Everyone who talks fast has been given the same advice, usually by someone who doesn't. Slow down. Take a breath. Pause more. And it works — for about eleven seconds, until your attention goes back to what you're actually saying and your rate snaps straight back to where it was.
That isn't a discipline failure. It's what happens when you treat an output as if it were an input.
Here's the sequence that does work:
- Record two minutes of yourself talking about something real, unscripted.
- Count your words per minute so you know whether you're genuinely fast or only feel fast.
- Read the transcript for tells — gasps, restarts, blurred word endings.
- Identify which of the four causes those tells point at.
- Apply the matching fix, not the generic one.
- Re-measure after two weeks against the same prompt.
- Rehearse under pressure, because a rate that holds when you're calm proves nothing.
Why doesn't "just slow down" work?
Because your speaking rate isn't something you set — it's something that falls out of four other systems: how much air you have, how activated your nervous system is, how far ahead of your mouth your thinking is, and what your motor patterns default to when you stop paying attention. Tell yourself to slow down and you're manually overriding all four at once, using the same attention you need for the actual content.
You can hold that override for a sentence or two. Then the content gets hard, attention shifts, and the underlying system reasserts itself. This is why people report the advice working in rehearsal and failing in the room — the rehearsal had spare attention and the room didn't.
The fix is to change one of the four inputs instead. Change your breathing pattern and the rate changes as a consequence, without supervision. That's the whole strategy.
How fast are you actually talking?
Before fixing anything, find out whether there's something to fix. Feeling fast and being fast are different problems, and nerves reliably inflate the first one — plenty of people convinced they race are sitting at a perfectly normal rate and simply hate the sound of their own recording.
You need a number to compare yourself against, and there's a well-established one.
The measurement takes four minutes: record yourself talking about your work for two minutes unscripted, transcribe it (any phone dictation app will do), count the words, divide by two. Then read the result against this band:
- Under 140 wpm — you're not a fast talker. Whatever is bothering you about your delivery is something else. Try how to be more articulate instead.
- 140–170 wpm — normal range. If listeners still struggle, the problem is almost certainly pauses, not speed. Skip to the pauses section below.
- Above 170 wpm — genuinely fast. Keep reading; the diagnosis section is for you.
One caveat worth knowing: rate is not stable across contexts. Measure yourself in the situation that actually causes trouble. If you only speed up in meetings, a calm kitchen-table recording will tell you nothing useful.
Which of the four causes is yours?
This is the step nearly every article on talking too fast skips, and it's the one that determines whether anything else works. The four causes produce different symptoms and respond to different fixes. Read your recording for the tells.
Most people have one dominant cause and a second one that shows up under stress. Fix the dominant one first — chasing all four at once is how you end up back at "just slow down."
What fixes each cause?
The four interventions below are not interchangeable. Each one changes a different input, and applying the wrong one is why the advice you've already tried didn't hold.
If it's breath: one sentence, one breath
Fast talkers typically ride a single lungful across three sentences, and the last one comes out compressed and unvoiced because there's nothing left behind it. Speed here is the body racing the air.
Set a rule instead of a target: take a breath at every full stop, whether or not you feel you need one. Breathe low, into the belly rather than the top of the chest. Then add volume — speak about 20% louder than feels necessary. Loud speech is physically hard to rush, because volume consumes air faster and forces you to refill. If your voice already tends to thin out, learning to project your voice fixes the pace problem sideways.
If it's arousal: down-regulate before you open your mouth
If you're only fast when you're being evaluated, the target isn't your speech at all. Your nervous system speeds up under observation — shorter breathing, higher heart rate — and speech rate follows the physiology up.
The move is a long exhale before you start: in for four, out for eight, three or four rounds. Extended exhales are the cheapest available lever on physical arousal, and unlike breathing exercises done mid-sentence, this one costs you no attention while you're actually speaking. Then script only your first thirty seconds. Rate spikes hardest at the opening, and a rehearsed opening gives the arousal somewhere to go. There's more on the physical side of this in how to calm presentation nerves.
If it's thought lead time: decide the point before you start the sentence
This is the cause people misread most often. You aren't talking too fast — you're talking too early, starting a sentence while the thought is still assembling, then accelerating to catch up with yourself. The tell is revision: restarts, "sorry, what I mean is," sentences that change direction halfway through.
Speaking slower won't help, because the gap isn't in your mouth. The fix is structural: land on your conclusion before your first word, say it, then explain it. One sentence of point, then support. If your sentences routinely outrun their own destination, how to stop rambling when you talk covers the structure side in full.
If it's habit: measure, drill, re-measure
If you're fast alone in your kitchen, it's motor patterning, and it needs reps rather than insight. Read aloud at deliberately exaggerated slowness — around 80% of your normal speed, over-articulating every consonant — for two minutes a day. You're not learning to talk like that. You're widening the range so your default lands somewhere slower inside it.
Habit is the slowest of the four to shift, and the one where measurement matters most, because the change is too gradual to feel. Expect four to eight weeks of near-daily practice, and judge it by the number, not by how it feels.
Fix one cause at a time and re-measure against the same two-minute prompt every two weeks. Running all four interventions simultaneously makes it impossible to tell which one is doing the work — and if you can't tell, you can't drop the three that aren't.
Is the problem speed, or missing pauses?
Here's the part almost nobody tells fast talkers, and it changes what you should work on: your listener is not overwhelmed by your raw speed. Research on time-compressed speech finds that people follow speech at 200 to 250 words per minute without undue difficulty, and comprehension actually holds up best somewhat above normal conversational rate (time-compressed speech research summary).
Read that against the ~150 wpm baseline and the arithmetic is uncomfortable. Even at a genuinely fast 180 wpm, you're comfortably inside what a listener can decode. So if people keep asking you to repeat yourself, raw rate probably isn't the culprit.
This appears to contradict the standard advice, and it's worth separating the two claims, because both are true. Toastmasters puts the conventional case well: your audience is hearing the point for the first time, so they need time to receive, weigh, and file it while you're already six sentences further on (Toastmasters International, 2021). That's a claim about retention, and it's correct. The research above is a claim about intelligibility, and it's also correct. Your listener can decode your words at speed; what they can't do is file them without a gap. So rushing genuinely does cost you — just not in the place most people try to fix.
Two things usually are. The first is missing pauses: comprehension needs gaps to consolidate the sentence that just landed, and a fast talker's real signature is not high wpm but the total absence of silence. The second is consonant collapse — at speed, your tongue can't reach every target, so word endings blur and "-sts," "-cts," and "-ths" disappear. That's an intelligibility problem wearing a speed costume.
Which reframes the goal. You're not trying to get slower. You're trying to get gappier and cleaner, and those are much easier to do on purpose than a global rate change. Pause a full beat at every paragraph break in your thinking, hit your consonants, and most listeners will describe you as having slowed down when your wpm has barely moved.
If you speak from notes, write the pauses into them. Put the literal word PAUSE where a point needs to land, and break the notes into short separate phrases rather than continuous prose — a page of unbroken text reads as one thing to sprint through, and a page of discrete bullets reads as a series of things with gaps between them.
How to stop talking too fast in the middle of a conversation
Everything above is training. This is triage — what to do when you're already three sentences deep and accelerating, which is a different problem with a shorter list of options.
- Finish the sentence, then stop. Don't brake mid-clause; it sounds like a malfunction. Land the current sentence, close your mouth, breathe once. The silence feels enormous to you and reads as considered to everyone else.
- Drop your pitch on the last three words. Rising, unresolved endings chain sentences together and pull you forward. A downward inflection is a full stop your body can hear.
- Hand the ball over. "Does that match what you're seeing?" A genuine question forces a real pause and resets your breathing, and it costs nothing socially.
- Slow only the next sentence. Not the whole talk — that's the override that never holds. One sentence at deliberate speed is achievable, and it resets the pattern more reliably than a resolution to be slower forever.
- If you're out of time, cut — don't sprint. This is the one situational trigger that has nothing to do with your four causes: you look up, realise you have twelve slides and four minutes, and accelerate. Speeding up doesn't buy you the content, because nothing delivered at a sprint gets retained. Drop the least important material instead, which is easier if you front-loaded your strongest points.
- Don't apologise for it. "Sorry, I'm talking really fast" adds a fast sentence to the pile and tells the room to listen for the thing you don't want them hearing.
The in-the-moment moves are patches, not fixes. They'll get you through today's meeting; only the cause-matched work stops the pattern from showing up in the next one. If you want the wider skill picture this sits inside, the how to improve speaking skills guide covers the full set of delivery habits.
When is fast speech a fluency disorder, not a habit?
Occasionally the answer isn't practice, and it's worth knowing where that line is. Cluttering is a recognised fluency disorder in which speech runs faster than the speaker's system can handle, producing collapsed or deleted syllables, an unusual number of ordinary disfluencies like interjections and revisions, and pauses in grammatically odd places that make the speech sound jerky (ASHA).
Two things distinguish it from ordinary fast talking. Rate is relative rather than absolute — someone can clutter at an average wpm if that rate still exceeds what their system can execute cleanly. And awareness is often reduced: people who clutter frequently don't hear the problem themselves, and their speech improves temporarily when asked to concentrate on it (Stuttering Foundation).
That last detail is the useful signal. If listeners regularly can't follow you, you're surprised when you hear your own recordings, and concentrating fixes it briefly but nothing sticks across weeks of honest practice, that's worth a speech-language pathologist's assessment rather than another drill. It's an evaluation, not a diagnosis — and cluttering responds well to treatment once it's correctly identified.
Key takeaways
- Speaking rate is an output of four systems — breath, arousal, thought lead time, and habit. "Just slow down" overrides all four with attention you need elsewhere, which is why it never holds.
- Measure before you fix: record two minutes, count the words, divide by two. Under 140 wpm you're not fast; above 170 you are. Conversational English averages about 150 wpm.
- Match the fix to the tell. Gasping means breath. Fast only under pressure means arousal. Restarts and self-corrections mean you're starting sentences too early. Fast everywhere means habit.
- Listeners comfortably follow 200–250 wpm, so raw speed is rarely the real problem — missing pauses and collapsed word endings are. Get gappier and cleaner rather than globally slower.
- If concentrating fixes it briefly but nothing sticks, and listeners routinely can't follow you, cluttering is worth ruling out with a speech-language pathologist.
Frequently asked questions
Why do I talk so fast?
Almost always one of four things: you're running out of air and racing your own breath, your nervous system is activated because you're being evaluated, you're starting sentences before the thought is finished and accelerating to catch up, or it's a long-standing motor habit. The tell is context. If you're only fast under pressure it's arousal; if you're fast alone in your kitchen it's habit; if you restart sentences constantly it's thought lead time.
What is a good words per minute speaking rate?
Conversational English averages around 150 words per minute, and roughly 140 to 170 reads as normal to listeners. Below 140 tends to sound hesitant; above 170 starts to register as rushed. Treat it as a band rather than a target — measure yourself in the situation that actually causes trouble, since rate shifts significantly between a relaxed conversation and a meeting where you feel watched.
How do I talk slower without sounding unnatural or slow?
Don't change your rate globally — add pauses and sharpen your consonants instead. Take a full beat of silence at each break in your thinking, and deliberately hit word endings rather than blurring them. Listeners register both as "slower" even when your words per minute barely moves, and neither one makes you sound laboured the way an evenly stretched delivery does.
Does speaking too fast in presentations make you sound less confident?
It reads as nervous more than incompetent, and mostly because of what comes with it — thin sentence endings, no pauses, and blurred consonants all signal someone trying to finish rather than someone in control. The good news is that the confidence signal is carried by the pauses and the landed endings rather than by rate itself, so fixing those two things changes how you're heard faster than a general slowdown does.
How long does it take to stop talking fast?
It depends entirely on the cause. Arousal-driven speed can improve in a single session once you add a long exhale before you start and script your opening thirty seconds. Breath-driven speed usually shifts within a couple of weeks of breathing at every full stop. Habitual fast speech is the slow one — expect four to eight weeks of near-daily drilling, and track it by measurement rather than by feel, because the change is too gradual to notice from inside.
Conclusion
Stop trying to talk slower and start working out why you're talking fast — the diagnosis is the whole job, and the fix follows from it in about a minute. Record two minutes today, count the words, and find your tell in the table above. Then run the one intervention that matches it, out loud, and re-measure in a fortnight.