| name | launch-retrospective |
| description | Use after a product launch to review results - structured analysis of metrics, what worked, what didn't, and concrete next actions |
Launch Retrospective
Overview
Structured review of what happened after launch. Produces honest analysis with concrete next actions, not vanity metrics.
Announce at start: "I'm using the launch-retrospective skill to review the launch."
Process
Step 1: Context gathering
- Read all existing marketing docs to understand the original plan
- Ask the output mode question (minimal vs full)
Step 2: Four areas (one at a time)
Metrics review
- Ask what data is available (don't demand specific tools or dashboards)
- Common metrics: signups, traffic, conversion rates, email open/click rates, social engagement, revenue, press mentions
- Present what was measured vs what was planned to measure
- Flag gaps in measurement
What worked
- Which channels outperformed expectations?
- Which messages resonated most?
- What surprised you?
- Probe for specifics: "social did well" → "which posts? what engagement numbers?"
What didn't work
- Which bets didn't pay off?
- What got skipped from the plan and why?
- What was the gap between plan and execution?
- No blame - honest assessment only
Lessons and next actions
- Each lesson gets a specific, actionable next step
- Not "do better at social" but "short-form video outperformed static posts 3:1 - shift 40% of content budget to video next quarter"
- Prioritize: what has the highest impact if changed?
Step 3: Write output
- Write
docs/marketing/launch-retrospective.md
- Full mode adds: metrics summary as JSON, updated persona notes with real learnings, revised channel priority ranking, next-quarter action checklist
Step 4: Handoff
- For v2 or next product: suggest cycling back to
globalcoder-marketing:market-research or globalcoder-marketing:positioning-strategy with learnings as input
Key Principles
- Honest over optimistic - a retro that says everything went great is useless
- Specifics over generalities - push for numbers, examples, evidence
- Lessons need actions - every insight must have a "so what do we do about it?"
- Work with available data - don't gatekeep on perfect analytics, use what exists
