Your brand voice document is a list of adjectives.

Open the brand voice section of almost any guidelines document and you'll find the same four words wearing different outfits. Confident. Authentic. Approachable. Bold. Sometimes there's a table: "We are confident, but not arrogant. We are playful, but not silly." The team that wrote it felt productive. The copywriter who inherited it got nothing.

Here's the test I run on every voice document I'm handed: could this describe your biggest competitor too? If yes, and it's almost always yes, you don't have a voice. You have a horoscope. It's written vaguely enough that everyone sees themselves in it.

Adjectives describe. They don't decide.

The job of a voice framework isn't to describe how the brand sounds. It's to help a person who isn't you make the same decision you would make, on deadline, without asking. That's the whole job. A word like "authentic" decides nothing, because no writer has ever sat at a keyboard trying to be inauthentic.

The decisions that actually come up are specific. Do we open the email with the offer or with the story? Do we make the joke? How do we tell a customer no? Can a headline be a sentence fragment? What do we sound like when something breaks? An adjective list answers none of these. So writers guess, every guess drifts a little differently, and eighteen months later someone declares the brand "inconsistent" and commissions a new adjective list.

Build spectrums, not synonym clouds

When I build voice frameworks, the core is a set of spectrums with a marked position. Not endpoints, positions. Formal to conversational: we sit here, and here's a sentence from either side of the line so you can see the edge. Reverent to irreverent: here. Plainspoken to lyrical: here, except in these two moments, where we're allowed to move.

The position matters less than the boundary. Writers don't need to be told what the middle of the brand sounds like. They can feel that from examples. They need to know where the line is, because the line is where every real argument happens. The Slack debate is never "should this sound confident?" It's "is this joke too much for a shipping-delay email?" A good framework settles that argument before it starts.

The before-and-after is the framework

The most useful page in any voice document I've ever shipped is the rewrite page. Take five real pieces of copy (an email subject line, a product description, an error message, a social caption, a headline) and show each one written wrong, then written right, with one sentence on what changed. Not invented examples. Real ones, pulled from the brand's own channels, because the gap between the brand's actual output and its aspirational voice is the entire reason the document exists.

When I did voice work on brand platforms at INDUSTRY, the rewrite pages got used long after the strategy decks were filed away. Writers don't reread philosophy. They pattern-match from examples. Give them patterns worth matching.

Voice is a set of refusals

One more thing the adjective list can't do: say no. A voice gets distinct not by what it claims but by what it refuses. No exclamation points. No puns in product copy. Never open with "we're excited to announce." We don't call customers "users." The refusals are falsifiable (you can audit a page of copy against them) and that's exactly what makes them useful.

So if your voice document fits on a mood board, rebuild it. Spectrums with positions. Real rewrites. A short list of refusals. It's less inspiring to read than "bold, but warm." It's also the only version anyone can actually write with.