Why ‘93% of communication is non-verbal’ is misleading

Imagine a team getting ready to announce a new staff expenses policy. Someone has drafted an email explaining when the change takes effect and how to submit a claim.

A colleague suggests a video of the manager instead.

'Words are only 7% of communication,' they say. 'The other 93% is how you say it.'

On that reasoning, a written announcement would leave most of its meaning out.

The claim combines 2 figures: 38% for tone of voice and 55% for facial expression. Add them together and you get 93%. Alongside the 7% assigned to words, these make up the 7-38-55 rule.

Those figures were proposed to describe how people judge a speaker's attitude: whether the speaker seems to like or dislike the person they are addressing. They do not measure how much of a message someone understands.

That distinction matters for the announcement. A manager might come across as welcoming while staff remain unsure how the new policy works. The research does not show that an email loses 93% of its meaning, or settle whether this team needs a video.

What the studies asked people to judge

In their 1967 study, Albert Mehrabian and Morton Wiener combined individual words with different tones of voice. Participants rated the speaker's attitude towards an imagined addressee. A positive word could be spoken in a negative tone, or a negative word in a positive tone.

A separate 1967 study by Mehrabian and Susan Ferris paired photographs of facial expressions with recordings of the word 'maybe'. The participants again judged the speaker's attitude towards another person.

Both experiments concerned inferred liking or disliking. They examined how people interpreted selected cues when making that judgement. Comprehension, recall and the ability to follow instructions were outside the measured outcome.

Those distinctions matter for the policy announcement. Staff might form an impression of their manager's attitude while still misunderstanding when the new process applies. Whether they perceived goodwill would not establish whether they understood the rule.

How attitude becomes a percentage of meaning

The familiar coefficients appear in the discussion of the Mehrabian and Ferris paper. The authors propose combining findings from the 2 studies into a model of simultaneous verbal, vocal and facial attitude communication. They did not run one experiment in which all 3 varied together.

Mehrabian's own explanation describes the equation in terms of liking: 7% verbal, 38% vocal and 55% facial. He limits its application to communications about feelings and attitudes.

The hypothetical brief changes what those coefficients describe. Relative contributions to a judgement of attitude become shares of everything a message means. Removing voice and expression is then assumed to leave writing with only 7% of the communication.

That calculation requires evidence the experiments do not provide. There was no measured total of message meaning from which the researchers removed voice and facial expression to see what remained in writing.

Attitude can be part of a message's meaning. Delivery can affect how words are interpreted, so a correction to the percentage claim is no reason to dismiss tone or expression. The limit is that one kind of judgement has been used to represent the whole communication.

For the expenses announcement, the brief would still have to explain what staff would lose in a written version. The percentages cannot identify an unclear instruction, an unanswered concern or a reason to hear directly from a manager.

The expenses announcement

In the expenses example, a new claim form takes effect on 1 December. Staff must use it for claims submitted from that date. The announcement needs to explain where to find the form and how to submit it.

A manager may also want to explain the reason for the change and answer questions about it. There could be a useful purpose for a briefing. Someone might need to ask how the rule applies to their situation, or want to hear the manager's response to a concern.

Those purposes deserve consideration in the brief. Calling the message '93% non-verbal' does not establish that a video will serve them. A recording of a manager, for example, does not itself provide an opportunity for staff to ask questions and receive an answer.

The information requirement also remains. Staff may need to check the effective date or the submission process later. The brief should say how they will find those details, including if a video or conversation is part of the proposal.

For this example, I would keep a written reference for the rules and consider a briefing if its purpose were clear. The revised brief would need to name what the manager should explain and where staff can check how to submit claims from 1 December.