How We Reduced Global AWS Spend by 32% While Scaling Multi-Country Digital Platforms

A FinOps Case Study on Governance, Automation, and Cloud Efficiency at Scale

Author: Aloisio Bastos Role: Product & Platform Strategy | FinOps | Cloud Governance | DevEx

Executive Summary

Managing cloud costs at scale is no longer just a financial exercise — it has become a strategic engineering capability.

In this case study, I’ll share how a global platform organization redesigned its cloud financial governance model to optimize AWS costs across multiple countries, products, and engineering squads.

The initiative combined:

• FinOps governance
• Platform engineering
• Automated cloud operations
• Cost allocation strategy
• Real-time observability
• Executive reporting

The result was a 32% reduction in AWS operational costs, improved engineering accountability, and the creation of a sustainable cloud governance framework.

Business Context

The organization operated a large-scale digital ecosystem across multiple countries, supporting:

• Microservices-based applications
• Real-time event streaming
• Containerized workloads
• Data platforms
• Search engines
• Payment and telemetry APIs

The cloud architecture relied heavily on:

• Amazon EC2
• Amazon ECS
• Amazon Aurora
• Amazon MSK (Kafka)
• Amazon OpenSearch
• AWS Transit Gateway
• Amazon S3
• AWS Athena
• AWS QuickSight

As the platform scaled, AWS costs began increasing faster than business growth.

The company faced several challenges:

Challenge Impact Lack of tagging standardization Impossible to allocate costs accurately Idle non-production environments. High waste during nights/weekends, Oversized EC2 and ECS workloadsLow utilization rates, Kafka and OpenSearch overprovisioning. Excessive infrastructure costs, Limited cost visibility, slow executive decision-making, high inter-region traffic, elevated networking expenses

Initial Architecture Overview

Global AWS Platform Architecture

Step 1 — Building a Global Tagging Governance Strategy

One of the first priorities was establishing a standardized tagging strategy.

Without consistent tagging, cloud costs become invisible.

We created mandatory tagging policies for:

Tag Purpose Domain Business area Squad Engineering ownership Country Regional allocation ProductProduct accountability Environment Prod / Staging / Dev Criticality Operational priority CostCenterFinancial ownership

AWS Tagging Governance Model

Step 2 — EC2 and ECS Optimization

After enabling visibility, we identified several workloads running with:

• CPU utilization below 20%
• Memory overprovisioning
• Idle ECS tasks
• Legacy EC2 instance families

Optimization Actions

EC2

• Migrated to Graviton-based instances where possible
• Rightsized underutilized instances
• Introduced Reserved Instances and Savings Plans
• Automated nightly shutdown for non-production workloads

ECS

• Reduced task overprovisioning
• Implemented autoscaling policies
• Optimized CPU/memory reservations
• Removed obsolete services

ECS Autoscaling Architecture

Step 3 — Kafka (MSK) and OpenSearch Cost Optimization

The organization relied heavily on event-driven architecture.

However, Kafka clusters and OpenSearch storage were significantly oversized.

Kafka Optimization

Improvements

• Reduced broker overprovisioning
• Optimized partition strategy
• Enabled data retention lifecycle policies
• Reduced replication where operationally acceptable
• Improved topic governance

OpenSearch Optimization

Improvements

• Migrated warm indexes to UltraWarm storage
• Reduced unnecessary replicas
• Archived historical data to S3
• Optimized shard sizing
• Implemented index lifecycle management

Event-Driven Cost Optimization Flow

Step 4 — FinOps Automation

One of the biggest gains came from automation.

The company previously depended on manual operational routines to manage environments.

This created:

• Human error
• Delayed shutdowns
• Unnecessary weekend costs
• Lack of governance consistency

Automated Shutdown Policy

We implemented serverless automation using:

• AWS Lambda
• CloudWatch Events
• AWS Systems Manager
• Tag-based policies
• Holiday-aware scheduling logic

Automation Rules

Environment Shutdown Policy, Development Automatic shutdown from 8 PM to 7 AM, Development Full shutdown during weekends and national holidays, Staging / Test Weekend and holiday shutdown, enabled Sandbox Auto-delete after inactivity Production Excluded.

The automation engine integrated calendar-aware logic to identify:

• Regional holidays
• Weekends
• Non-business operational windows

This allowed the organization to eliminate unnecessary cloud consumption during periods with no engineering activity.

The solution dynamically evaluated environment tags and business calendars before triggering ECS service scaling or EC2 shutdown operations.

AWS Automation Flow

Step 5 — Anomaly Detection and Budget Governance

Cost optimization is not sustainable without proactive monitoring.

We implemented:

• AWS Budgets
• Cost anomaly detection
• Slack integrations
• Executive notifications
• Forecasting alerts

The platform engineering teams began receiving alerts in real time whenever abnormal consumption patterns appeared.

This transformed cloud cost management from a reactive process into a continuous operational discipline.

FinOps Monitoring Architecture

Step 6 — Executive Dashboards with Athena and QuickSight

Visibility drives accountability.

To improve decision-making, we developed executive dashboards using:

• AWS Athena
• AWS Cost Explorer
• Amazon QuickSight

The dashboards provided:

• Cost by country
• Cost by product
• Cost by squad
• Daily consumption trends
• Forecasted monthly spend
• Infrastructure utilization indicators
• Savings tracking

Example Dashboard KPIs

KPIResultGlobal AWS Spend Reduction32%ECS Optimization Savings18%Kafka Infrastructure Savings24%OpenSearch Savings29%Non-Production Waste Reduction65%Cloud Cost Allocation Accuracy95%

Executive Dashboard Layout

Cultural Transformation

One of the most important outcomes was not purely financial.

The initiative changed engineering behavior.

Before the transformation:

• Cost discussions happened only with finance teams
• Engineers had limited visibility into cloud consumption
• Optimization was reactive

After implementing FinOps governance:

• Engineering squads owned their cloud spend
• Product leaders gained real-time visibility
• Platform teams established architectural standards
• Cloud efficiency became part of the software delivery lifecycle

This created a sustainable FinOps culture.

Key Lessons Learned

1. Visibility Comes Before Optimization

Without tagging and dashboards, optimization becomes guesswork.

2. Automation Delivers Consistency

Manual governance does not scale.

Serverless automation dramatically improved operational discipline.

3. FinOps Is a Product Capability

Cloud cost management should not be treated as a finance-only activity.

It must be integrated into:

• Platform engineering
• Product management
• DevOps
• Architecture reviews

4. Shared Accountability Changes Behavior

When squads can see their costs clearly, engineering decisions improve naturally.

Final Results

The global FinOps initiative achieved:

✅ 32% AWS cost reduction ✅ Improved infrastructure efficiency ✅ Executive-level cloud visibility ✅ Automated operational governance ✅ Sustainable financial accountability ✅ Stronger collaboration between product, engineering, and finance

Most importantly, the organization created a scalable governance model capable of supporting future international growth.

Recommended AWS Services Used

Area AWS ServicesComputeEC2, ECS, LambdaDatabasesAuroraStreamingAmazon MSKSearchAmazon OpenSearchAnalyticsAthena, QuickSightMonitoringCloudWatchGovernanceAWS Budgets, Cost ExplorerNetworkingTransit Gateway, VPCNotificationsSNS

Final Thoughts

FinOps is no longer optional for organizations operating at scale.

Companies that combine:

• Governance
• Engineering excellence
• Automation
• Data visibility
• Product thinking

will build sustainable and cost-efficient cloud ecosystems.

The future of cloud operations belongs to organizations capable of transforming financial governance into an engineering discipline.

About the Author

Aloisio Bastos is a Product & Platform Strategy professional with experience in:

• Product Management
• FinOps
• Platform Engineering
• DevEx
• Cloud Governance
• CI/CD Modernization
• AWS Architecture
• Observability and Operational Excellence

Focused on building scalable, efficient, and data-driven digital platforms.