Market Brand — Voice, Tone, and Archetype

Analyze how a brand communicates across every touchpoint and produce a usable voice guide that a new copywriter could pick up and write on-brand within an hour.

Source material (in priority order)

1. Homepage — the most curated representation
2. About page — how they describe themselves
3. Product / pricing pages — how they sell
4. Blog (3–5 recent posts) — extended voice
5. Social profiles — conversational voice
6. Error / empty states / microcopy — unguarded voice
7. Email or newsletter if accessible
8. Job postings — internal culture signal

Use scripts/analyze_page.py to pull raw text from each page.

The 4 Voice Dimensions (score each 1–10)

Formal ↔ Casual

Signals: contractions, sentence length, "we/you" vs "the company/one", greetings, slang.

Serious ↔ Playful

Signals: exclamation marks, emoji, puns, metaphors, error message style, self-deprecation.

Technical ↔ Simple

Signals: jargon density, acronym explanation, detail depth, assumed expertise.

Reserved ↔ Bold

Signals: hedged claims ("may help") vs direct claims ("guaranteed"), opinionated stances, competitor mentions.

Every score needs 3 quoted examples from the source material — no score without evidence.

Voice map (ASCII visualization)

Formal    |----------[●]-----------| Casual     (6/10)
Serious   |-------[●]--------------| Playful    (4/10)
Technical |-----------------[●]----| Simple     (7/10)
Reserved  |--------------[●]-------| Bold       (6/10)

Archetype Detection (Jung's 12, collapsed to 12 for brand use)

Map the brand to 1 primary and optionally 1 secondary archetype:

Give the archetype score with evidence. Many brands blend two — e.g., Notion is Creator + Sage; Liquid Death is Outlaw + Jester.

Tone by Context

Voice is consistent; tone shifts. Map how the brand should sound in each context:

Vocabulary

Signature words (15–20)

Words the brand uses repeatedly, grouped by action verbs, descriptive adjectives, value words, industry terms.

Words to avoid

Off-brand words, clichés, competitor terminology.

Signature phrases

Taglines, recurring openers, linguistic patterns (e.g., "always starts with a verb").

Do / Don't Chart

Consistency Audit

For each channel, rate Consistent / Mostly / Inconsistent with a one-line note. Identify the top 1–2 divergences (e.g., "homepage is bold and confident, but support emails are formal and hedged — this erodes trust").

8 Copy Samples

Generate these so writers have references:

1. Homepage H1
2. Product one-liner
3. Feature description (100 words)
4. Blog post intro (50 words)
5. Social post (platform-appropriate)
6. Email subject line (3 variants)
7. CTA button (primary + secondary)
8. Error message / empty state

Output: BRAND-VOICE.md

# Brand Voice — [Brand]
**Analyzed:** [YYYY-MM-DD]  |  **Sources:** [list]

## Voice Summary
[2–3 sentences: archetype, dimensions, what makes this voice distinct.]

## Voice Dimensions
### Formal ↔ Casual: X/10
Evidence:
> "quoted example 1"
> "quoted example 2"

... repeat for all 4 dimensions ...

### Voice Map
[ASCII visualization]

## Archetype
**Primary:** [name] — [why, with evidence]
**Secondary:** [name] — [why]
**Fit:** [Strong / Moderate / Weak]

## Tone by Context
[Table]

## Vocabulary
### Signature words
### Avoid
### Signature phrases

## Do / Don't Chart
[Table]

## Copy Samples
[8 examples]

## Consistency Audit
[Channel-by-channel + overall score]

## Recommendations
### This week
### This quarter

Quality Bar

• Voice is diagnosed, not invented. Don't tell the brand who they should be — tell them who they already are and where they're inconsistent.
• Every dimension score and archetype choice needs quoted evidence.
• The 8 copy samples must sound like the brand, not like generic marketing. Write them after the analysis, not before.