Service Cost Analysis

Overview

Service cost analysis is a systematic process for estimating infrastructure and operational costs for software projects. Research current pricing, model costs across usage scenarios, calculate ROI projections, and identify optimization opportunities. Use this skill to make informed budget decisions backed by real pricing data.

When to Use This Skill

This skill should be used when:

• Estimating infrastructure costs for new projects
• Comparing cloud provider pricing (AWS, GCP, Azure)
• Analyzing third-party service costs (SaaS, APIs)
• Creating cost projections across usage scenarios
• Calculating ROI and payback periods
• Identifying cost optimization opportunities

Visual Enhancement with Project Diagrams

When documenting cost analysis, include visualizations.

Use the project-diagrams skill to generate:

• Cost breakdown charts
• Scenario comparison charts
• ROI projection graphs
• Cost optimization decision trees

python .claude/skills/project-diagrams/scripts/generate_schematic.py "diagram description" -o diagrams/output.png

---

Cost Analysis Framework

Cost Categories

Category	Examples	Variability
Compute	VMs, containers, serverless	Usage-based
Storage	Object storage, block storage, databases	Capacity + operations
Networking	Data transfer, load balancers, CDN	Traffic-based
Databases	Managed databases, caching	Capacity + throughput
Third-Party	APIs, SaaS tools, services	Usually usage-based
Development	CI/CD, monitoring, logging	Often fixed + usage
Personnel	Development, operations, support	Fixed (but critical)

Research Protocol

For every service, follow this protocol:

1. Find Official Pricing

   • Use WebFetch or WebSearch to get current pricing
   • Look for pricing calculators
   • Note pricing model (per-hour, per-request, per-GB, etc.)

2. Identify Cost Drivers

   • What metrics drive cost?
   • Are there minimum commitments?
   • What are the tiered pricing breakpoints?

3. Document Assumptions

   • Usage estimates (requests, storage, users)
   • Growth projections
   • Geographic distribution

4. Calculate Scenarios

   • Low (conservative usage)
   • Mid (expected usage)
   • High (aggressive growth)

Service Cost Specification Schema

service_cost:
  # Identity
  service_name: "string"
  provider: "aws | gcp | azure | other"
  category: "compute | storage | database | networking | third_party | development"

  # Pricing
  pricing_model: "per_hour | per_request | per_gb | per_user | flat_rate | tiered"
  pricing_source: "URL to official pricing"
  pricing_date: "YYYY-MM-DD"

  # Cost Estimates
  monthly_cost:
    low: number
    mid: number
    high: number

  # Assumptions
  assumptions:
    - "string - assumption 1"
    - "string - assumption 2"

  # Usage Estimates
  usage:
    metric: "string (e.g., requests, GB, hours)"
    low: number
    mid: number
    high: number

  # Notes
  notes: "string - important considerations"
  optimization_opportunities:
    - "string - potential optimization"

Cloud Provider Cost Analysis

AWS Cost Analysis

Common Services:

Service	Pricing Model	Key Cost Drivers
EC2	Per-hour + data transfer	Instance type, hours, data out
RDS	Per-hour + storage + IOPS	Instance type, storage, multi-AZ
S3	Per-GB + requests	Storage class, requests, data transfer
Lambda	Per-request + duration	Invocations, memory, duration
ECS/Fargate	Per-vCPU-hour + memory-hour	Task size, hours running
CloudFront	Per-request + data transfer	Requests, data out, regions

AWS Pricing Research:

# Use WebSearch to find current pricing
WebSearch: "AWS [service] pricing 2025"
WebFetch: "https://aws.amazon.com/[service]/pricing/"

Example AWS Cost Specification:

service_costs:
  - service_name: "AWS RDS PostgreSQL"
    provider: "aws"
    category: "database"
    pricing_model: "per_hour"
    pricing_source: "https://aws.amazon.com/rds/postgresql/pricing/"
    pricing_date: "2025-01-06"

    monthly_cost:
      low: 50      # db.t3.micro, 20GB, single-AZ
      mid: 250     # db.t3.medium, 100GB, multi-AZ
      high: 800    # db.r5.large, 500GB, multi-AZ, provisioned IOPS

    assumptions:
      - "US East (N. Virginia) region"
      - "Reserved instances not applied (on-demand pricing)"
      - "Low: Development environment"
      - "Mid: Production with moderate traffic"
      - "High: Production with high availability requirements"

    usage:
      metric: "instance hours + storage GB"
      low: "730 hours db.t3.micro + 20GB"
      mid: "730 hours db.t3.medium + 100GB + multi-AZ"
      high: "730 hours db.r5.large + 500GB + multi-AZ + 3000 PIOPS"

    notes: |
      Consider Aurora for better scalability above 500GB.
      Reserved instances can reduce costs by 30-60%.
      Multi-AZ doubles compute cost but is recommended for production.

    optimization_opportunities:
      - "Reserved instances for predictable workloads"
      - "Aurora Serverless for variable workloads"
      - "Read replicas instead of larger instance"

GCP Cost Analysis

Common Services:

Service	Pricing Model	Key Cost Drivers
Compute Engine	Per-second (min 1 min)	Machine type, sustained use
Cloud SQL	Per-hour + storage	Instance type, HA configuration
Cloud Storage	Per-GB + operations	Storage class, operations
Cloud Functions	Per-invocation + compute time	Invocations, memory, duration
Cloud Run	Per-request + vCPU-second	Requests, CPU, memory
BigQuery	Per-TB scanned + storage	Query bytes, storage

GCP Pricing Research:

WebSearch: "GCP [service] pricing 2025"
WebFetch: "https://cloud.google.com/[service]/pricing"

Azure Cost Analysis

Common Services:

Service	Pricing Model	Key Cost Drivers
Virtual Machines	Per-hour	VM size, hours, data transfer
Azure SQL	DTU or vCore model	Compute tier, storage
Blob Storage	Per-GB + operations	Access tier, redundancy
Azure Functions	Per-execution + duration	Executions, memory
Container Apps	Per-vCPU-second	vCPU, memory, requests
Cosmos DB	RU/s + storage	Provisioned RUs, storage

Third-Party Service Costs

Common Categories:

Category	Examples	Typical Pricing
Authentication	Auth0, Clerk, Firebase Auth	Per-MAU
Payments	Stripe, PayPal	% of transaction
Email	SendGrid, Postmark, SES	Per-email
Search	Algolia, Elasticsearch Cloud	Per-search + records
Monitoring	Datadog, New Relic	Per-host + features
Error Tracking	Sentry, Bugsnag	Per-event
Analytics	Mixpanel, Amplitude	Per-MTU

Example Third-Party Cost:

service_costs:
  - service_name: "Stripe Payment Processing"
    provider: "stripe"
    category: "third_party"
    pricing_model: "per_transaction"
    pricing_source: "https://stripe.com/pricing"
    pricing_date: "2025-01-06"

    monthly_cost:
      low: 150      # $5,000 GMV
      mid: 750      # $25,000 GMV
      high: 3000    # $100,000 GMV

    assumptions:
      - "Standard pricing: 2.9% + $0.30 per transaction"
      - "Average transaction: $50"
      - "Low: 100 transactions/month"
      - "Mid: 500 transactions/month"
      - "High: 2000 transactions/month"

    notes: |
      Volume discounts available above $100K/month.
      Additional fees for international cards (+1.5%).
      Subscription billing may have different rates.

Cost Estimation Templates

Monthly Cost Summary

monthly_cost_summary:
  project: "[Project Name]"
  date: "YYYY-MM-DD"
  currency: "USD"

  scenarios:
    low:
      description: "MVP / Development"
      users: "< 100"
      total: 0
    mid:
      description: "Initial Production"
      users: "1,000 - 10,000"
      total: 0
    high:
      description: "Scale / Growth"
      users: "> 10,000"
      total: 0

  by_category:
    compute:
      services: []
      low: 0
      mid: 0
      high: 0
    storage:
      services: []
      low: 0
      mid: 0
      high: 0
    database:
      services: []
      low: 0
      mid: 0
      high: 0
    networking:
      services: []
      low: 0
      mid: 0
      high: 0
    third_party:
      services: []
      low: 0
      mid: 0
      high: 0
    development:
      services: []
      low: 0
      mid: 0
      high: 0

  totals:
    monthly:
      low: 0
      mid: 0
      high: 0
    annual:
      low: 0
      mid: 0
      high: 0

ROI Analysis Template

roi_analysis:
  project: "[Project Name]"

  investment:
    development_cost: 0
    infrastructure_setup: 0
    training_and_onboarding: 0
    total_initial_investment: 0

  ongoing_costs:
    monthly_infrastructure: 0
    monthly_third_party: 0
    monthly_personnel: 0
    total_monthly_operating: 0
    annual_operating: 0

  revenue_projections:
    year_1:
      monthly_revenue: 0
      growth_rate: "% per month"
      annual_revenue: 0
    year_2:
      annual_revenue: 0
    year_3:
      annual_revenue: 0

  metrics:
    payback_period_months: 0
    year_1_roi: "percentage"
    year_3_roi: "percentage"
    break_even_users: 0

  assumptions:
    - "Revenue assumption 1"
    - "Cost assumption 2"
    - "Growth assumption 3"

Cost Optimization Strategies

Compute Optimization

Strategy	Savings	When to Use
Reserved Instances	30-60%	Predictable, steady workloads
Spot/Preemptible	60-90%	Fault-tolerant, batch jobs
Right-sizing	20-40%	Over-provisioned resources
Auto-scaling	Variable	Variable traffic patterns
Serverless	Variable	Sporadic, unpredictable loads

Storage Optimization

Strategy	Savings	When to Use
Tiered Storage	40-80%	Infrequently accessed data
Lifecycle Policies	Variable	Data with known access patterns
Compression	20-50%	Compressible data types
Deduplication	Variable	Redundant data

Database Optimization

Strategy	Savings	When to Use
Reserved Capacity	30-60%	Predictable workloads
Read Replicas	Variable	Read-heavy workloads
Serverless	Variable	Variable traffic
Query Optimization	20-50%	Pay-per-query models

Third-Party Optimization

Strategy	Savings	When to Use
Annual Contracts	10-30%	Committed usage
Volume Discounts	Variable	High volume
Alternative Providers	Variable	Comparable services
Self-hosting	Variable	High volume, DevOps capacity

Cost Analysis Report Structure

# Cost Analysis Report: [Project Name]

## Executive Summary
- Total estimated monthly cost: $X - $Y
- Primary cost drivers: [List top 3]
- Key optimization opportunities: [List top 3]

## Cost Breakdown by Category

### Compute ($X/month)
[Service details and cost justification]

### Storage ($X/month)
[Service details and cost justification]

### Database ($X/month)
[Service details and cost justification]

### Third-Party Services ($X/month)
[Service details and cost justification]

## Scenario Analysis

### Low Scenario (MVP)
- Users: X
- Monthly cost: $Y
- Suitable for: [Use case]

### Mid Scenario (Production)
- Users: X
- Monthly cost: $Y
- Suitable for: [Use case]

### High Scenario (Scale)
- Users: X
- Monthly cost: $Y
- Suitable for: [Use case]

## ROI Analysis
[If applicable]

## Optimization Recommendations
1. [Recommendation with potential savings]
2. [Recommendation with potential savings]
3. [Recommendation with potential savings]

## Assumptions and Risks
- [Key assumption 1]
- [Key assumption 2]
- [Cost risk 1]

## References
- [Pricing source 1]
- [Pricing source 2]

Quality Checklist

Before completing cost analysis:

• All major services identified
• Current pricing verified from official sources
• Pricing dates documented
• Usage assumptions documented
• Three scenarios calculated (low/mid/high)
• Cost optimization opportunities identified
• ROI calculated (if applicable)
• Total monthly and annual costs calculated
• Primary cost drivers identified

Best Practices

Do's

• Always cite official pricing sources
• Include pricing date (pricing changes frequently)
• Document all assumptions explicitly
• Consider all cost categories (don't forget networking)
• Calculate both monthly and annual costs
• Identify optimization opportunities proactively

Don'ts

• Don't use outdated pricing data
• Don't forget data transfer costs (often overlooked)
• Don't ignore startup vs. steady-state costs
• Don't assume linear scaling of costs
• Don't forget development and operational tooling
• Don't skip personnel costs in TCO analysis