How to reverse engineer a client's brand voice in four steps

Executive overview

Copywriters lose clients not because of weak strategy or poor conversion techniques — but because the copy doesn't sound like the client. The client can't articulate why, which makes the feedback impossible to act on.

A four-step reverse engineering process — collect, measure, analyse, document — gives you objective data on what a client's voice actually is, so you can match it precisely and defend your choices.

Measuring sentence length, grade level, and tone with free tools turns "sounds like me" from a gut feeling into a replicable system.

Voice vs tone vs messaging

• Voice is brand personality — how a brand communicates, consistently, across all channels.
• Tone shifts by context; voice stays fixed.
• Messaging is what you say; voice is how you say it.
• Voice is made up of four elements: tone, sentence structure, language, and style.

Why voice matching matters for copywriters

• Voice mismatch is the number one client objection — more common than audience misunderstanding or lack of personalisation.
• Clients who feel heard don't edit you, don't need to train you, and refer more work.
• Voice capability lets you write across niches rather than specialising.
• It justifies higher fees and enables brand voice guides as a standalone or add-on deliverable.

Step 1: Collect source material

• Get 3–5 pieces of copy written by the client themselves — not by a previous copywriter.
• Podcasts or recordings work if the client speaks freely; note that spoken content is less edited than written.
• More samples improve accuracy; 100 sentences is a reliable baseline for tool analysis.

Step 2: Measure with three tools

• Analyze My Writing — reveals average sentence length and punctuation frequency. More commas generally signal longer, more complex sentences.
• Hemingway App — shows reading grade level, sentence difficulty, and active vs passive voice use. Useful for defending your copy against "this doesn't sound like me" with data.
• IBM Tone Analyzer — identifies emotional tones (e.g. joy, fear, anger) at sentence level. Interpret tones in context: a copywriter mirroring audience fear about a deadline isn't writing fearful copy — they're reflecting the reader back.
• Measure your own voice in the same tools so you know how far to adjust.

Step 3: Analyse for style, language, and content

• Style: note use of italics, bold, underline, colour, emojis, capitalisation, and Oxford comma.
• Language: flag recurring phrases, slang, swear words, sentence rhythm. Find the signature phrase — the one line that only this client would say.
• Content: what does the client reference? Family, pop culture, personal stories, geographic markers?
• Context: who is the client arguing for or against? Where do emotional tones appear — and why?
• This, not that: define the voice at its edges — "confident, not arrogant"; "fun, not brash".

Step 4: Build a cheat sheet

• Compile all findings into a Google Doc.
• Reference it while writing, not after.
• Add a final pass when reviewing copy: "Would my client actually say this?"
• The same document can be packaged and sold as a brand voice guide.