Ingress / Case Studies / Cloud Optimization

From $3.1M to $0.7M. Same workload.

8-week FinOps engagement for a Series C SaaS company. Tagging, rightsizing, Savings Plans, and governance. $2.4M annual savings identified without capacity reduction.

SectorSaaS
ServiceCloud FinOps
Duration8 weeks
The Situation

Cloud costs grew faster than the company

A Series C SaaS company at $45M ARR and 180 employees watched their AWS bill balloon from $400K per year to $3.1M in just 18 months. Headcount had doubled, but costs had grown nearly 8x. The CTO knew something was wrong. They had no FinOps process, weak tagging discipline, and no leverage on their infrastructure.

The hidden costs of rapid growth

Fast-growing companies pay cloud at on-demand rates while they figure out how to optimize. By the time they realize the cost, they are locked into architectures that are hard to change. This company had no visibility into what was driving spend.

  • No tagging discipline. Only 20% of resources were tagged. Cost allocation was guesswork. Finance could not chargeback to teams.
  • Instance oversizing. Developers spun up large instances for testing. No discipline on shutting them down. Memory and CPU utilization tracked nowhere.
  • No Reserved Instances or Savings Plans. Everything on on-demand pricing. With $3.1M spend, that meant $1M+ in unrealized savings.
  • Spot Instance unused. CI/CD and non-production workloads could run on Spot at 70% discount. No one had implemented it.
  • Data transfer costs hidden. Egress from VPC to third-party APIs charged at $0.02/GB. Optimization potential not understood.
How We Optimized

Visibility, rightsizing, commitments, then automation

01
Cost Attribution & Tagging Gap Analysis
Analyzed current tagging to identify untagged resources by cost and criticality. Designed a tag taxonomy tied to cost centers, teams, and applications. Automated tagging enforcement via AWS Config rules on resource creation.
Week 1Discovery
02
Rightsizing Analysis & Implementation
Used CloudWatch metrics and AWS Compute Optimizer to identify oversized instances. Found 140 instances running at less than 20% CPU and memory. Right-sized 95% of them. Estimated 18% compute cost reduction.
Week 2-3Analysis
03
Savings Plans & Reserved Instance Strategy
Modeled Reserved Instance and Savings Plan options. Recommended All Upfront for predictable production workloads. Identified $1.2M in RI coverage for core database and application servers. Modeled 3-year break-even.
Week 3-4Purchasing
04
Spot Instance Strategy for Non-Prod
Analyzed CI/CD and development workloads for Spot Instance eligibility. Implemented Spot fleets for Jenkins agents and testing environments. Reduced non-prod compute by 65%.
Week 4-5Optimization
05
Chargeback Model & FinOps Governance
Designed a chargeback model allocating cloud costs to application teams and cost centers. Built cost dashboard in AWS Cost Explorer. Established monthly FinOps council to review optimization progress and set targets.
Week 5-6Governance
06
Automation & Ongoing Optimization
Built Lambda functions to terminate unused resources and scale down non-prod at end of day. Set up AWS Budgets and SNS alerts for cost overruns. Established quarterly optimization reviews with product teams.
Week 6-8Operations
What We Delivered

Cost reduction without capacity loss

$2.4M
Annual savings identified
78% of the cost overrun eliminated through rightsizing, commitments, and Spot optimization. Applied without reducing capacity or performance. Payoff on engagement in 2 weeks.
140
Instances rightsized
From over-provisioned on-demand to appropriately sized. Average rightsize reduced cost by 35% per instance. Performance maintained or improved through better utilization.
94%
Tagging compliance within 30 days
All resources tagged by cost center, team, and application. Automated enforcement via AWS Config. Zero new untagged resources created after week 1.
Why It Stuck

Ownership and visibility drive behavior

This company did not need a one-time cost reduction. They needed a FinOps culture. We built that by making costs visible, holding teams accountable, and automating the boring parts.

FinOps as a team sport

Engineering teams care about cost when they see it. We gave them a dashboard. Finance cares about allocation when they can track it. We gave them chargeback. The FinOps council made optimization a priority because the data showed it mattered.

  • Tagging at creation. AWS Config rules prevent untagged resources. Developers tag on creation, not retroactively.
  • Team-level dashboards. Each team sees their own costs daily. Incentive to optimize is immediate.
  • Automation, not policies. Instead of telling developers to shut down resources, Lambda scales them down automatically.
  • Quarterly business reviews. FinOps council meets monthly to approve optimizations and set targets. Cost tied to board metrics.
  • Capacity planning with cost. Infrastructure scaling discussions now include FinOps impact. Cost became part of architecture reviews.
Technology Stack

AWS tools for cost discipline

AWS Cost Explorer
Dashboard for cost attribution by tag, account, and service. Real-time visibility into trends and anomalies. Monthly FinOps council reviewed dashboards.
AWS Compute Optimizer
Machine learning recommendations for rightsizing. Identified oversized instances with low utilization. Recommendations integrated into monthly review process.
AWS Config + Lambda
Automated enforcement of tagging standards. Lambda functions for resource cleanup and non-prod auto-scaling. Infrastructure as code in Terraform.
AWS Budgets & Alerts
Monthly cost budgets by team and service. SNS alerts for forecast overruns. Gives FinOps council data for monthly reviews.
Start a conversation

Tell us what's worth doing.

// 30 minutes → a written brief.

Bring the problem. We'll come back with a written brief: what to build, what to defer, and where AI actually moves the number. No deck pitches.

Book a 30-min call →All case studies
Emailconnect@ingressits.com
GSA MAS#47QTCA26D000K
Reply< 24 hrs