From 0 to 30% Savings: A 90-Day FinOps Journey

"How long until we see savings?"

It's the first question every CFO asks. Here's the honest answer: you can see meaningful savings in 90 days if you focus on the right things in the right order.

This is the 90-day playbook I use with every new client.

Days 1-14: Foundation

Week 1: Visibility

Before you optimize, you need to see.

Day 1-2: Enable Cost Allocation Tags

Go to AWS Billing → Cost Allocation Tags. Activate:

• aws:createdBy (who created it)
• Environment (prod/staging/dev)
• Service (application name)
• Owner (accountable person)

Day 3-4: Set Up Cost Explorer

Create saved reports for:

• Daily costs by service
• Weekly costs by account
• Monthly costs by tag

Bookmark these. You'll use them constantly.

Day 5: Budget Alerts

Create AWS Budgets for:

• Total monthly spend (alert at 80%, 100%, 120%)
• Per-account or per-team budgets
• Service-specific budgets for your top 5 services

Day 6-7: Anomaly Detection

Enable AWS Cost Anomaly Detection. It's free and catches weird spending patterns automatically.

Week 2: Assessment

Day 8-10: The Zombie Hunt

Find and document:

• Stopped EC2 instances with EBS
• Unattached EBS volumes
• Unused Elastic IPs
• Old EBS snapshots
• Empty S3 buckets
• Idle load balancers

Don't delete yet. Document.

Day 11-12: Commitment Inventory

What Savings Plans and Reserved Instances do you have? What's your coverage percentage? When do commitments expire?

Day 13-14: Top 10 Deep Dive

Your top 10 spending services are 80% of your bill.

For each one, understand:

• What it does
• Who owns it
• Whether it's sized appropriately
• What optimization options exist

Days 15-45: Quick Wins

Week 3-4: Delete the Waste

Now delete (carefully) what you found:

Zombie resources: Delete or archive. For anything you're uncertain about, snapshot it first. Unattached volumes: If unclaimed for 14 days, delete. Old snapshots: Implement lifecycle policies. Delete anything past retention.

Expected savings: 3-8% of total spend.

Week 5-6: Right-Size the Obvious

Look at CloudWatch metrics for your top compute consumers:

• Average CPU under 30%? Probably oversized.
• Memory consistently under 40%? Probably oversized.

Start with non-production. Downsize by one tier. Monitor for a week. Then production.

Expected savings: 5-15% of compute spend.

Week 7: Commitment Optimization

Based on your assessment:

If coverage is under 60%:

• Buy Compute Savings Plans for 70% of stable baseline
• Start with 1-year terms

If you have RDS/ElastiCache without RIs:

• Purchase Reserved Instances for databases

Expected savings: 20-40% on committed resources.

Days 46-75: Architectural Quick Hits

Week 8-9: gp2 → gp3 Migration

Identify all gp2 EBS volumes. Migrate to gp3.

This is:

• 20% cheaper
• No downtime required
• Takes minutes per volume

Expected savings: 20% on EBS storage.

Week 10-11: Graviton Where Possible

Identify workloads that can run on Graviton:

• Containerized applications
• Supported languages (Node, Python, Java, Go)
• No x86-specific dependencies

Start with non-production. Validate. Move to production.

Expected savings: 20-40% on migrated compute.

Week 12: Dev/Test Scheduling

Implement auto-shutdown for non-production:

• Stop resources nights and weekends
• Use AWS Instance Scheduler or Lambda-based solution
• Keep override mechanism for exceptions

Expected savings: 60-70% on non-production compute.

Days 76-90: Sustain

Week 13: Process & Governance

Weekly cost review: 30 minutes, review anomalies and trends. Monthly optimization check:

• Review commitment utilization
• Run zombie hunt
• Check for new quick wins

Quarterly deep dive:

• Assess architectural opportunities
• Review commitment expirations
• Plan next quarter's optimizations

Week 14: Documentation & Training

Document:

• What you optimized
• How to maintain it
• Who owns what

Train the team:

• How to read Cost Explorer
• How to respond to alerts
• Where to find help

The 90-Day Scorecard

Here's what a typical 90-day journey achieves:

| Category | Savings | |----------|---------| | Zombie resources | 3-5% | | Right-sizing | 5-10% | | Commitments | 10-15% | | Storage optimization | 2-3% | | Graviton migration | 3-5% | | Dev/test scheduling | 3-5% | | Total | 26-43% |

The exact numbers depend on how much waste you're starting with. Companies with no prior optimization see higher savings. Companies with some maturity see lower (but still meaningful) gains.

Common Pitfalls

Going Too Fast

Resist the urge to change everything at once. A botched right-sizing that causes an outage will set your entire FinOps program back.

Skipping the Foundation

Visibility and budgeting feel like overhead. They're not. You can't sustain savings without them.

Declaring Victory at Day 90

Day 90 is the beginning, not the end. Without ongoing maintenance, costs creep back up.

Not Celebrating Wins

Every saved dollar should be celebrated. This builds momentum and culture.

After 90 Days

The first 90 days capture the "easy" savings. What comes next:

• Deeper architectural optimization
• Application-level efficiency
• Contract negotiations (EDPs)
• Advanced automation
• Unit economics optimization

These take longer but can deliver another 20-30% over the following year.

Start Now

The best time to start was yesterday. The second best time is today.

Don't wait for perfect visibility. Don't wait for organizational buy-in. Start with what you can control.

Delete one zombie resource today. That's progress.