Typical AWS bills: 30–60% leak.
Audit → quick wins → structural fixes. No feature loss. Same engineer who cut Imohub infra 70% and Cuez 40%.
Who this is for
CTO or CFO with AWS bill growing faster than revenue — $20k–$200k/month with no clear path to cut, feature work cannot stop for the cleanup.
The pain today
- AWS bill climbed 30%+ year-over-year while usage didn't
- Reserved Instances and Savings Plans 'someone should set up' never scheduled
- Data transfer costs mysterious and growing
- Cost Explorer overwhelming; no clear 'these are your top 5 to fix'
- FinOps platform evaluated but adds cost without reducing bill
The outcome you get
- 30–60% AWS bill reduction within 4–6 weeks, no feature loss
- Reserved Instances and Savings Plans correctly sized and scheduled
- Data transfer audit with specific fixes for top cost centers
- Auto-scaling and right-sizing applied where safe
- Cost monitoring so the bill doesn't creep back up
Where AWS bills leak money
Patterns across multiple cost audits. Compute: oversized EC2 or RDS instances running 24/7 when auto-scaling would cut 40%. Storage: unlimited S3 retention, EBS volumes on terminated instances, snapshots years old. Data transfer: inter-AZ traffic, NAT gateway egress, CloudFront vs direct S3 decisions. RDS: overprovisioned read replicas, Multi-AZ when single-AZ would suffice for non-prod, IOPS set to max because defaults felt scary. Lambda: warm Lambda at high concurrency costing more than EC2 would. The specific leaks vary per architecture; the pattern is consistent — 30–60% of most mid-sized AWS bills is leak, not value.
Quick-win levers vs structural
Quick wins (week 1–2, 15–30% typical reduction). Right-sizing (audit instance utilization, resize where usage is consistently below 40%). Scheduling (stop non-prod outside business hours, 40–60% savings on dev/staging). S3 lifecycle policies (transition cold data to Glacier, typical 70% storage savings on archives). Unused resources (terminated-instance EBS volumes, unused Elastic IPs, old snapshots). Structural wins (week 3–6, 15–30% additional). Reserved Instances or Savings Plans for committed workloads (30–50% compute savings, 1–3 year commitment). Database right-sizing and Multi-AZ review. Application-layer optimization (caching, query reduction — overlaps with API refactor work). Data transfer architecture review (move chatty services to same AZ, optimize CloudFront usage).
Multi-region, Savings Plans, Reservations
Savings Plans vs Reserved Instances: Savings Plans are the modern default (more flexible across instance family, operating system, region), Reserved Instances sometimes better for very stable workloads. Commitment tier: 1-year partial upfront usually the sweet spot — 40% savings with only 50% of the cash commitment. 3-year all-upfront for maximum savings (55%+) when cash-rich and load is stable. Multi-region: usually cost-adds, not cost-cuts, because traffic crossing regions costs more. I recommend multi-region only when business requires it (data residency, DR), not as a cost-optimization tactic. Savings Plans need recalculation quarterly as workload evolves.
Case studies: Cuez 40%, Imohub 70%
Cuez: infrastructure cost cut ~40% as a secondary outcome of the API performance work (3s → 300ms). Better queries meant fewer resources needed to serve the same traffic. Application-layer optimization drove infra savings. Imohub: 70% infrastructure cost reduction as a primary outcome of the rebuild (legacy Imóveis SC → Imohub on Next.js + Laravel + Meilisearch + MongoDB + AWS). Architecture-level optimization — right tool for right job (Meilisearch for search instead of ElasticSearch), right instance sizing, efficient data patterns. Both engagements combined quick wins with structural changes. Quick wins alone would have saved maybe 25%; structural changes drove the rest.
Pricing
AWS cost reduction engagements fit the Fractional CTO service at Advisory ($4,500/mo) for audit + quick wins + recommendation, or Fractional ($8,500/mo) for deeper involvement implementing structural changes alongside your team. Typical engagement: 4–6 weeks to deliver initial cost reduction, then optional ongoing monitoring and quarterly recalculation. ROI math: if your current AWS spend is $20k/mo and the engagement cuts 40% ($8k/mo saved), the retainer pays for itself in weeks. For sub-$10k/mo AWS spend, the math often doesn't justify the engagement — I'll say honestly if your bill is too small for consulting to make sense.
What I don't touch
AWS cost optimization has dead-end paths. I don't rewrite applications purely for cloud cost — the risk vs savings math rarely works for $5k/month optimizations. I don't migrate off AWS for cost unless the business case is clear (very-rare, usually requires commitment to lock-in on replacement). I don't recommend spot instances for production workloads without careful analysis (interruption cost vs savings). I don't engage with vendor savings plans (Reserved Virtual Machines from random middlemen) that promise big savings for unclear operational risk. Honest scoping keeps the engagement focused on durable savings rather than one-time accounting wins.
Recent proof
A comparable engagement, delivered and documented.
Rebuilt a real estate portal at a fraction of the cost
Rebuilt Imóveis SC's real estate portal as ImoHub — a faster, more scalable successor — handling 120k+ properties with sub-second search and drastically reduced AWS costs.
Frequently asked questions
The questions prospects ask before they book.
- What % savings should I expect?
- Typical 30–60% of monthly bill on first engagement. Quick wins alone usually deliver 15–30%; structural changes deliver the rest. Very well-optimized environments see less (maybe 10–20%). Very leaky environments see more (sometimes 60%+). I share a realistic target range after week-1 audit.
- Do I need a FinOps platform?
- For mid-size AWS bills ($10k–$100k/mo), the AWS Cost Explorer, Compute Optimizer, and Cost Anomaly Detection are usually enough once properly configured. FinOps platforms (Cloudzero, Vantage, Zesty) add value above $100k/mo where multi-dimensional analysis and automated recommendations compound. I recommend vs. building based on your scale.
- Will you work with our AWS account directly?
- Yes, with least-privilege IAM access scoped to what I need (Cost Explorer, read-only on resources, write access only when mutually agreed for specific changes). Or I can work advisory-only with screenshare sessions where your team makes the changes. Many clients prefer advisory to limit access.
- Can you do this without access to the application code?
- Partially. Infrastructure-level wins (right-sizing, scheduling, RIs) don't need code. Application-level wins (caching, query optimization) do. For code-free engagements, expect savings to cap around 30–40% rather than the 60%+ that includes application optimization.
- What about GCP or Azure?
- Same principles, different specific levers. GCP Committed Use Discounts, Azure Reserved Instances. I've worked primarily with AWS but the patterns transfer. For a pure GCP or Azure engagement, I'd scope the first week as learning-your-environment rather than diving into known AWS patterns.
Ready to start?
Tell me what you need in 60 seconds. Tailored proposal in your inbox within 6 hours.