| name | campaign-plan |
|---|---|
| description | Multi-channel campaign orchestrator. Produces a campaign brief with goal, audience, channel mix, content calendar entries, success metrics, and explicit kill criteria. Use when the user says "plan a campaign for X", "we're launching Y, plan the rollout", "what's the marketing motion for Z", or anything that needs multi-channel coordination over a window of days/weeks. |
Inspired by anthropics/marketing/campaign-planning. The orchestrator: takes a fuzzy "we want to launch X" and produces a brief tight enough that other skills (content-draft, content-repurpose, competitor-scan, sales-enablement) can execute against it.
- "plan a campaign for {launch / topic / audience}"
- "how should we roll out {feature/release/event}"
- "marketing motion for {thing}"
- New module, integration, or feature shipping in the next 4–8 weeks
- What's being launched/promoted — feature, integration, content piece, event, partnership
- Goal — awareness / consideration / activation / retention / advocacy. One primary goal. If two, pick the bigger one.
- Target audience — executive / technical / end-user; ideally with persona detail
- Window — start date, end date, key milestone dates
- Constraints — budget (if any), partner co-marketing requirements, regulatory or legal review needs, blackout dates
- Existing assets — prior launches, related content, customer stories that apply
- Confirm goal. One primary goal, in one sentence, with a measurable success criterion. If the user hasn't given one, ask.
- Frame the audience. Who specifically — not "everyone." If multiple personas, plan rollout by persona (e.g. data leader first, IT second).
- Pick the channel mix. Per Orama's playbook:
- Anchor channel — typically blog or product page. Where the canonical claim lives.
- Distribution channels — LinkedIn, X/Bluesky, email, partner co-marketing
- Sales enablement leg — internal collateral so the sales team can ride the campaign (drives
sales-enablementskill output) - Always include 1 measurement leg — what we'll watch
- Produce a calendar. Rolling 4–6 week window with specific dates per artifact. Each entry: date, channel, asset, status, owner, depends-on. Lands as new rows in
briefs/calendar.md(don't overwrite — append). - Commission the assets. For each calendar entry, name which skill produces it. e.g. "Day 0 blog →
content-draft"; "Day +2 LinkedIn →content-repurposefrom blog"; "Day +7 battlecard →competitor-scanpaired withsales-enablement." - Define success metrics. Three max. Each tied to the primary goal. Use real numbers (e.g. "300 qualified signups in 14 days," not "good engagement").
- Define kill criteria. Explicit. "If after 7 days we see < 50% of the Day-7 metric target, we cut spend on channel X and reallocate." Without kill criteria, campaigns drift.
- Risks and reversibility. What could go wrong, how easy to back out. If reversibility is low (paid media, partner commits), flag the gates explicitly.
- Run brand-voice lint on the brief itself (it'll be read by humans and reused).
- Write the brief to
briefs/{YYYY-Q?}-{slug}.md. Updatebriefs/calendar.md. - Append to tasks. One line per major asset commissioned, plus the campaign-level entry.
- Post to
#marketer-approvewith the 5-field gate (What / Where / Why / Risks / Reversibility) before locking.
---
campaign: {slug}
status: draft | approved | active | closed
window: {YYYY-MM-DD} → {YYYY-MM-DD}
goal: {one sentence}
audience: {persona, named}
owner: {name}
created: {ISO}
created_by: campaign-plan
---
# {Campaign title}
## Goal
{One sentence. Outcome, measurable.}
## Audience
{Specific persona. Why now for them. What they care about.}
## Anchor claim
{One sentence — the central thing this campaign argues.}
## Channels
- **Anchor**: {blog | landing} — {asset name}
- **Distribution**: {list}
- **Sales enablement**: {assets}
- **Measurement**: {dashboards, tags, UTMs}
## Calendar
| Date | Channel | Asset | Skill | Status | Notes |
|---|---|---|---|---|---|
| ... | ... | ... | ... | ... | ... |
## Success metrics
1. {Metric} — target {N} by {date}
2. {Metric} — target {N} by {date}
3. {Metric} — target {N} by {date}
## Kill criteria
- If by Day {N} we see less than {N}% of {metric} target, {action}.
- If {risk event}, {action}.
## Risks and reversibility
- {Risk} — reversibility: high/medium/low. Mitigation: {action}.
## Dependencies
- {Asset/decision/approval that must land first}
## Open questions
- {Things needing answers before lock}- Don't plan more than one primary goal per campaign. If a request implies two, ask the user to pick.
- Don't generate vanity metrics (impressions alone, follower growth alone). Metrics tie to goal.
- Don't lock a brief without kill criteria and
#marketer-approvesign-off. - Don't bundle unrelated launches into one campaign just because they ship close in time.