How to Cut Your AWS Bill by 40% in 30 Days: A Practical Guide

When a fintech client came to us last year, their AWS bill had grown from $8,000/month to $23,000/month in 18 months — without a proportional growth in users or revenue. Sound familiar?

After a 30-day optimization sprint, we brought it down to $13,800/month — a 40% reduction. Here’s exactly what we did, and how you can apply the same framework to your own infrastructure.

Step 1: Get Visibility First (Days 1–3)

You can’t optimize what you can’t see. Before touching anything, enable these two tools:

• AWS Cost Explorer — filter by service, region, and tag to find your top 5 cost drivers
• AWS Compute Optimizer — automatically identifies over-provisioned EC2 instances, Lambda functions, and EBS volumes
• Trusted Advisor — flags idle load balancers, underutilized instances, and unattached EIPs

In our fintech client’s case, 3 services (EC2, RDS, and data transfer) accounted for 78% of the bill. That’s where we focused.

Step 2: Right-Size Your EC2 Instances (Days 4–10)

Over-provisioning is the #1 source of AWS waste. Most teams provision for peak load and forget about it. Here’s the process:

1. Pull 30-day CPU and memory utilization from CloudWatch for every running instance
2. Any instance averaging below 20% CPU is a right-sizing candidate
3. Use Compute Optimizer’s recommendations as a starting point — but test in staging first
4. Consider moving to Graviton (ARM-based) instances for compatible workloads — typically 20% cheaper for the same compute

Step 3: Switch to Reserved Instances or Savings Plans (Days 11–15)

On-demand pricing is the most expensive way to run AWS. If you have predictable workloads (and most businesses do), switching to Reserved Instances or Compute Savings Plans can cut those costs by 40–60%.

• 1-year Reserved Instances (no upfront): ~30% discount vs on-demand
• 1-year Reserved Instances (all upfront): ~40% discount
• Compute Savings Plans: More flexible — covers EC2, Fargate, and Lambda with ~25% savings

Rule of thumb: Only commit to instances that have been running stably for 2+ months. Don’t commit to anything you might resize or retire.

Step 4: Clean Up Orphaned Resources (Days 16–20)

These are the silent budget killers — resources that were provisioned and forgotten:

• Unattached EBS volumes — run aws ec2 describe-volumes --filters Name=status,Values=available
• Old snapshots — set a lifecycle policy to delete snapshots older than 90 days automatically
• Idle load balancers — any ALB/NLB with 0 healthy targets for 7+ days
• Unused Elastic IPs — AWS charges $0.005/hour for unassigned EIPs (adds up fast)
• Stopped EC2 instances — still paying for attached EBS storage and EIPs

Step 5: Optimize S3 and Data Transfer (Days 21–28)

S3 is deceptively expensive at scale. Two quick wins:

• Lifecycle policies: Move objects older than 30 days to S3-IA (40% cheaper), and older than 90 days to Glacier (80% cheaper). Most businesses have years of logs sitting in S3 Standard.
• Data transfer optimization: If you’re transferring large amounts of data between regions, use S3 Transfer Acceleration only when needed. Better yet, architect your data to live closer to where it’s consumed.

The 30-Day Results

Where to Start Today

If your AWS bill is growing faster than your business, don’t wait. Even a quick 2-hour audit will likely surface $500–$2,000/month in low-hanging fruit.

The 5 steps above require no architectural changes — just cleanup, right-sizing, and smarter pricing. Most teams can execute this without disrupting production.