Forecast Skill

Project future cloud costs with trend analysis and scenario modeling.

Purpose

Enable proactive cost planning through:

• •Historical trend analysis
• •Multi-scenario forecasting
• •Capacity planning cost models
• •Budget recommendations
• •Forecast accuracy tracking

Commands

Process

1. Gather Historical Data

Required (6+ months): Monthly costs, service breakdown, usage metrics, business metrics

Sources: wicked-mem, cloud cost APIs, past invoices

wicked-mem recall "finops/monthly-costs/*" --limit 12

2. Analyze Trends

Methods: Growth rate, linear regression, seasonal adjustment

See Forecasting Methods.

3. Model Scenarios

Best Case: Lower growth, optimizations succeed Likely Case: Expected growth, stable pricing Worst Case: Higher growth, delays

4. Incorporate Business Context

wicked-product recall "growth-projections"

Factors: User growth, launches, expansion, campaigns Translation: +20% users → +15% compute

5. Model Capacity Planning

Current: 100 servers @ $50/mo = $5,000
Growth: +20% per quarter
With optimizations: Net growth 35% (vs 108%)

See Capacity Models.

6. Generate Forecast

Monthly breakdown with base, variable, projects Confidence ranges: Best (-20%), Likely, Worst (+30%)

7. Set Budget Recommendations

Budget = Likely Forecast × 1.10 (10% buffer)

Alert thresholds: 50%, 80%, 100%

8. Track Accuracy

Error = |Forecasted - Actual| / Actual × 100%
Targets: 1-mo <5%, 3-mo <10%, 12-mo <20%

Integration

wicked-mem: Store/retrieve forecasts wicked-product: Align with business plans FinOps Analyst: Historical data Cost Optimizer: Planned optimizations

Output Format

## Cloud Cost Forecast: {Period}

### Executive Summary
**Current**: ${current}
**Forecasted**: ${forecast}
**Budget**: ${budget}
**Growth**: {%}%/month
**Confidence**: {HIGH|MEDIUM|LOW}

### Historical Trend
{6 months data}

### Scenarios
**Best**: ${best} - {assumptions}
**Likely**: ${likely} - {assumptions} ⭐
**Worst**: ${worst} - {assumptions}

### Monthly Breakdown
| Month | Base | Variable | Projects | Total | Budget |
{rows}

### Cost Drivers
1. Business growth: {%}%
2. Infrastructure changes
3. Optimizations: -{%}%

### Budget Recommendation
**Budget**: ${budget}
**Alerts**: 50%, 80%, 100%

### Risks
| Risk | Probability | Impact | Mitigation |
{rows}

### Assumptions
{list}

### Validation
**Next Review**: {date}

Forecasting Techniques

Linear Regression: Clear trend Exponential Smoothing: Recent data emphasis Seasonal Decomposition: Repeating patterns Scenario Modeling: Uncertainty

See Methods for details.

Events

Published:

• •[finops:forecast:started:success]
• •[finops:forecast:completed:success]
• •[finops:forecast:validated:success]

Configuration

forecasting:
  default_period: quarterly
  scenarios:
    best_case_factor: 0.80
    worst_case_factor: 1.30
  buffer_percent: 10

Tips

1. •Use Real Data: Trends beat assumptions
2. •Provide Ranges: Not single points
3. •Document Assumptions: Be explicit
4. •Track Accuracy: Learn from misses
5. •Update Frequently: Quarterly minimum
6. •Separate Components: Base vs variable
7. •Business Context: Align with growth

Common Pitfalls

• •Straight-line extrapolation (ignores changes)
• •Overconfidence (forecasting is uncertain)
• •No seasonality (Q4 ≠ Q2)
• •Forgetting one-time costs
• •No validation

Reference Materials

• •Forecasting Methods
• •Capacity Planning Models
• •Scenario Analysis Guide