Free, and the cheapest honesty you can get. Watch the video after a day away from it, or hand it to a friend with one instruction: tell me where you got lost. The catch is the one we opened with. From inside your own head, fast feels normal, so your ear alone will lie to you.

OPTION 02

With a transcript and a stopwatch

Accurate and honest. Pull the transcript, count the words, time the talking, divide. Now you have a real number to compare against the table. The cost is effort, and it only gives you an average. It will not point at the one rushed stretch at 0:41 that is actually losing people.

OPTION 03

With a coach in one pass

Hand the file or a link to CutScore. It measures your speaking rate from the audio, marks the stretches that run too fast or go flat, and folds it into a 0 to 100 craft score with timestamped evidence and fixes. Delivery is one of the families we check. See a sample report.

How CutScore reads your delivery CutScore is an AI video quality coach for pre-publish QC. For pacing it measures your speaking rate from the audio, finds the stretches that sprint past the comfortable band or flatten into a monotone, and tells you exactly where on the timeline they happen. It judges the craft of how you deliver, the same way it judges loudness or exposure, and reserves AI for the genuinely subjective parts. One score, the evidence behind it, and a prioritised list of fixes, before anyone else hits play. More on the method and the standards.
QUESTIONS

Frequently asked.

Around 150 to 160 words per minute is the comfortable middle for most on-camera talking. That is roughly conversational, a little brighter than a slow lecture and a long way short of an auctioneer. Tutorials and explainers sit a touch slower so people can follow along; high-energy shorts and ads run faster. If your pace lands in that band and you still vary it, you are fine.
Neither at the extremes. A flat, slow delivery makes viewers scroll away because nothing feels urgent. A wall of fast speech with no pauses exhausts them because they cannot keep up. The trick is not one speed, it is contrast: a brisk default with deliberate slowdowns on the important lines. Pace that moves keeps people watching far better than any single number.
Nerves, mostly. On camera your body treats the lens like a threat, so you rush to get it over with, often 20 to 30 words a minute faster than you think. Cutting pauses out in the edit makes it worse. The fix is to slow your start, breathe at the end of sentences, and leave a little air between thoughts instead of trimming every gap.
Pull the transcript, count the words, divide by the minutes of actual talking, and you have your words per minute. Do not count silent b-roll or music-only sections. For a faster read, CutScore measures your speaking rate from the audio, flags the stretches that run too fast or too flat, and gives you the timestamps, so you fix the delivery instead of guessing at it.
EARLY ACCESS

Stop guessing how you sound.

CutScore measures your pace, your filler words and the rest of your delivery, then tells you exactly what to fix, with the evidence to back it up. Join the waitlist for early access.

Join the waitlist