Venue voice profile builder by Hive Backstage

---
name: venue-voice-profile-builder
description: Read 5 to 10 pieces of a venue's strongest past content and produce a venue-specific voice profile that downstream Backstage skills (pre-show announcement, recap, logistics, segmentation) will use as their voice source of truth. Trigger this skill whenever an operator says "build a voice profile", "set up our voice", "onboard a new venue", "I have past emails, what's our voice", or pastes a folder of past announcements, recaps, social captions, or web copy and wants a usable profile back. Also trigger when another Backstage skill cannot find an existing venue voice profile and needs one before it can write in the venue's voice. This is the foundation skill for every email and content skill in the Backstage library, so prefer running it once up front rather than guessing a voice on the fly.
---
# Venue voice profile builder

## What this skill does

This skill takes a venue's real past content and produces a written voice profile that every other Backstage email and content skill will read before drafting anything. The point is to write in the venue's voice, not in a generic event marketing voice, and to do that consistently across announcements, recaps, logistics, social, and outreach.

The output is a single document. Other skills load it. Operators read it. It is the artifact that turns one good copywriter's instincts into reusable house style.

## When to trigger

Trigger when an operator is onboarding a new venue, when an operator says they want to "set up our voice" or "lock in how we sound", or when they paste samples of past content and ask for a profile. Also trigger when another Backstage skill is about to draft an email and no voice profile exists for the venue, because writing without a profile produces generic copy that does not match the room.

If the operator only has one or two samples, ask for more before drafting. Five is the floor. Ten is better. The profile is only as honest as the source material.

## Required inputs

Ask for these before drafting. Do not invent any of them.

The venue name and a one-line sense of the room, e.g. "550-cap indie club, mostly rock and indie touring acts". The operator should provide this; if they do not, ask once.

Five to ten pieces of past content. Mix the formats. Strongest blend is two or three on-sale announcement emails, two recap or post-show emails, two or three social captions, and one piece of web copy (about page, FAQ, manifesto). If the operator pastes only emails, ask once for one or two social captions and a web blurb so the profile covers more than one channel.

Optional but valuable: a list of writers behind the samples (single author voice is different from rotating staff voice), and any past style guides the venue has used.

## Workflow

Read every sample end to end first. Do not start writing the profile until you have read everything. The profile reflects the body of work, not the first email.

While reading, take notes on six dimensions:

Tone. Formality, emotional register, authority posture, humor frequency, directness. Mark these as positions on a spectrum, with evidence quoted directly from the samples.

Language. Reading level, signature phrases that repeat across samples, jargon comfort, words the venue clearly loves, words the venue clearly avoids. Look for repeated phrases across two or more samples before you call something a signature phrase. One use is a coincidence.

Structure. How announcements open. How recaps open. How emails close. Where the CTA lives. Average sentence length. Average paragraph length. Whether the venue uses headers, bullets, fragments, or runs everything as prose.

Audience. Who the venue assumes is reading. What knowledge they assume. Point of view (you, we, third person). Whether the venue speaks to fans, ticketholders, members, or some other label.

Hard rules. Things the venue consistently does and consistently avoids. Em-dashes or no. Emoji or no. Exclamation points or no. All-caps for emphasis or no. Subject line conventions. CTA conventions. Anything that, if violated, would look obviously off-brand.

Voice samples. Three to five lines pulled directly from the samples that capture the voice at its strongest. These become reference quotes for downstream skills.

## Output structure

Always use this structure. Other skills depend on the section headers being consistent.

```markdown
# Voice Profile: [Venue Name]

**Built:** [date] | **Samples analyzed:** [count]
**Sources:** [list of sample titles or short descriptions]

## Voice Summary
[One paragraph. The shortest accurate description of how this venue sounds. Should be quotable.]

## Tone Profile
[Table with five rows: Formality, Emotional Register, Authority, Humor, Directness. Each row has a Position column and an Evidence column with a direct quote from the samples.]

## Language & Vocabulary

### Reading Level
[Grade range. One sentence on sentence style.]

### Signature Phrases & Patterns
[Bullet list. Only include phrases that appear in two or more samples. Quote them exactly.]

### Jargon & Technical Depth
[One paragraph. What industry terms the venue uses naturally and which it avoids.]

### Words They Love
[Short bullet list, four to eight items. Evidence based.]

### Words They Avoid
[Short bullet list, four to eight items. Evidence based.]

## Structure & Formatting

### Typical Announcement Structure
[Numbered steps. Five or so.]

### Typical Recap Structure
[Numbered steps. Five or so.]

### Sentence & Paragraph Style
[Average sentence length, average paragraph length, notable patterns.]

### Formatting Habits
[Headers, lists, bold, emoji, CTAs. One line each.]

## Audience & Persona

### Target Reader
[Who this venue is writing to.]

### Knowledge Assumptions
[What the reader is assumed to know and not know.]

### Point of View
[You, we, third person. One sentence on the dynamic this creates.]

## Hard Rules
[Numbered list of dos and donts pulled from the samples. Examples: No em-dashes anywhere. Subject lines never use all caps. CTA is always a single button at the bottom.]

## Voice Samples
[Three to five direct quotes from the source material, with a one-line note on what each quote captures.]
```

## Hard rules for the profile itself

Pull every trait from evidence in the samples. If a trait does not appear in the samples, do not include it. The profile is not aspirational. It documents what the venue already does well.

Quote directly. When you cite evidence, paste the line verbatim. Never paraphrase a sample line into the profile and call it evidence.

If the samples contradict each other, name the contradiction in the Voice Summary and let the operator decide which version to lock in. Do not paper over it.

If fewer than five samples are provided, stop and ask for more before drafting. A two-sample profile is a guess wearing a suit.

Default to the operator's existing rules. If their past emails never use em-dashes, the profile says no em-dashes. If they always close with the same sign-off, the profile says always close with that sign-off. The skill documents the venue, it does not redesign it.

## What to deliver

Return the completed profile in the exact structure above, saved as a markdown file or pasted into the operator's chosen home for it. Then tell the operator two things in plain text, no bullets:

First, which sections you felt confident about and which you flagged as thin because the samples did not cover them well.

Second, the three pieces of content you would ask for next to strengthen the profile if they have time, e.g. "two more recap emails would let us lock down the post-show voice with more confidence."

Keep that closing note to three or four sentences. Do not walk the operator through every word choice in the profile.