Cost Savings Calculator

Use this calculator to estimate your savings with the implemented optimizations.

Variables

D = Number of developers
Q = Questions per developer per day  
WD = Working days per month (typically 22)
S = Percentage of queries that are summaries (0-100)
C = Cache hit rate expected (typically 20-40%)

Formulas

Monthly Query Volume

Total Queries = D × Q × WD

Baseline Cost (All Sonnet)

Baseline = Total Queries × $0.016

Cost After Strategy 2 (Smart Models)

Summary Queries = Total Queries × (S / 100)
Deep Queries = Total Queries × ((100 - S) / 100)

Cost After S2 = (Summary Queries × $0.0065) + (Deep Queries × $0.016)
Savings S2 = Baseline - Cost After S2

Cost After Strategy 4 (Caching)

Uncached Queries = Cost After S2 × (1 - C/100)
Cached Queries = Cost After S2 × (C/100)

Cost After S4 = Uncached Queries + (Cached Queries × $0)
             = Cost After S2 × (1 - C/100)
Savings S4 = Cost After S2 × (C/100)

Total Savings

Total Savings = Savings S2 + Savings S4
Final Cost = Cost After S2 - Savings S4
Savings % = (Total Savings / Baseline) × 100

Examples

Example 1: Small Team (5 developers)

Variables:
  D = 5
  Q = 5 questions/day
  WD = 22 working days
  S = 25% summaries
  C = 30% cache hit

Calculations:
  Total Queries = 5 × 5 × 22 = 550

Baseline:
  Cost = 550 × $0.016 = $8.80

After Strategy 2 (Smart Models):
  Summaries = 550 × 0.25 = 137.5
  Deep = 550 × 0.75 = 412.5
  Cost S2 = (137.5 × $0.0065) + (412.5 × $0.016)
         = $0.89 + $6.60 = $7.49
  Savings S2 = $8.80 - $7.49 = $1.31 (15%)

After Strategy 4 (Caching at 30%):
  Cost S4 = $7.49 × (1 - 0.30) = $5.24
  Savings S4 = $7.49 × 0.30 = $2.25

Total Savings:
  Total = $1.31 + $2.25 = $3.56
  Final Cost = $8.80 - $3.56 = $5.24
  Savings % = ($3.56 / $8.80) × 100 = 40.5%

✅ Small Team Saves: ~$3.50/month (40% reduction)

Example 2: Medium Team (20 developers)

Variables:
  D = 20
  Q = 10 questions/day
  WD = 22 working days
  S = 30% summaries
  C = 35% cache hit

Calculations:
  Total Queries = 20 × 10 × 22 = 4,400

Baseline:
  Cost = 4,400 × $0.016 = $70.40

After Strategy 2 (Smart Models):
  Summaries = 4,400 × 0.30 = 1,320
  Deep = 4,400 × 0.70 = 3,080
  Cost S2 = (1,320 × $0.0065) + (3,080 × $0.016)
         = $8.58 + $49.28 = $57.86
  Savings S2 = $70.40 - $57.86 = $12.54 (18%)

After Strategy 4 (Caching at 35%):
  Cost S4 = $57.86 × (1 - 0.35) = $37.61
  Savings S4 = $57.86 × 0.35 = $20.25

Total Savings:
  Total = $12.54 + $20.25 = $32.79
  Final Cost = $70.40 - $32.79 = $37.61
  Savings % = ($32.79 / $70.40) × 100 = 46.6%

✅ Medium Team Saves: ~$33/month (47% reduction)

Example 3: Large Scale (100 developers)

Variables:
  D = 100
  Q = 8 questions/day
  WD = 22 working days
  S = 35% summaries
  C = 40% cache hit

Calculations:
  Total Queries = 100 × 8 × 22 = 17,600

Baseline:
  Cost = 17,600 × $0.016 = $281.60

After Strategy 2 (Smart Models):
  Summaries = 17,600 × 0.35 = 6,160
  Deep = 17,600 × 0.65 = 11,440
  Cost S2 = (6,160 × $0.0065) + (11,440 × $0.016)
         = $40.04 + $183.04 = $223.08
  Savings S2 = $281.60 - $223.08 = $58.52 (21%)

After Strategy 4 (Caching at 40%):
  Cost S4 = $223.08 × (1 - 0.40) = $133.85
  Savings S4 = $223.08 × 0.40 = $89.23

Total Savings:
  Total = $58.52 + $89.23 = $147.75
  Final Cost = $281.60 - $147.75 = $133.85
  Savings % = ($147.75 / $281.60) × 100 = 52.5%

✅ Large Scale Saves: ~$148/month (52.5% reduction)

Your Scenario

Fill in your values:

D = _____ developers
Q = _____ questions/dev/day
WD = _____ working days/month (default: 22)
S = ____% summaries (estimate: 20-40%)
C = ____% cache hit rate (expect: 25-40%)

Total Queries = D × Q × WD = _____________

Baseline Cost = Total Queries × $0.016 = $_____________

Cost After S2 = (S% × $0.0065) + ((100-S)% × $0.016) = $_____________
             = $_____________

Cost After S4 = Cost After S2 × (1 - C%) = $_____________

Total Savings = Baseline - Cost After S4 = $_____________
Savings % = (Total Savings / Baseline) × 100 = _______% 

Final Monthly Cost = $_____________

Breakeven Analysis

At what scale do optimizations pay for themselves?

Infrastructure Cost vs Savings

Monthly Infrastructure:
  VPS UAT: $2.50
  Domain: $1.00
  Total: $3.50

If saving $3.50+/month → Optimizations pay for themselves

Breakeven Point:

Small team (5 devs): Saves $3.50/month immediately ✅
Any team size pays for infrastructure

Larger Infrastructure (Fly.io Production)

Monthly Infrastructure:
  Fly.io APIs + Workers: $40
  Supabase: $25
  Total: $65

If saving $65+/month → Optimizations pay for themselves

Breakeven Point:

Need ~15 developers asking ~5 questions/day
Typical medium team meets this threshold

Monthly Cost Scenarios

Scenario 1: Startup (2 devs, light usage)

Baseline: 220 queries × $0.016 = $3.52/month
With Optimization: ~$1.80/month
Savings: $1.72/month (49%)

Scenario 2: Pilot Team (10 devs, 5 Q/day)

Baseline: 2,200 queries × $0.016 = $35.20/month
With Optimization: ~$16.50/month
Savings: $18.70/month (53%)

Scenario 3: Department (50 devs, 7 Q/day)

Baseline: 7,700 queries × $0.016 = $123.20/month
With Optimization: ~$57.50/month
Savings: $65.70/month (53%)

Scenario 4: Enterprise (500 devs, 10 Q/day)

Baseline: 110,000 queries × $0.016 = $1,760/month
With Optimization: ~$800/month
Savings: $960/month (55%)

ROI Timeline

Example: 50-person team with $250/month baseline

Month 1 (No optimization):
  Cost: $250
  Savings: $0
  Cumulative: $0

Month 2-3 (Optimization live, low cache hit):
  Cost: $200/month (20% savings)
  Savings: $50/month
  Cumulative: $100

Month 4-6 (Cache warming up):
  Cost: $130/month (48% savings)
  Savings: $120/month
  Cumulative: $460

Month 7+ (Mature cache, full optimization):
  Cost: $110/month (56% savings)
  Savings: $140/month
  Cumulative: $900+ (breakeven by month 7)

Important Notes

Cache Hit Rate builds gradually:
- Week 1: 5-10%
- Month 1: 20-30%
- Month 3+: 40-50%
Summary Detection depends on team patterns:
- Conservative estimate: 20% summaries
- Typical: 30-35%
- High-summary teams: 40-50%
Seasonal Variation:
- Higher queries during sprints/releases
- Lower during planning/meetings
- Adjust monthly estimates accordingly
Cost per Model (Sonnet 3.5 is default):
- If you switch to Opus: costs 5x higher
- If you switch to Haiku: costs 2.5x lower
- Mixed model strategy gives best savings

Implementation Impact

Once implemented, you'll see:

✅ Week 1: Baseline metrics established, cost tracking enabled ✅ Week 2-3: Smart model routing active, summaries routed to Haiku ✅ Week 4: First cache hits appearing (5-10%) ✅ Month 2: Cache warming up (20-30% hit rate) ✅ Month 3+: Full optimization (40-50% hit rate, 50%+ cost reduction)

Questions?

Check the implementation guide: COST_OPTIMIZATION.md Check monitoring queries: COST_OPTIMIZATION.md → "Monitoring and Alerting"

Ready to deploy and start saving!

Cost Savings Calculator ​

Variables ​

Formulas ​

Monthly Query Volume ​

Baseline Cost (All Sonnet) ​

Cost After Strategy 2 (Smart Models) ​

Cost After Strategy 4 (Caching) ​

Total Savings ​

Examples ​

Example 1: Small Team (5 developers) ​

Example 2: Medium Team (20 developers) ​

Example 3: Large Scale (100 developers) ​

Your Scenario ​

Breakeven Analysis ​

Infrastructure Cost vs Savings ​

Larger Infrastructure (Fly.io Production) ​

Monthly Cost Scenarios ​

Scenario 1: Startup (2 devs, light usage) ​

Scenario 2: Pilot Team (10 devs, 5 Q/day) ​

Scenario 3: Department (50 devs, 7 Q/day) ​

Scenario 4: Enterprise (500 devs, 10 Q/day) ​

ROI Timeline ​

Example: 50-person team with $250/month baseline ​

Important Notes ​

Implementation Impact ​

Questions? ​