Search “brand voice examples” and you’ll get the same twelve logos on every site. Mailchimp. Oatly. Apple. Wendy’s Twitter. Each one described in three adjectives: witty, bold, authentic.
That’s not a voice analysis. That’s a vibe mood board.
The reason most brands can’t replicate what those companies do, even after studying them, is that adjectives don’t explain mechanisms. “Witty” doesn’t tell you how to write a subject line. “Bold” doesn’t tell you what to do when a client asks you to soften the copy.
Turning those mechanisms into something repeatable is what a good brand voice template is for: rules and examples, not adjectives.
Voice is a system. Here’s how to actually read it.
What separates real voice from performed tone
The brands people actually remember don’t have a tone. They have a point of view, a belief about their category that shows up in everything they publish, whether it’s a tweet, a job listing, or a terms of service page.
That’s the test. Not “does this sound fun?” but “could only this brand have written this?”
There’s a useful tool for this: the logo removal test. Take a piece of content, strip the brand name and logo, and ask a stranger who wrote it. If they can’t tell, the voice isn’t working. If they say “that sounds like Oatly” or “that’s definitely Basecamp,” the voice is a real asset.
Most brands fail the test. Not because they haven’t thought about voice, but because they’ve confused tone (how you say things) with POV (what you actually believe). Tone without POV is costume jewelry. It looks fine until someone looks closely.
Let’s run nine brands through the real analysis: what makes each one work (or not), what the actual mechanism is, and what you can steal.
The brands that pass
1. Oatly: category challenger
POV: The dairy industry is environmentally indefensible and everyone knows it and nobody’s saying it out loud.
Signature move: Breaks the fourth wall. Talks about the ad being an ad. Makes the packaging argue with you.
Vocabulary: Post-milk. Wow no cow. It’s like milk, but made for humans. They coin terms competitors can’t use without referencing Oatly.
What they’d never say: “A delicious dairy alternative for the health-conscious consumer.” That sentence belongs to a committee. Oatly fires committees.
The test: Pull any Oatly sentence. Could a competitor publish it? No, because no competitor is willing to hold the same belief. The voice is downstream of the belief, not the other way around.
What to steal: Pick the belief your competitors are too scared to state. Then state it in every channel, not just advertising.
2. Linear: opinionated tool
POV: Software should have opinions. Tools that try to work for everyone work for no one.
Signature move: States the principle before naming the feature. Never leads with what it does. It leads with why the approach exists. Every product decision is framed as a philosophical position.
Vocabulary: Clarity. Speed. Craft. Opinionated. They have a word list that other PM tools would find arrogant, and that’s exactly the point.
Sentence structure: Short. Declarative. No hedging. “We build for the people who care about their craft.” Not “We believe we’ve built a product that may be useful for teams who prioritize quality.”
The test: Could Jira write a Linear sentence? Not without sounding like it’s trying to be Linear. Their voice is only available to them because it requires believing things Jira’s product org would never sign off on.
What to steal: Stop softening your claims. “We think this might help” is not a brand voice. “Here’s what we believe and why” is.
3. Basecamp: mission-led
POV: The hustle culture around SaaS growth is actively harmful and we built the opposite of it.
Signature move: Argues with conventional wisdom in the first paragraph. Never starts neutral. Every piece of content positions itself against something: VC culture, growth-at-all-costs, the “move fast” mentality.
Consistency as the test: Their pricing page has a POV. Their job listings have a POV. Their books have a POV. Their Twitter has a POV. The 404 page probably has a POV. That’s what consistency actually means: not using the same color palette, but holding the same belief everywhere.
What it costs them: They alienate VCs and growth-obsessed buyers deliberately. The voice isn’t for everyone. Brands that try to be for everyone end up sounding like HubSpot.
What to steal: Find the thing your ideal customer believes but almost never hears a vendor say out loud. Say it. On every channel. In every format.
4. Patagonia: enemy-defined
POV: The apparel industry is one of the most destructive forces on the planet. We’re part of it, and that fact should make everyone uncomfortable, including us.
Signature move: Turns self-criticism into authority. The 2011 “Don’t Buy This Jacket” ad didn’t hurt sales. It made the voice bulletproof. A brand willing to say that is a brand you can trust to mean everything else it says.
Vocabulary: Worn wear. Burden of ownership. Environmental cost. They don’t use sustainability marketing language. They use environmental accountability language, and the difference runs deep.
The enemy: Patagonia writes like it has a named enemy: fast fashion, extraction capitalism, the idea that consumption is neutral. Their audience doesn’t need to know the enemy’s name. They feel it in every sentence.
What to steal: Name your enemy. Not a competitor, but a behavior or belief your customers hate. Then write like you’re fighting it.
5. Liquid Death: character-driven
POV: Canned water is absurd. Wellness marketing is absurd. We’re going to be absurd at each other and everyone will have a better time.
Signature move: The mock-metal aesthetic isn’t the voice. It’s the costume. The actual voice mechanism is commitment to the bit. They never break character. The tagline is “Murder Your Thirst.” The product is water. The commitment to that gap is the brand.
What it requires: A character-driven voice only works if leadership defends it when a brand manager gets nervous. Liquid Death has been clear it would rather stay uncomfortable than become “approachable,” and that conviction is what makes the voice real.
The danger: Character-driven voices are the most fun to copy and the hardest to sustain. Without a founding-level commitment to the character, they drift into “quirky” and die.
What to steal: Pick one character trait and commit to it past the point of comfort. Quirky content that hedges is worse than no character at all.
6. Stripe: precision as voice
POV: Most infrastructure is bad because it was built by people who didn’t care about the developer experience. We care about it the way architects care about buildings.
Signature move: Uses complexity to signal respect. Where most B2B SaaS simplifies for fear of losing the reader, Stripe writes like the reader is intelligent. The documentation is the brand. The developer who reads a Stripe API doc doesn’t think “this is well-written.” They think “this company gets me.”
Vocabulary: Precise, technical, unhedged. They name things what they are. No “powerful” or “comprehensive.” Just: here’s what it does, here’s how, here’s the edge case.
The result: Stripe’s voice made paying for infrastructure feel like belonging to something. That’s not small. That’s the reason developers fight for Stripe in their org.
What to steal: Precision is a voice. Saying exactly what something does, without fluff, without hedging, is rarer and more distinctive than most brands realize.
7. Duolingo: emotion-led
POV: Learning a language is hard, boring, and you keep quitting. So we’ll make ourselves impossible to ignore. The character will haunt you until you come back.
Signature move: The owl guilt-trips you. That’s the whole mechanic. Most brand mascots are friendly. Duo is passive-aggressive. “You said you’d practice today.” That’s a sentence no other ed-tech brand would publish, because it requires believing that friction can be charming if the character is strong enough.
What makes it work: Consistency between channels. The app, the push notifications, the social media, and the brand campaigns all run the same emotional logic. The voice didn’t live in a brand deck. It lived in the product.
The risk they took: Annoying people on purpose. Some users hate the notifications. That’s fine. The ones who find it funny are the ones who stick around and tell people about the app. Voice should polarize slightly. Unanimous approval means the edges are too soft.
What to steal: Let your voice live in the product, not just the marketing. Push notifications, error messages, empty states: these are voice opportunities that most brands leave blank.
8. Notion: community brand
POV: Software should be a canvas, not a cage. We’re not going to tell you what your workflow looks like.
Signature move: Reflects the user back at themselves. Notion’s marketing doesn’t say “here’s how Notion works.” It says “here’s what people like you have built.” The voice is warm, curious, slightly in awe of its own users, which creates a specific feeling: I should be building something too.
What’s interesting: Notion’s voice is less about what the company believes and more about what the community has become. That works for community brands, but it’s a fragile model. When the community’s identity shifts, the brand has to shift with it.
The vocabulary: Build. Create. Your workflow. Personal OS. Second brain. They borrowed language from their community and made it feel official, which deepened community ownership.
What to steal: Find the language your most vocal customers use to describe what your product does. Use it before they expect you to. It signals: we’re listening.
9. HubSpot (post-scale): committee brand
POV: None detectable.
Vocabulary: Comprehensive. Helpful. Professional. All the words that mean nothing because everyone in the category uses them.
How it happened: HubSpot had a strong POV when it was small. They invented “inbound marketing” and held a genuine belief that interruption advertising was broken. That was a real position. As the company scaled and needed to serve a broader market, the edges softened. The vocabulary expanded. The belief got diluted into a content program. The voice became a content factory.
The test: Swap their homepage copy onto a competitor’s site. It fits. That’s the failure, not that the content is bad, but that it’s interchangeable. The logo removal test reveals nothing distinctive to remove.
What to learn from it: Voice erosion is the default trajectory for growing companies. It takes active effort to preserve the POV as headcount and stakeholder count grow. Most brands don’t do that work. Most brands sound like HubSpot.
The warning: Every committee brand used to have a voice. The committee is what happened to it.
What the ones that work have in common
Every brand with a voice people actually remember has the same thing: a controversial belief stated clearly enough that someone could disagree with it.
Oatly’s belief makes dairy companies uncomfortable. Linear’s belief makes product managers who love Jira roll their eyes. Basecamp’s belief makes VCs close the tab. Patagonia’s belief makes fast fashion executives defensive. Liquid Death’s belief makes wellness brands confused.
That discomfort is the voice working. A voice that everyone agrees with isn’t a voice. It’s a press release.
The second pattern: the voice lives in the product, not just the marketing. Stripe’s documentation. Duolingo’s push notifications. Basecamp’s pricing page. Oatly’s oat milk carton. The brand didn’t apply a voice to their content. They built a product that expressed the voice.
The third: they know exactly who they’re not writing for. Liquid Death isn’t for people who find their packaging off-putting. Linear isn’t for teams who want infinite configurability. Basecamp isn’t for growth-stage startups optimizing for scale. Knowing your anti-audience is what gives you permission to hold a position. If you need to convert everyone, you can’t say anything real.
The question isn’t “what adjectives describe our tone?” The question is: what do we believe about our category that our competitors would publicly disagree with, and are we willing to say it everywhere?
How to read any brand voice in 10 minutes
If you want to understand how any of these brands actually work, not what adjectives describe them, but what mechanisms create them, read their content with these five questions:
1. What do they believe that their competitors would dispute? This is the POV. If nothing comes to mind, there isn’t one yet.
2. What do they call things that no one else calls that? Owned vocabulary is a voice asset. “Post-milk generation.” “Murder your thirst.” “Personal OS.” These aren’t slogans. They’re category claims in disguise.
3. What would they never say? The negative space of a voice is often more instructive than the positive. Linear would never write “we’re excited to announce.” Patagonia would never write “treat yourself.” The things they refuse define the edges.
4. Does the voice show up in the product, or only in the marketing? Marketing-only voice is vulnerable. Product-embedded voice is the real moat.
5. Could this have been written three years ago and still sound like them? Voices that age well have a position, not a trend. Voices built on trend vocabulary expire with the trend.
Run any brand through those five questions and you’ll know within ten minutes whether the voice is real or performed.
Why AI makes this harder, and more important
Every company is now generating more content than their team could have written five years ago. The volume problem is solved. The sameness problem just got worse.
Nearly every entry in this roundup of AI writing tools is built to solve that volume problem; not one of them fixes sameness on its own.
When everyone uses the same AI tools with no voice instructions, the output converges. The same sentence structures. The same opener patterns. The same word choices that exist because they appear frequently in training data, not because they’re interesting or true to any specific brand.
Search engines will wave that convergence through, which is the uncomfortable part: Google won’t downrank your AI content, your readers will.
The brands that looked like Oatly or Linear before AI look like everyone else now, unless they’ve done the work to capture their voice in a form that can actually instruct the tools they’re using.
Capturing the voice starts a layer down, with the brand personality framework underneath it: the belief and the enemy that make a brand disagreeable on purpose.
Adjective lists don’t do it. A PDF brand guide doesn’t do it. What works is a machine-readable voice profile: the POV, the vocabulary list, the sentence patterns, the banned words, the specific moves that only this brand makes. Give that to any AI tool and the output shifts. Without it, the output defaults to “sounds like the average of the internet.”
That is also why a PDF brand guide stalls where a usable one travels; these brand guidelines examples show which rules a model can actually read.
That’s the gap. The brands that close it first will be the ones people study in five years the way we study Oatly now.
For a B2B team shipping at that volume, the same discipline is what a real B2B content marketing strategy is built on: a distinct voice at scale, not more undifferentiated output.
Find your brand's POV.
Boombrand runs you through the voice discovery process: research first, then a targeted conversation, and outputs a machine-readable voice profile you can give to any AI tool. The result sounds like you, not like everyone else.
Discover your voice →
