Brand Voice (Global)
The foundation skill of the content cluster. Every other content skill — 36-content-brief-global, 37-social-caption-global, 04-script-video-global, 05-ad-copy-global, 14-email-marketing-global, 38-community-seeding-global — reads this document before writing. Without it, output is technically correct and tonally inconsistent.
Read .agents/product-marketing-context-global.md and any existing brand guideline first. If information is missing, ask up to 4 questions:
- What does the brand sell, to whom, at what price point — and what should a reader feel after reading its content?
- Which 3 adjectives describe the brand's actual character — and how must the brand never be perceived?
- Is there a brand or writer whose voice is close to the target? (reference only, never to copy)
- What words do customers use when they talk about the brand or category? Paste real reviews, comments, support tickets, or call notes — this is the single most valuable vocabulary source.
Principles
- Voice is a living document. Review quarterly or whenever the brand pivots. Do not write it once and file it.
- It is a compass, not a statute. Use it to check work, and flex by context.
- The core voice is constant across channels. Only format and length change per platform — never the personality.
- Vocabulary comes from real customer language. Run
09-customer-insight-global first so word choices are sourced, not invented.
- Final test: cover the logo and it still reads as you. If it could be any competitor in the category, it is not done.
- Every claim in the voice must be substantiable. A confident voice does not license unsupported superiority or results claims — FTC advertising rules apply to organic posts, not just ads.
Workflow
1. Lock brand personality: 3 adjectives
Pick exactly 3 core adjectives, each with one sentence explaining why it reflects the brand truthfully. More than 3 dilutes the voice.
Illustrative example (a DTC skincare brand):
- Direct: names the actual problem in the first line, no throat-clearing.
- Practical: every claim comes with something the reader can do today.
- Evidence-led: cites the ingredient, the study, or the number — never "clinically proven" without the clinic.
Also record 3 things the brand must never be perceived as (for example: not luxury-aloof, not pressure-selling, not vague-wellness).
2. Define tone of voice across 5 dimensions
For each dimension write both columns. The "is not" column matters as much as the first — it is what stops voice drift.
| Dimension | The brand is... | The brand is not... |
|---|
| Energy | [warm / measured / high-energy] | [...] |
| Formality | [casual / semi-professional / formal] | [...] |
| Person and address | [we-you / I-you / third person] | [...] |
| Use of evidence | [always a number / story first, data second] | [...] |
| Handling objections | [acknowledge then explain / rebut with proof] | [...] |
3. Build the vocabulary: use / avoid / banned
-
Use often: words that carry the personality, plus real customer phrasing pulled from reviews and support (source it from 09-customer-insight-global).
-
Avoid: words that contradict the personality or overclaim in the category.
-
Banned words — hard stop, enforced at copy review. Default global marketing-cliche list, plus the brand's own additions:
"game-changer", "revolutionary", "unlock your potential", "don't miss out", "act now", "hurry — limited time only", "take it to the next level", "elevate your [noun]", "in today's fast-paced world", "we're excited to announce", "best-in-class", "world-class", "seamlessly", "one-stop shop", "must-have", "literally life-changing", "no-brainer", "synergy", "leverage" (as a verb in customer-facing copy).
-
Canonical product names: one agreed spelling and capitalization per product or plan. Not a new variation per post.
4. Set sentence-level rules
Answer all 5:
- Average sentence length: short / mixed / long — and why.
- Rhetorical questions: when they are allowed, when they are not.
- Bullets vs paragraphs: ratio and the contexts each fits.
- Standard opening patterns: pain hook / curiosity / data point / story.
- Standard closing patterns: specific CTA / question / summary.
5. Write 5 before/after examples
Each example needs: an off-voice version, an on-voice version, and one line naming the difference. All 5 situations are mandatory, written for real — never left as placeholders. Writers use this section more than any other.
Illustrative set, same DTC skincare brand (Direct / Practical / Evidence-led):
1. Product launch caption
- Before: "Introducing our revolutionary new serum — a total game-changer for your skin. Don't miss out!"
- After: "New: a 0.3% retinal serum. Same active family as the prescription route, without the four-week peeling phase."
- Difference: names the formula and the trade-off it solves instead of stacking adjectives.
2. Replying to a negative comment
- Before: "We're so sorry to hear that! Please DM us so we can make it right."
- After: "Fair complaint. Purging past six weeks usually means the strength is too high for your barrier, not that the product failed. Reply here and we'll swap you to the 0.15%."
- Difference: answers the actual problem first; the service gesture comes second, not instead.
3. Promotion announcement
- Before: "HUGE SALE! 30% OFF EVERYTHING! Act now before it's gone!"
- After: "30% off through Sunday. It's overstock from the December run — same formula, dented boxes."
- Difference: gives a real reason for the discount, so the discount does not reset the perceived price.
4. Educational post
- Before: "Unlock the secret to glowing skin with these amazing tips!"
- After: "Vitamin C oxidizes. If your serum has turned orange, it has stopped working — which is the whole reason ours ships in an opaque pump."
- Difference: teaches one checkable fact instead of promising a secret.
5. Testimonial / case study
- Before: "Our customers absolutely LOVE us. Best product ever!"
- After: "Priya, 11 weeks in: 'The texture on my cheeks is gone. My jaw still breaks out around my period.' That second part is expected — hormonal acne needs a different lever."
- Difference: publishing the unresolved part is what makes the resolved part believable.
Core voice is unchanged. Only format, length, and register shift.
| Channel | Adjust vs. core | Keep constant |
|---|
| Facebook | Longer form allowed, conversational, light emoji if on-brand | Tone, vocabulary |
| Instagram | Visual leads, first line carries the hook, tighter caption | Personality, banned words |
| LinkedIn | One notch more professional, first-person point of view, no hard sell | Directness, evidence habit |
| TikTok | 1-2 lines, spoken-word rhythm, keyword-forward for search | Core message |
| X | Compressed to a single idea, no wind-up | Point of view, precision |
| Email | Greeting and sign-off, slightly warmer, one clear ask | Directness, actionability |
Output structure
File name: brand-voice-[brand]-[YYYYMMDD].md
# Brand Voice Document — [Brand]
## I. Brand personality
- [Adjective 1]: [one-sentence rationale]
- [Adjective 2]: [one-sentence rationale]
- [Adjective 3]: [one-sentence rationale]
- Never perceived as: [3 things]
## II. Tone of voice
| Dimension | The brand is... | The brand is not... |
|-----------|-----------------|---------------------|
| Energy | | |
| Formality | | |
| Person and address | | |
| Use of evidence | | |
| Handling objections | | |
## III. Vocabulary
**Use often:** [words / phrases, with source where customer-derived]
**Avoid:** [words / phrases]
**Banned words:** [hard-stop list]
**Canonical product names:** [agreed spelling per product/plan]
## IV. Sentence rules
- Average sentence length: [short / mixed / long] — [why]
- Rhetorical questions: [when yes / when no]
- Bullets vs paragraphs: [ratio + context]
- Opening pattern: [pain / curiosity / data / story]
- Closing pattern: [CTA / question / summary]
## V. Before / after — 5 examples
### 1. Product launch caption
- Before: "[off-voice]"
- After: "[on-voice]"
- Difference: [one line]
### 2. Replying to a negative comment
[same format]
### 3. Promotion announcement
[same format]
### 4. Educational post
[same format]
### 5. Testimonial / case study
[same format]
## VI. Platform adaptation
| Channel | Adjust vs. core | Keep constant |
|---------|-----------------|---------------|
| | | |
## VII. Copy approval checklist
- [ ] Read it aloud — does it sound like a person or like a report?
- [ ] Any banned words present?
- [ ] Does the first line earn the second?
- [ ] Any word that can be cut without losing meaning? Cut it.
- [ ] Is the next action obvious to the reader?
- [ ] Is every claim substantiable if challenged?
- [ ] Cover the logo — does it still read as us?
How other skills consume this document
- Every written asset (caption, script, ad copy, email, community post) checks sections II and III before drafting, and section VII before handoff.
- Copy reviewers approve against section VII, not against personal taste.
- When onboarding a freelance writer, agency, or creator, send sections I-V as the briefing pack.
09-customer-insight-global: run first — real customer language from reviews and calls is the raw material for vocabulary.
36-content-brief-global: every brief cites the tone and banned words from here.
37-social-caption-global, 04-script-video-global, 05-ad-copy-global, 14-email-marketing-global: read this document before producing output.
product-marketing-context-global: source of positioning, audience, and price point that the voice must fit.
Quality checklist
- All 7 sections present, from personality through copy approval checklist.
- Exactly 3 adjectives, each with a specific rationale — no generic descriptors.
- Tone table covers all 5 dimensions and includes the "is not" column.
- Vocabulary has all three groups plus canonical product names.
- All 5 before/after examples are written for real in the brand's voice — no placeholders.
- Platform adaptation covers every channel the brand actually runs.
- At least some vocabulary is sourced from real customer language, with the source noted.
- No unsupported superiority or results claims anywhere in the examples.
- File name follows
brand-voice-[brand]-[YYYYMMDD].md.