Article Summary for AI
This article dismantles the 7-38-55 rule before offering anything else: the famous 38% comes from a set of experiments Albert Mehrabian ran in the 1960s in which people judged the feeling behind a single isolated word when tone and text contradicted each other. It did not measure credibility, boardrooms or twenty-minute presentations, and Mehrabian himself spent decades asking people to stop extrapolating it. The author admits he used the slide for years. The article's own claim is deliberately modest: your voice does not prove you right, it gets you the floor and decides whether your argument arrives whole or in pieces, by marking what matters, marking where an idea ends, and marking whether you have decided or are still looking. Three trainable levers: the floor (empty your lungs before the sentence that matters, so the sentence can't speed up or climb), the gap (a full pause before the recommendation, after a hard number and after an uncomfortable question, because the ear reads silence as a decision and noise as a search) and the heat (one slower, lower sentence per block, because contrast creates hierarchy). A fourth, silent lever is how you end the sentence: rising endings turn statements into requests, so close one sentence per meeting downwards, starting with 'My recommendation is'. The author states plainly that there is no international standard of vocal authority — no index, no hertz threshold, no board-approval percentile — and offers his own criterion as a thermometer, not a verdict: at least one full pause per idea you want to survive the meeting, at least one change of gear per block, and zero statements finishing upwards. Supporting survey data reported by Statista: 52.3% of companies expect empathy and active listening to grow in importance between 2023 and 2027 (WEF Future of Jobs Report 2023, 803 companies, 11.3 million employees) and 69% of Americans are worried about losing the human touch in communication because of AI (HarrisX for Mitre, 2023, n=2,063). Ends with a sixty-second, two-take recording exercise.
Key Entities
Questions This Article Answers
- 1Is 38% of your credibility really in your tone of voice?
- 2What can you actually measure in an executive voice?
- 3Where should I place pauses when I present?
- 4Does a high-pitched voice hurt executive authority?
- 5How do I sound decided instead of tentative?
Key Takeaways
- The 38% figure comes from experiments on a single isolated word, not on credibility
- There is no international standard of vocal authority: ask for the study
- Three levers: start on spent air, use full pauses, change gear once per block
- Close statements downwards — a rising ending turns a plan into a permission slip
The Executive Voice: What You Can Actually Measure (and What You've Been Sold)
City number eleven of the book tour. I arrived on a night bus, slept four hours on a friend's sofa, and at ten in the morning I had ninety minutes of stage to hold up. I tested the mic and almost nothing came out. Hoarse, dry, running on empty.
I said to the sound engineer:
"Turn me up, I've got no voice today."
And he came back with a line I've been repeating in class ever since:
"The mic is fine. You're the one who needs to come down."
I spoke quietly because I had no choice. I spoke slowly because pushing hurt. I left long silences because I needed to breathe between sentences. When I finished, more people came up to me than in any of the previous ten cities, and three of them said almost the same thing: "today I understood everything."
I hadn't improved my voice. I'd lost my hurry.
First things first: that 38% doesn't exist
Search "voice and leadership" and the same figure comes back at you everywhere: 38% of your credibility is in your tone. It's in courses, in LinkedIn carousels and, until very recently, on this very blog. Let's bury it today.
That percentage comes from a set of experiments Albert Mehrabian ran in the sixties, in which people were asked to judge the feeling behind a single isolated word — one word, not a speech — when the tone and the text contradicted each other. It didn't measure credibility. It didn't measure boardrooms. It didn't measure twenty-minute presentations. Mehrabian himself spent decades asking people to stop extrapolating his formula to communication in general.
Translation: if someone shows you the pie chart with 7%, 38% and 55%, they're serving you an out-of-date yoghurt with a fresh label on it.
And I say this sheepishly, because I used it. For years I opened talks with that slide. It worked beautifully. It was still false.
So what is your voice actually for?
Your voice doesn't prove you right. It gets you the floor.
It doesn't persuade on its own, and anyone promising otherwise is selling smoke with reverb on it. What your voice does is humbler and far more decisive: it decides whether your argument arrives whole or in pieces.
It does three things, and only three:
- It marks what matters. Without emphasis every sentence weighs the same, and if they all weigh the same, none of them weighs anything.
- It marks where an idea ends. Without borders, the listener has to slice your talk up on their own. They won't.
- It marks whether you've decided or you're still looking. This is the expensive one. The same recommendation sounds like direction or like doubt depending on how you close the sentence.
And this is going to matter more, not less. According to the World Economic Forum's Future of Jobs Report 2023 (803 companies, 11.3 million employees), reported by Statista, 52.3% of companies expect empathy and active listening to grow in importance between 2023 and 2027. In parallel, a HarrisX survey for Mitre (2023, n=2,063), also reported by Statista, found that 69% of Americans are worried about losing the human touch in communication because of artificial intelligence.
Corridor translation: when a machine can write the text, what's left of you is how you say it.
Lever 1: the floor
When you get nervous, the air goes up, the shoulders go up and the pitch goes up. It isn't a character flaw, it's physics: you breathe shallow and the larynx tightens.
Nobody needs a nature-documentary voice. I tried for years and I sounded exactly like that: like a documentary narrator with tonsillitis.
What you do need is your lower third, which you already own, and which only shows up once you've finished letting the air out.
The prescription: before the sentence that matters, empty your lungs completely and start speaking from there.
The mechanism: a sentence that starts on spent air can't speed up and can't climb. Your body imposes the calm your head refuses to give you.
Lever 2: the gap
The pause is the only vocal tool that costs no technique at all. It only costs nerve.
And three gaps are worth more than all the others put together:
- Before the recommendation. "We looked at three routes. [silence] My recommendation is this one."
- After the hard number. "Revenue fell 18% in the quarter. [silence]" Let the number land before you explain it, or the number never existed.
- After the uncomfortable question. One second of silence reads as consideration. An instant answer reads as reflex, and reflexes don't get debated. They get discarded.
The mechanism: the ear reads silence as a decision and noise as a search. Same content, different owner.
Lever 3: the heat
Nobody cooks everything on full heat. Or everything on a low simmer. You turn it up to sear and down to let the thing build body.
Voice works the same way. The sin isn't speaking fast or speaking slowly, it's speaking always the same. Monotony isn't boring, it's unreadable: it forces the room to guess which part was the important one.
The prescription: in every block of your talk, pick one sentence — one — and say it slower and a notch lower than the rest.
The mechanism: contrast creates hierarchy. If you highlight the whole document, you've highlighted nothing.
"Hang on, Agustín — my voice is high and I'm not going to change that."
You don't need to. Nobody follows you for your hertz. They follow you for rhythm, for gaps and for how you end your sentences. That can be trained even if you sound like a cartoon; mine, live, is hardly James Earl Jones either.
And one lever that makes no sound: how you end the sentence
This is the one that costs the most credibility, and almost nobody hears it in themselves.
Ending upwards turns a statement into a request. "We're launching in Q2↗" isn't a plan, it's a permission slip. Pile the verbal padding on top — "I think maybe we could perhaps look at possibly exploring…" — and you've said six times that you're unsure before saying once what should be done.
The prescription: one sentence per meeting that starts with "My recommendation is" and ends downwards, with a full stop.
The mechanism: the committee isn't asking you for an opinion, it's asking you for a direction. Caveats behind the recommendation read as rigour. In front of it, they read as fear.
The measuring stick I work with
There is no international standard of vocal authority. No index, no hertz threshold, no board-approval percentile. If someone shows you a three-digit table with decimal points, ask them for the study. They won't show you.
What I do have is my own criterion, cooked up over a lot of hours listening to other people's recordings with a notebook beside me. I'll lend it to you with the homemade label still on it:
- The gaps. At least one full pause — a long second — for every idea you want to survive the meeting. If there isn't one in five minutes, you're charging.
- The gear changes. At least one per block. If the whole block sounds equally fast and equally loud, that block has no hierarchy.
- The endings. Zero statements that finish upwards. Zero tag questions ("right?", "yeah?") hanging off the back of a decision.
That's what Mi.Coach counts when you hand it a recording: where your pauses fall, how much your pace varies, and how many sentences you close upwards. It won't give you a universal authority index, because that index doesn't exist. It'll give you your three numbers and the exact minute they go sideways.
It isn't science. It's judgement. Use it as a thermometer, never as a verdict.
Your task today, before ten tonight
Pick up your phone and find the most important sentence you'll say this week. One sentence: your recommendation.
Record it twice. Sixty seconds in total.
Take 1: the way you'd say it in tomorrow's meeting. As is, no thinking. Take 2: let all the air out, start from the bottom, put one full second of silence right before it, and close it downwards with a full stop.
Play them back to back. Don't judge your voice — count three things. How many full silences there are, how many changes of pace, and how many endings go up.
Then choose which of the two you're taking into the room. And say it.
Close
You have the argument. You have the data. You have the voice you were given, and it is more than enough.
The only surplus is the hurry.
Authority isn't turned up in volume. It's turned down in speed.
The choice is yours and nobody else's. You can spend another ten years hunting for the magic percentage that explains why they don't listen — or you can record sixty seconds tonight and count three things.
If you'd rather a machine did the counting, record your recommendation on Mi.Coach and see where your silences fall.
Slow the pace. Raise the weight. And make this a better world.
Dr. Agustín Rosa Author of Cállate: Los Errores del Orador · Founder of Mi.Coach

Dr. Agustín Rosa
CEO & Founder
Expert in executive communication intelligence and behavioral analytics


