Brand Voice

Origin & evolution

The idea that a brand has a personality predates the writing discipline, but the working separation of voice from tone is what made it usable for content teams. The clearest public articulation is Mailchimp's Content Style Guide, published openly so other teams could borrow it. Its formulation has become the reference point: the voice stays roughly the same from day to day; the tone changes all the time depending on the situation and the reader's emotional state.

Mailchimp defines its own voice as plainspoken, genuine, and a little dry-humoured, then shows the voice flexing into different tones across contexts. A celebratory tone when a user sends their first campaign. A plain, reassuring tone when a payment fails. Same personality, different register, chosen to match how the reader is feeling. That table of situations, each with a recommended tone, turned an abstract trait into something a writer can apply at three in the morning under deadline.

The practical instrument most teams adopt is the voice chart and its companion, the "we are / we are not" table. The voice chart lists each personality trait beside a do-this, avoid-that pair, so "confident" gets pinned down as "state things plainly" against "hedge and over-qualify." The "we are / we are not" table draws the boundary by contrast, naming the adjacent personality the brand keeps mistaking itself for. Without those guardrails, voice guidelines collapse into a list of flattering adjectives that no two writers interpret the same way.

How it works in practice

A developer-tools company finds its docs, its changelog🕰️ChangelogProduct SpecificationA record of changes shipped in a releaseView reference →, and its marketing site sound like three different companies. The docs are terse to the point of curt, the marketing is breathless, and the changelog is robotic. They write a voice chart with four traits: direct, technically precise, warm without being chummy, never condescending. Each trait carries a do/avoid pair and a real before-and-after rewrite pulled from their own copy.

The "we are not" column does the heavy lifting. "We are warm, we are not your buddy" kills the forced exclamation marks in marketing. "We are precise, we are not pedantic" stops the docs from explaining what a variable is. Tone still varies: an outage post-mortem is sober and specific, a feature⭐FeatureProduct SpecificationA product capability or featureView reference → launch carries a little lift. The personality underneath stays fixed, so a reader moving from the marketing site into the docs feels one company instead of three.

Brand voice vs. its neighbours

• Brand voice vs. tone is the load-bearing distinction. Voice is the constant personality; tone is the situational register that personality adopts. One brand, one voice, many tones. Treat them as the same thing and the guidelines fail in exactly the moments they were written for.
• Brand identity💠Brand IdentityBrand IdentityThe root brand identity entityView reference → is the full system of signals, visual and verbal, that a brand projects. Voice is the verbal slice of that system, sitting alongside colour and typography. A wordmark and a sentence are doing the same recognition work in different media.
• Content strategy✍️Content StrategyGo-To-MarketA content strategy for thought leadershipView reference → governs what gets published, for whom, and why. Voice governs how it sounds once written. Strategy decides the article should exist; voice decides whether it reads like the brand wrote it.

In the graph

In the Unified Product Graph, brand_voice🎙️Brand VoiceBrand IdentityBrand voice and tone guidelines sits in the brand domain and connects upward to the identity it serves and outward to the words it shapes. The edge brand_identity_speaks_with_brand_voiceBrand Identityspeaks withBrand Voicehierarchy makes voice a defined facet of the brand's projected self, the verbal counterpart to colour and type. The edge messaging_aligns_with_brand_voiceMessagingaligns withBrand Voicecross-domain ties the abstract personality to concrete claims, so a positioning🎪PositioningGo-To-MarketProduct positioning statementView reference → line or a campaign message can be checked against the voice that is meant to carry it. Holding voice as its own node, with its own edges, lets a team query every piece of messaging✉️MessagingGo-To-MarketMessaging framework and key messagesView reference → that should conform to it and catch the surfaces, like a neglected changelog, that quietly drifted off. Buried as a property inside identity, voice would be far harder to audit that way.