# /cm:retro — Campaign / Sprint Retrospective

Structured post-mortem for campaigns, launches, or sprints.

## What It Does

A deeper analysis than `/cm:compound`. While compound captures quick learnings, retro does a full post-mortem: what happened, why it happened, what to change. Best run after campaigns, launches, or sprint cycles.

## Time Investment

20-30 minutes

## Process

### 1. Set the Scope

- What are we reviewing? (Campaign name, sprint #, launch, experiment)
- What was the timeline? (Start → end)
- What was the original goal?
- What was the actual outcome?

### 2. Results vs. Expectations

| Metric | Target | Actual | Delta | Verdict |
|--------|--------|--------|-------|---------|
| [Primary metric] | ... | ... | ... | 🟢/🟡/🔴 |
| [Secondary metric] | ... | ... | ... | 🟢/🟡/🔴 |
| [Secondary metric] | ... | ... | ... | 🟢/🟡/🔴 |

**Overall:** Did we hit the goal? Why or why not?

### 3. What Worked (Keep Doing)

List things that went well. Be specific — not "content was good" but "the comparison blog post drove 340 signups at $2.40 CAC."

For each:
- What specifically worked?
- Why did it work?
- How can we do more of this?
- Can this be systematized?

### 4. What Didn't Work (Stop Doing)

List things that failed or underperformed. No blame — just facts.

For each:
- What happened?
- Why didn't it work? (Root cause, not surface)
- How much time/money was spent?
- Should we try again differently, or kill it?

### 5. What We'd Change (Start Doing)

Based on the above:
- What would we do differently if we ran this again?
- What new approach should we try?
- What process changes would improve execution?

### 6. Surprises & Insights

Things you didn't expect — both good and bad:
- Unexpected channel that worked
- Audience segment that responded differently
- Content format that surprised
- Timing or sequencing insight

### 7. Action Items

**Concrete next steps:**

| Action | Owner | Deadline | Priority |
|--------|-------|----------|----------|
| [Action] | [Who] | [When] | High/Med/Low |
| ... | ... | ... | ... |

### 8. Compound the Learnings

**Run `/cm:compound` to document:**
- Key insights worth keeping
- Patterns to remember
- Frameworks to reuse
- Data points for future decisions

## Output

```markdown
## Retro: [Campaign/Sprint Name] — [Date]

### Summary
[1-2 sentence summary of what happened]

### Scorecard
| Metric | Target | Actual | Verdict |
|--------|--------|--------|---------|
| ... | ... | ... | ... |

### Keep (What Worked)
1. ...
2. ...

### Stop (What Didn't Work)
1. ...
2. ...

### Start (What to Change)
1. ...
2. ...

### Key Insight
[The single most important thing we learned]

### Action Items
1. ...
2. ...
```

## When to Run

- After every campaign ends
- At sprint boundaries (end of 2-week sprint)
- After product launches (1-2 weeks post-launch)
- After A/B tests conclude
- After any significant marketing initiative

## Related Workflows

- `/cm:compound` — Quick learning capture (use for small tasks)
- `/cm:sprint` — Sprint planning (retro feeds into next sprint)
- `/cm:weekly` — Weekly review (smaller scope than retro)
- `/cm:audit` — Quarterly audit (bigger scope than retro)

## Common Mistakes

- Skipping the retro (most common — and most costly)
- Making it about blame (it's about learning, not fault-finding)
- Not capturing action items (retro without actions is just venting)
- Doing it too late (memory fades — run within 1 week of campaign end)
- Not reading previous retros before planning new work