Building a Brand Voice That Sounds Like You (Not Everyone Else)

Why Brand Voice Is Your Last Real Competitive Moat

Almost everything else about your marketing can be copied by lunch. A competitor can match your pricing this afternoon, clone your landing page layout this week, and run the same Google Ads keywords tonight. What they can’t replicate is the way you sound — the specific personality that makes a customer read three sentences of an email and know it’s from you before they check the sender. That’s what a brand voice is: the consistent, recognizable personality in everything you write, from homepage headlines to the auto-reply that goes out when someone files a support ticket.

The business case is straightforward. Buyers see dozens of brands a day that all describe themselves the same way — innovative, customer-obsessed, passionate, results-driven. When every company sounds like the same press release, the one with an actual personality gets remembered, and memory is what gets you into the consideration set when the buying moment finally arrives. Voice also compounds: every blog post, ad, and email written in a consistent voice deposits into the same recognition account, while inconsistent copy starts the relationship over every time.

And yet most companies have never defined their voice at all. It lives in the founder’s head, approximated by whoever happens to be writing that day, drifting with every new hire and every new freelancer. The result is a website that sounds corporate, social posts that sound like a different company, and sales emails that sound like a third. This guide is the workshop framework we use to fix that: how to define voice and tone deliberately, build a messaging hierarchy on top of them, and document the whole thing so it survives contact with real deadlines.

Voice vs. Tone: One Personality, Many Situations

The two words get used interchangeably, and the confusion causes real problems, so pin them down first. Voice is your brand’s personality — stable, consistent, the same across every channel and every year. Tone is how that personality flexes for the situation. You are the same person at a wedding and at a funeral, but you don’t speak the same way at both. Voice is who you are; tone is how you adjust.

This distinction matters because the most common objection to defining a voice is some version of “but we can’t be playful in a billing dispute.” Correct — and you don’t have to be. A brand whose voice is warm and plainspoken stays warm and plainspoken when a customer is angry; it just dials the playfulness to zero and leads with accountability. The voice attributes don’t change. The intensity does.

Practically, that means your voice guidelines need two layers. The first defines the fixed personality: the three or four traits that show up everywhere. The second is a tone map: a short table of common situations — a launch announcement, an apology, a pricing page, a 404, a churned-customer winback — with notes on which traits to amplify and which to mute in each. A typical failure mode is documenting only the first layer, which leaves writers guessing in exactly the high-stakes moments where guessing is most expensive. The tone map is what turns a poster on the wall into an instruction you can follow at 4:55 p.m. on a Friday when the outage email needs to go out.

The Brand Voice Workshop: Who’s in the Room and What You Bring

Voice can’t be assigned to one copywriter to invent alone, and it can’t be decided by a committee of twelve. The workshop that works sits in between: a 90-minute to two-hour session with four to six people — a founder or senior leader who carries the company’s origin story, someone who talks to customers daily (sales or support, not just marketing), and whoever actually writes most of the public-facing words. Smaller companies can run it with two people and a whiteboard; the structure matters more than the headcount.

Do the homework before anyone enters the room, because the raw material for your voice already exists in two places. First, your own best moments: pull five to ten pieces of past copy that felt right — the email that got replies, the page a customer quoted back to you, the post that sounded like the founder on a good day. Pull the worst offenders too, the copy that makes everyone wince. Second, customer language: review transcripts, support tickets, sales call notes, and the words real customers use in Google reviews to describe you. When customers repeatedly call you “straight shooters” or say “they explained it without making me feel dumb,” that’s not feedback — that’s your voice, observed from the outside.

Bring competitor copy as well. Print the homepages of your three closest competitors with the logos removed and read them aloud. In most categories they’re interchangeable, and hearing that sameness out loud is the single most motivating moment of the workshop: the room realizes the bar for sounding distinct is on the floor, and the open territory becomes obvious.

The Workshop Exercises That Actually Produce a Voice

Four exercises, run in order, take the room from vague vibes to usable attributes.

Start with the brand-as-person exercise, because it bypasses corporate vocabulary. If your company were a person at an industry event, who are they? How do they dress, what do they joke about, what would they never say? Push past the flattering answers — everyone says “confident and knowledgeable” — by asking the inverse: what kind of person are we definitely not? Not the clown, not the professor, not the salesman working the room. Negative space defines a personality faster than positive claims do.

Second, run the admired-voices exercise. Ask everyone to name brands — from any industry, the further from yours the better — whose writing they actually enjoy, and read examples aloud. The point isn’t to copy them; it’s to extract what’s admirable. “They explain complicated things without jargon” and “they’re funny without trying too hard” are attribute candidates in disguise.

Third, do a card sort. Put forty or so personality adjectives on cards or sticky notes — warm, direct, irreverent, precise, scrappy, authoritative, dry, earnest — and have the group sort them into three piles: us, not us, and aspirational. The arguments are the deliverable. When sales insists the brand is “bold” and support insists it’s “gentle,” you’ve surfaced a real strategic tension that was already confusing your customers; resolve it in the room.

Finish by drafting against reality. Take one real, scheduled piece of upcoming copy — next week’s email, the new service page intro — and rewrite its opening lines live in the proposed voice. If the group can’t produce two sentences that sound like the attributes, the attributes are wrong, and it’s better to find out now.

Defining Voice Attributes: The “This, Not That” Test

The output of the workshop is three or four voice attributes. Not seven — nobody can write toward seven traits at once, and a long list quietly means no list. But a bare adjective is useless on its own, because every word has a failure mode. “Confident” curdles into arrogant. “Friendly” slides into unprofessional. “Direct” becomes blunt to the point of rude. So every attribute ships in three parts: the trait, the boundary, and an example.

The format that survives real-world use is “this, not that,” with a sentence of behaviour attached. Something like: confident, not cocky — we state recommendations plainly and own our calls, but we show the reasoning and admit what we don’t know. Or: plainspoken, not dumbed-down — we explain technical work in everyday words, but we never patronize a reader who knows the subject. Or: warm, not chummy — we write like a helpful colleague, not a buddy fishing for a reply. The boundary word does most of the work; it tells a writer under deadline exactly where the line is.

Then attach a real example to each attribute: one sentence of on-voice copy and one sentence of off-voice copy, ideally drawn from your own past writing. Writers don’t internalize adjectives — they internalize examples. A junior hire or a new freelancer should be able to read your three attributes with their paired examples and produce passable on-voice copy on day one, without a single meeting. That’s the test of whether the attributes are specific enough: if they require oral tradition to interpret, they aren’t done.

Messaging Hierarchy: Deciding What You Say, Not Just How

Voice is how you say things; the messaging hierarchy is what you say, in what order, and it belongs in the same document because the two fail together. A distinctive voice delivering scattered messages is a charming person with nothing to say.

The hierarchy has three levels. At the top sits the core message: the one sentence you’d keep if you could only keep one — who you serve, what you change for them, and why you over the alternatives. It’s not a tagline and it’s rarely customer-facing verbatim; it’s the internal sun the rest of the copy orbits. If your team can’t recite it, your customers certainly can’t infer it.

The second level is three to five supporting pillars: the recurring arguments that prove the core message. For a service business these are typically things like speed of response, transparency of reporting, depth in a niche, or pricing without surprises — each pillar paired with the proof behind it (the process, the guarantee, the numbers you genuinely have). Pillars are where most messaging discipline lives: every page, campaign, and deck should lean on one or two pillars, not all five at once, and nothing important should rest on a message that isn’t a pillar.

The third level is proof points and language for each pillar — the specific phrases, examples, and details writers can pull from. This level is also where you keep the banned list: the empty intensifiers and category clichés you’ve agreed never to use. The hierarchy’s payoff is consistency at the level buyers actually feel. Voice makes your emails sound related; the hierarchy makes your website, your sales deck, and your proposal argue the same case — which is the difference between a brand and a pile of content.

Writing Voice Guidelines People Actually Open Twice

The workshop’s findings die in two formats: the 60-page brand book nobody reads, and the inspirational one-pager too vague to apply. The usable middle is a short, example-heavy working document — typically four to eight pages or the equivalent in a shared doc — organized for a writer with a deadline, not for a brand director with an afternoon.

The structure that earns repeat visits: the voice attributes with their this-not-that boundaries and paired examples; the tone map for high-stakes situations; the messaging hierarchy with pillars and proof points; a short mechanics section covering the practical choices that cause daily debates — contractions or not, sentence length habits, how you refer to customers, emoji policy, capitalization of product names; and a before-and-after gallery, real past copy rewritten into the voice, because side-by-side rewrites teach faster than any rule. End with the banned list. Writers love a banned list; it’s the fastest rule in the book to apply.

Two properties separate guidelines that get used from guidelines that get archived. First, searchability — a living shared document beats a designed PDF, because the writer mid-draft needs to find “how do we sign off emails” in ten seconds. Second, honesty about edge cases: include a short FAQ of the genuinely hard calls, like how the voice handles a price increase or responds to a public complaint, with worked examples. The high-stakes copy is exactly where teams abandon the voice and retreat into legalistic mush, and it’s where a customer is paying the most attention. If the guidelines only cover the easy moods, they’ll be ignored in the moments that matter most.

Keeping Your Voice When AI Writes Half the Internet

There’s a new reason the moat metaphor has teeth. AI writing tools have made average copy effectively free, which means the volume of competent, beige, interchangeable content in every market is exploding. The default output of these tools is the statistical centre of all writing — polished, agreeable, and identical to everyone else’s default output. If your brand publishes unedited generic AI copy, you are paying to sound like everyone, which is the opposite of what content is for.

This doesn’t mean refusing the tools; it means your voice guidelines become the input. A well-built voice document — attributes with boundaries, on-voice and off-voice examples, the banned list, the messaging pillars — converts almost directly into instructions an AI assistant can follow, and the difference between prompted-with-a-voice-doc and prompted-cold output is dramatic. Teams that invested in defining their voice now get more leverage from the same tools than teams that didn’t, which is a pleasant inversion: the most human asset you own is also the thing that makes the machines useful.

The discipline that has to survive is the edit. Treat AI drafts the way you’d treat a competent freelancer’s first pass: the structure may be fine, but the voice pass — sharpening the openings, cutting the hedging, injecting the specific opinions and examples only your team has — is where the brand happens. A practical habit is to keep humans on the copy where voice carries the most weight: homepage and pricing pages, apologies and incident comms, founder emails, anything a customer reads at a moment of high emotion. Let the tools accelerate the middle; keep the voice-critical edges handwritten.

Rollout, Governance, and Catching Voice Drift Early

A voice that exists only in a document is a voice that doesn’t exist. Rollout starts with a working session, not an email attachment: walk the team through the attributes, then have everyone rewrite a piece of real current copy in the voice and compare results. An hour of practice builds more fluency than a quarter of reminders. Give the guidelines an owner — one named person who answers judgment calls, approves additions to the banned list, and keeps the examples fresh — because unowned standards decay at the speed of staff turnover.

Then sequence the rewrite by exposure rather than attempting everything at once. A typical order: homepage and top service pages first, then the email templates and automated sequences customers actually receive, then ads and social templates, then the long tail of support macros and proposals. The automated copy deserves more priority than it usually gets — onboarding emails, invoices, confirmation messages — because it’s sent thousands of times and almost always sounds like it was written by the software vendor rather than by you.

Finally, audit for drift on a schedule. Twice a year, pull one recent example from every channel — a blog post, a sales email, three social posts, an ad, a support reply — strip the logos, and read them in one sitting. Ask the only question that matters: would a customer know these came from the same company? Score each against the attributes, note where the voice is eroding and which new situations the tone map doesn’t cover, and update the document. Voice isn’t a project that ends; it’s an asset you maintain. The brands you can recognize from a single sentence got there exactly this way — they decided who they were, wrote it down, and refused to drift.