Emotion · about 4 minutes
The 7% Room
Seven per cent of your message is the words. Thirty-eight per cent is your tone. Fifty-five per cent is your face and your body. Which is to say: ninety-three per cent of communication is nonverbal.
You have met this number. It is in the presentation course, the sales training and the management book, and it is on a truly remarkable number of slides. Before we tell you where it came from, we'd like you to do the one thing nobody quoting it ever seems to do.
Six messages. For each one you are on the far side of a loud, crowded room from the person who needs it, and you get exactly one channel to say it with. Take whichever one actually gets the message there.
Message 1 of 6
Say it with…
What you just did
Nobody had to teach you that. And notice what it is not: it isn't that one channel is richer than the other. It's that they carry different cargo. Words carry Thursday. Faces carry how we're doing.
Now let's do it the way the slide does it
The rule doesn't sort anything. It says every message arrives in the same fixed proportions, whatever the message happens to be. Here is one of yours, drawn to scale. Try the others.
7% the words 38% the tone 55% the face
Two studies, 1967, both rather small
The first is Albert Mehrabian and Morton Wiener, Decoding of inconsistent communications. Nine single words -- three warm (dear, thanks, honey), three neutral (maybe, oh, really), three cold (brute, don't, terrible) -- each recorded by two female speakers in three different tones of voice. Three groups of ten listeners judged one thing: how much the speaker liked the person being spoken to. When the word and the tone disagreed, the tone won.
The second is Mehrabian and Susan Ferris, Inference of attitudes from nonverbal communication in two channels. This time the single word maybe, spoken three ways, paired with photographs of three faces. The face beat the tone by about half again.
From those, coefficients: .07 for the words, .38 for the voice, .55 for the face. Mehrabian put them together in a book called Silent Messages in 1971, and the number left home.
Five things it left behind
1. They were single words. Not sentences, not conversations, not meetings. One word at a time, and one of them was "maybe."
2. The channels were rigged to disagree. That's what "inconsistent communications" means in the title -- the design deliberately set the word against the tone. It measures which channel you trust in a conflict. That is a real and interesting question, and it is not remotely the same question as how much of a message each channel carries.
3. Nobody ever tested all three at once. One study did words against tone. The other did tone against faces. The famous three-slice pie was assembled out of two two-slice experiments, and the papers never actually show the arithmetic that joined them.
4. The 55% is photographs. Still photographs, of three faces. There is no body language anywhere in the body-language statistic. No gesture, no posture, no hands, no stance, nobody moving at all. When you chose "tone, face and body language" across your crowded room, you were granting the claim more than its own evidence ever contained.
5. It was about liking, and only liking. Every judgment every subject made was how much does this speaker like the person they are talking to. Not how much did you understand. Not how much arrived.
And all of it -- the two speakers, the three photographed faces, the thirty listeners -- was women. Which the authors said at the time, in print, as a limitation.
What its author has been saying ever since
"Please note that this and other equations regarding relative importance of verbal and nonverbal messages were derived from experiments dealing with communications of feelings and attitudes (i.e., like–dislike). Unless a communicator is talking about their feelings or attitudes, these equations are not applicable."
-- Albert Mehrabian, on his own website, for years
He has been saying versions of that for more than fifty years. It has made no difference whatsoever.
The awkward part, and why this room is here
Two rooms ago you turned three dials and watched rage become terror without moving two of them. Those dials are the PAD model -- pleasure, arousal, dominance -- and one of the two people who built it is Albert Mehrabian. The machine in Exhibit № 38 has its whole inner life measured on his axes. This wing runs on his work, and it is good work.
So let's be careful about what the story is. It is not a story about a bad scientist. He ran two small, careful experiments, reported what he found, published the limitations himself, and has spent half a century asking people to stop stretching it. He is the most reliable witness against his own famous number.
It's a story about what happens to a finding after it leaves the building. And the genuinely uncomfortable part is that nothing in that process was unusual. A result gets a memorable shape -- three numbers, and they add to a hundred. The shape is portable. The conditions are not. So the shape travels, gets rounded, gets a pie chart, gets a slide, and arrives somewhere it was never measured, wearing all of its original confidence and none of its original sentence.
Thirty people. Single words. Nineteen sixty-seven. A million slides.
You are carrying numbers like that right now. Several, probably. They came to you the same way -- pre-shaped, unsourced, from someone who was repeating someone who was repeating someone. That isn't a reason for despair, and it certainly isn't a reason to believe nothing. It's the reason for the small, cheap habit this entire museum is really about: when a number is that quotable, go and look at what it was measured on.
There's a gauge for that, just below. This is the one room in the building where it is pointing at us as much as at the science.
Take it further
The two papers are Mehrabian & Wiener, Decoding of inconsistent communications (Journal of Personality and Social Psychology, 1967) and Mehrabian & Ferris, Inference of attitudes from nonverbal communication in two channels (Journal of Consulting Psychology, 1967). The book that launched the number is Silent Messages (1971). Wikipedia keeps a decent account of the mess under Albert Mehrabian.
For the good work by the same hand, the room next door but one is Exhibit № 37, and the machine running on it is Exhibit № 38.
One honest note about what you just did, since this is the room for it: six messages sorted by one person is a demonstration, not an experiment. We chose those six precisely because they split cleanly. That is exactly the kind of thing you should now be suspicious of, and you should be suspicious of it here too.