This skill should be used when the user asks to "estimate impact", "design confidence framework", "create impact scoring", "define confidence levels", "build ROI analysis", or needs frameworks for estimating improvement impact in playbook analysis.
Design confidence frameworks with clear criteria for estimating improvement impact.
Impact estimation provides:
| Level | Confidence | Variance | Criteria |
|---|---|---|---|
| High | 90%+ | ±10 points | Direct evidence, clear causal mechanism |
| Medium | 70-89% | ±20 points | Logical connection, some uncertainty |
| Low | <70% | ±30 points | Speculative benefit |
Medium Confidence (70-89%):
Low Confidence (<70%):
## Impact Analysis
### Item: {ID}
**Current state:** X%
**With fix:** Y%
**Estimated lift:** +Z points
**Confidence:** {High/Medium/Low}
**Reasoning:**
{Why this estimate is credible}
**Evidence:**
- {Evidence point 1}
- {Evidence point 2}
## Aggregate Impact Summary
| Improvement | Items | Avg Lift | Total Points | Confidence |
|-------------|-------|----------|--------------|------------|
| {Fix 1} | N | +X | +N×X | High |
| {Fix 2} | N | +Y | +N×Y | Medium |
| {Fix 3} | N | +Z | +N×Z | Low |
**Total estimated improvement:** +{sum} points
**Weighted by confidence:** +{adjusted} points
## Impact Analysis: {Opportunity}
| Item | Current | With Fix | Lift | Confidence |
|------|---------|----------|------|------------|
| {ID1} | 45% | 75% | +30 | High |
| {ID2} | 60% | 80% | +20 | Medium |
| {ID3} | 55% | 70% | +15 | Low |
**Summary:**
- Total items: 3
- Average lift: +21.7 points
- Confidence-weighted lift: +18.5 points
When you have clear causal links:
### Estimation Method: Direct Evidence
**Applicable when:**
- Fix addresses root cause directly
- Similar fixes have known outcomes
- Trace shows clear mechanism
**Process:**
1. Identify items with this root cause
2. Calculate current score
3. Estimate score if root cause removed
4. Assign High confidence
When you can compare similar cases:
### Estimation Method: Comparative
**Applicable when:**
- Similar items have different outcomes
- Difference attributable to specific factor
- Pattern is consistent
**Process:**
1. Find similar items with/without factor
2. Calculate average scores for each group
3. Difference = estimated impact
4. Assign Medium confidence
When direct evidence is limited:
### Estimation Method: Extrapolation
**Applicable when:**
- Limited direct evidence
- Logical mechanism exists
- Must estimate from principles
**Process:**
1. Identify affected items
2. Estimate mechanism impact
3. Apply conservative estimate
4. Assign Low confidence
Priority = (Items Affected × Quality Improvement) / (Effort × Risk)
| Factor | Scale | Description |
|---|---|---|
| Items Affected | 1-50 | Count of items this addresses |
| Quality Improvement | 1-30 | Expected point improvement |
| Effort | 1-5 | Implementation difficulty |
| Risk | 1-3 | Potential for negative impact |
| Tier | Score Range | Action |
|---|---|---|
| P1 | >50 | Implement immediately |
| P2 | 20-50 | Plan for next cycle |
| P3 | <20 | Consider for backlog |
## Priority Analysis
| Opportunity | Items | Lift | Effort | Risk | Score | Priority |
|-------------|-------|------|--------|------|-------|----------|
| {Opp 1} | 15 | +20 | 2 | 1 | 150 | P1 |
| {Opp 2} | 10 | +15 | 3 | 2 | 25 | P2 |
| {Opp 3} | 5 | +10 | 4 | 2 | 6 | P3 |
## ROI Analysis
### Investment
- Development: {hours/cost}
- Testing: {hours/cost}
- Deployment: {hours/cost}
- **Total:** {total}
### Return
- Items improved: N
- Average lift: +X points
- Total improvement: +N×X points
### ROI
- Improvement per unit effort: {points/hour}
- Payback: {timeframe}
High Impact
│
Quick Wins│ Major Projects
(Do first)│ (Plan carefully)
│
──────────────┼──────────────
│
Fill-ins │ Reconsider
(If time) │ (Maybe don't)
│
Low Impact
Low Effort ──> High Effort
**Assumptions:**
1. {Assumption 1}
2. {Assumption 2}
**If assumption invalid:**
- Impact could be: {range}
**Calculation:**
- Current score: 45%
- Missing 3 key observations worth ~10 points each
- Fix addresses 2 of 3 observations
- Expected improvement: +20 points (2 × 10)
- Confidence: Medium (mechanism clear, but observation value estimated)
Before finalizing estimates:
For impact estimation patterns:
${CLAUDE_PLUGIN_ROOT}/references/playbook-pattern.md - Section 4.8