No invented statistics. Every number is sourced or labelled as an example. Our standards

Myth Busting

Myth: "93% of communication is non-verbal" — what Mehrabian actually measured

The claim:
Only 7% of communication is the words you use; 38% is tone of voice and 55% is body language, so in a sales meeting what you say barely matters.
Misapplied

Anyone who's been on a presentation skills course has probably seen the pie chart: 7% words, 38% tone of voice, 55% body language. In sales training it's often used to argue that your handshake, posture and eye contact matter far more than your proposal, your pricing or your answers to objections. Unlike many sales statistics, this one does come from real, published research by a named psychologist. The trouble is what it gets used for. If you take it literally, you'd conclude a prospect could follow a sales call with the sound off. Anyone who's tried to explain a pricing model in mime knows that's not true.

Where the claim comes from

The figures come from Albert Mehrabian, a psychologist and emeritus professor at UCLA. In 1967 he published two papers with colleagues. Both are listed in his own publications list.

He brought the findings together in his book Silent Messages, first published in 1971, with a second edition in 1981. That book is where the 7-38-55 split reached a wide audience.

What we could actually find

This is one of the rare sales myths where we can go straight to the researcher. Mehrabian's own page about Silent Messages sets out the formula as Total Liking = 7% Verbal Liking + 38% Vocal Liking + 55% Facial Liking. Directly under it, he explains that this equation and others like it came from experiments on how people communicate feelings and attitudes, specifically liking and disliking. He says that unless someone is talking about their feelings or attitudes, the equations don't apply. The same page describes the book's discussion as being about inconsistent messages, where someone's words say one thing and their voice or face says another.

So the formula is about something quite narrow. It measures how much a listener relies on each channel when judging whether a speaker likes or dislikes something, in situations where the channels conflict.

The experiments were narrow too. According to Olivia Mitchell's detailed write-up and the Wikipedia summary, both studies used single words rather than conversations. In the first, listeners heard words spoken in a tone that could clash with the word's meaning, and rated the speaker's feelings. In the second, one fairly neutral word, "maybe", was paired with photographs of facial expressions. The 7-38-55 ratio came from combining the results of the two studies. None of it involved a real conversation, a decision, or anything like a sales meeting.

Mehrabian has also spoken about how the finding gets misused. Olivia Mitchell reproduces an email he sent to Max Atkinson in 2002, which Atkinson published in his book Lend Me Your Ears. In it, Mehrabian says he's uncomfortable with how his work is misquoted. He says he has tried from the start to explain the limits of his findings, and he criticises self-styled image and leadership consultants with little psychology training for spreading the wrong version. We haven't seen Atkinson's book ourselves, but it matches the caveat on Mehrabian's own site.

That's why our verdict is Misapplied. The numbers are real and properly published. The claim that they describe communication in general, or the persuasiveness of a sales pitch, is something the researcher himself has explicitly rejected.

What the evidence does support

There's a genuinely useful lesson in the research once you put it back in context. When someone's words and delivery don't match on how they feel, listeners tend to believe the tone and the face over the words. In sales, that's relevant in two places.

What the research doesn't show is that words carry 7% of the meaning in a sales conversation. Your explanation of the problem, the numbers in your proposal and the answers you give to hard questions are the content. The research doesn't say they barely matter, and nothing in it covers that kind of communication.

What to do instead

If you train others, you can still use Mehrabian. Just use the accurate version: when feelings and words don't match, people trust the tone. It's more defensible, and more useful, than the pie chart.

Sources we checked

Spotted a mistake or a number we can't back up? Tell us and we'll fix it.