Audience segmentation brief by Hive Backstage

---
name: audience-segmentation-brief
description: Given a specific show and artist, return a ranked, one-page segmentation brief that a CRM operator can hand straight to the send tool. Covers first-party affinity, behavioral history, recency windows, geo, and exclusions, with rough size estimates and a recommended top-line segment. Trigger whenever an event marketer or CRM operator says "who do we send this to", "build the segment for [artist]", "draft the segmentation brief", "what's the audience for the [artist] show", or describes a show and asks who should get the dedicated send. Also trigger when they say things like "I don't want to blast the whole list, who should I pull" or "what's our best slice for this one". This skill produces a hand-off doc, not the email itself. Use it before any dedicated artist-specific send. Do not use it for full-list announce emails (use `pre-show-announcement-email`).
---
# Audience segmentation brief

## What this skill does

This skill produces a one-page segmentation brief for a single show. The brief is a hand-off doc for whoever runs the CRM. It tells them which fans to pull, in what order of priority, with what exclusions, and roughly how many sends each segment should produce.

This is not the email. This is the input upstream of the email. Most venues blast the full list because building segments is slow. The brief makes the build fast. Done well, it turns a 30,000-name blast into a 4,000-name dedicated send that out-converts the blast by a wide margin.

## When to trigger

Trigger when an operator asks who should get a dedicated send for a specific show, when they want to avoid blasting the full list, or when they say something like "build me a segment for [artist]" or "who's the audience for this one". Trigger also when the operator pastes show details and asks for the right slice of the list.

Do not trigger this skill for full-list announce emails (use pre-show-announcement-email) or for loyalty-tier emails based purely on attendance count (use loyalty-tier-email). This skill is for artist- and show-specific segmentation only.

## Required inputs

The brief is only as good as the inputs. Ask for these in one message before drafting if anything is missing.

Artist name and short context. Genre, scene, era, anything that frames who the artist's fan looks like in the database. If the operator says "indie folk", that goes in. If they say "the kind of fan who buys vinyl", that goes in too.

Date and city of the show, and the venue or room.

Similar past artists at the venue. The single most useful input. Two to four prior shows at the venue whose attendees would plausibly come to this one. Names, dates, and room if multiple venues.

Recent behavior signal the operator cares about. Has the fan opened the last three emails. Has the fan clicked any tour-related link in the last 90 days. Has the fan bought a ticket in the last twelve months. Pick what the venue tracks.

Geo constraint. City, metro, radius from the venue. Some venues pull at a 50-mile radius. Some pull tighter. Confirm before drafting.

Any known exclusions. Already bought tickets to this show. Unsubscribed from artist-specific sends. Bounced in the last 30 days. Marked as VIP and going through a different channel. The operator should name these. Do not assume.

Estimated total list size and rough breakdown if known. A list of 80,000 with 30,000 active in the last 12 months segments differently than a list of 80,000 with 5,000 active. If the operator does not know, ask once. If they cannot get it, draft the brief in shape but note the size estimates are placeholders.

## Workflow

Draft segments from most relevant to least relevant. The CRM operator should be able to read top-down and stop pulling when the segment size hits the venue's target for the send.

Stack signals rather than replace them. The strongest segments combine two or more inputs: "fans who attended [Similar Artist 1] OR [Similar Artist 2] AND are within 50 miles AND have opened any email in the last 90 days". Single-signal segments are weaker and should go further down the brief.

Recency windows matter. A fan who attended a similar show in the last 18 months is a stronger signal than one who attended five years ago. Tighten recency on the top segment.

Always include exclusions as a separate section, not inside each segment. Exclusions apply globally to the send.

If two segments overlap heavily (which they often do), note the overlap and recommend a dedupe order. Usually the CRM tool dedupes by send, so the order in the brief should match the priority order.

## Output structure

Always use this exact structure. The CRM operator should be able to scan it in 90 seconds.

```markdown
# Segmentation Brief: [Artist] at [Venue], [Date]

**Show one-liner:** [Two sentences from the operator on who this artist's fan is. Genre, scene, anything useful.]
**Send goal:** [One sentence. The objective of the send.]

## Top segment (priority 1)
**Definition:** [Stacked criteria, written as a logical statement.]
**Estimated size:** [Number or range. Note the assumption.]
**Why this segment:** [One or two sentences on the signal logic.]

## Second segment (priority 2)
**Definition:** [Stacked criteria.]
**Estimated size:** [Number or range.]
**Why this segment:** [One or two sentences.]

## Third segment (priority 3)
**Definition:** [Stacked criteria.]
**Estimated size:** [Number or range.]
**Why this segment:** [One or two sentences.]

[Add a fourth segment only if the operator wants it. The brief should not exceed five segments. Past five, the venue is back to blasting.]

## Global exclusions
[Bullet list of exclusions applied to every segment.]

## Dedupe order
[A single line naming the order in which segments should dedupe against each other.]

## Total estimated reach
[One line. Combined estimated reach after dedupe and exclusions.]

## Recommended top-line
[Two to three sentences. Which single segment the skill would lead with if the operator only had time to send to one segment, and why.]
```

## Hard rules

Stack signals. Single-criterion segments like "everyone in the city" are weak. The brief defaults to combined signals.

Name the assumptions. If a size estimate is a guess, say it is a guess. If the brief assumes the venue's CRM has a "last attended show" field, state it. The CRM operator should never be surprised by a hidden assumption.

Pick one top-line. The brief ends with a single recommended top-line segment, not three options. Editorial judgment is the point.

Do not invent fields the venue's CRM does not have. Ask the operator which fields are available in their system if uncertain. Building a brief around fields the operator cannot pull is wasted work.

Recency windows default to 18 months for past-show signals and 90 days for engagement signals (opens, clicks), unless the operator says otherwise.

Exclusions are global. Do not stack exclusion logic into each segment. The CRM operator applies them once across the send.

## What to deliver

Return the completed brief in the exact structure above. Then add two to three sentences at the end naming what is still soft. Specifically: which size estimates are guesses, which fields you assumed the CRM has, and the single follow-up question that would tighten the brief the most.

Do not walk the operator through every segment choice. The brief is the doc. The closing note is short.