Tone-of-voice documents have a familiar moment.
After the research, the workshops and the careful discussion about personality, the brand eventually becomes:
Confident. Human. Clear. Warm. Expert.
None of these words are wrong.
That is part of the problem.
Most companies would happily describe themselves with some version of the same five adjectives. The words sound sensible in a presentation, but they rarely help when somebody actually has to write something.
How warm?
How confident?
What happens when confidence starts sounding arrogant? What happens when “human” becomes a caption full of forced jokes? What happens when clarity requires saying something the company would rather soften?
A useful brand voice should help answer those questions.
It should tell us how the brand behaves when the situation changes, not only how it sounds in the paragraph explaining the brand voice.
A voice needs range
People do not speak in one register all day.
We do not use exactly the same language in a client presentation, over dinner with an old friend, in a family WhatsApp group and when explaining to a bank that they have made an error.
The personality remains broadly the same. The context does not.
A brand should be allowed the same intelligence.
This is where many tone-of-voice systems become too narrow. They define an ideal mood rather than a usable range.
The brand is warm.
Fine. But what does warm sound like when a customer is angry?
The brand is playful. What happens when the company has made a mistake?
The brand is expert. Can it explain something complicated without sounding as though it is trying to win an argument?
And if the brand has nothing useful to add, is it allowed to be quiet?
That last one matters more than most tone-of-voice documents admit.
A real voice has to survive more than the launch campaign.
Same person, different room
Erving Goffman's work on self-presentation is useful here. People adjust how they present themselves according to the social situation and the audience in front of them. The internet later gave us a related problem: context collapse, where very different audiences are flattened into one imagined crowd.
Brands do this to themselves all the time.
They invent an enormous abstract creature called “our audience” and then try to speak to it everywhere.
The same sentence moves from LinkedIn to Instagram to X. The crop changes. The character count changes. The behaviour does not.
But the person seeing the brand may be in a very different mode each time.
The same person can want detail in one place, distraction in another and a quick answer somewhere else. They can tolerate a long explanation on Tuesday morning and have absolutely no patience for one on Friday night.
This does not mean every platform needs a different personality. It means context belongs inside the communication decision.
Consistency should protect the character of the brand, not freeze its behaviour.
The safest voice is usually the least interesting one
There is another reason brands begin to sound the same: organisations are very good at sanding off variation.
One draft becomes three rounds of comments. Three rounds become six people on an email. Legal removes the risky word. Leadership softens the opinion. Marketing adds the product message back in. Someone asks whether the sentence could be “a little more engaging”.
By the end, nobody strongly objects to it.
Nobody strongly feels anything about it either.
This is not because approval is unnecessary. Brands need standards, accuracy and judgement. The problem is what happens when “safe” becomes the only principle capable of surviving the process.
When many different audiences and internal stakeholders have to be satisfied at once, language tends to drift towards the safest common denominator.
The result is familiar corporate communication: nothing incorrect, nothing offensive, nothing particularly alive.
A strong brand voice needs enough definition to survive approval without being averaged into neutrality.
That requires more than adjectives.
It requires decisions.
The centre has to be stronger than the rules
The most useful voice systems are clear about what must remain stable and what is allowed to move.
The stable part is not a list of preferred phrases.
It is the brand's point of view. Its standards. What it knows. What it believes. What it refuses to pretend. The level of intelligence it assumes in the audience. The kinds of claims it will and will not make.
Once that centre is strong, quite a lot can change around it.
Formality can move.
Pace can move.
Humour can appear and disappear.
A subject can be explained in depth in one place and reduced to one sharp observation in another.
The brand can be serious without suddenly sounding like a legal notice. Informal without pretending to be eighteen. Funny without arriving at a meme after the meme has already filed for retirement.
This is why consistency and sameness should not be treated as synonyms.
Without a strong centre, consistency requires rules.
With a strong centre, it can survive variation.
That is a more demanding system to build, but it is also a much more useful one.
A voice should be tested under pressure
One of the easiest ways to find out whether a tone-of-voice system is real is to stop testing it on easy copy.
Do not start with the homepage headline.
Try the difficult email.
The service failure.
The price increase.
The complicated product explanation.
The post where the company has a genuinely strong opinion.
The response to criticism.
The announcement that should be clear but not celebratory.
The moment when humour would be completely inappropriate.
If the only guidance available is “warm, confident and human”, somebody still has to invent the answer from scratch.
This is why examples matter more than adjectives.
Not polished sample captions designed to make the document look finished. Real tensions.
How direct are we willing to be?
What do we never joke about?
When do we explain and when do we get out of the way?
How much expertise can we assume?
What does confidence sound like when the answer is “we got this wrong”?
Those decisions are where a voice becomes operational.
Platform behaviour comes later
This is also why we prefer not to begin brand voice work with a channel matrix.
“LinkedIn is professional. Instagram is visual. TikTok is playful.”
That sort of guidance ages quickly and usually tells you more about the stereotype of the platform than the brand.
The better sequence is the other way around.
First, understand the character.
Then understand its range.
Then decide how that character behaves in different environments.
A research finding might become a detailed argument on LinkedIn, a visual explanation on Instagram and a longer conversation on YouTube. The idea remains connected, but the expression changes because the context changes.
Sometimes the adaptation is substantial.
Sometimes almost nothing needs to change.
And sometimes the correct version of a piece of content for another platform is not to publish it there at all.
Calendars tend not to enjoy that answer.
Strategy sometimes does.
A brand should sound like somebody, not like a template
At Recreative Digital, we are less interested in giving brands a perfectly controlled “tone” than in helping them develop a recognisable way of thinking and behaving.
The distinction matters.
A tone can become a setting.
A voice has judgement.
It knows when to explain, when to simplify, when to push, when to soften and when to leave the sentence alone.
It can change its volume without changing its character.
That is what allows a brand to appear in different places, speak to different people and deal with different situations without sounding as though a new personality has arrived every time.
The goal is not consistency for its own sake.
The goal is recognition.
If your brand can only sound like itself in one mood, it does not really have a voice.
It has a setting.