Overview
- Client: Fortune 500 Retailer
- Industry: Retail / E-Commerce
- Cloud Provider: AWS
- Challenge: Rising infrastructure costs across 12,000+ EC2 instances, 500TB+ S3 storage, and thousands of EBS volumes with no visibility into waste
One of the largest retailers in the United States had scaled its AWS footprint aggressively to support a rapidly growing e-commerce platform, in-store digital systems, and supply chain analytics. Annual cloud spend had crossed $11M across EC2, S3, and EBS alone, but the infrastructure team lacked granular visibility into where waste was accumulating. Leadership needed actionable data -- not dashboards -- to drive measurable cost reduction without impacting performance.
The Challenge
Years of rapid growth and seasonal scaling had created a sprawling cloud environment with deep inefficiencies embedded across three core AWS services:
- EC2 sprawl -- The fleet had grown to over 12,000 instances across 14 AWS accounts. Many instances had been provisioned for peak holiday traffic and never right-sized afterward, leaving thousands running at 15-30% average utilization.
- S3 storage accumulation -- More than 500TB of data sat across hundreds of S3 buckets with no lifecycle policies in place. Log archives, outdated media assets, and legacy data exports were stored in S3 Standard, costing over $2M per year in storage fees alone.
- EBS volume waste -- Infrastructure changes and instance terminations had left over 800 EBS volumes completely unattached. An additional 1,200 volumes were provisioned as gp3 with performance characteristics far exceeding their actual workload requirements.
- Tooling gaps -- Existing cost management tools provided account-level summaries and basic utilization flags but could not detect workload-level waste patterns, cross-service inefficiencies, or provide the migration-safe recommendations that engineers needed to act with confidence.
The Solution
PointFive's agentless platform connected to the retailer's AWS environment in under four hours through a read-only IAM role -- no agents, no code changes, and no impact on production workloads. The DeepWaste Detection Engine immediately began analyzing usage telemetry, billing data, and workload behavior patterns across all three services simultaneously.
Key findings and recommendations included:
- EC2 right-sizing -- PointFive identified 2,400+ instances over-provisioned by 40-70% based on actual workload behavior over 30-day windows. Each recommendation included a specific target instance type and a zero-downtime migration path aligned with the retailer's deployment tooling.
- S3 lifecycle optimization -- Analysis revealed $1.8M per year in storage costs tied to cold data that had not been accessed in 90+ days. PointFive recommended tiered lifecycle policies to move data into S3 Glacier and Intelligent-Tiering, with projected savings validated at the bucket level.
- EBS volume cleanup -- The platform flagged 800+ unattached volumes representing pure waste and identified 1,200+ attached volumes where switching from gp3 to gp2 or st1 would reduce costs by 35-50% without affecting I/O performance requirements.
Results
PointFive delivered its initial findings within 48 hours of deployment, giving the retailer a prioritized action plan across all three services:
- $4.2M in annualized savings identified and validated within the first 48 hours
- 38% reduction in combined EC2, S3, and EBS spend within 60 days of implementation
- ROI exceeding 500% within the first 30 days of acting on recommendations
- Zero downtime throughout the entire optimization process across all 14 AWS accounts
- Ongoing automated detection now continuously monitors for new waste patterns, preventing cost drift as the environment scales for future growth
Conclusion
Generic cost tools had given this retailer a false sense of optimization. Surface-level utilization alerts and account-level summaries missed millions in waste buried across over-provisioned instances, cold storage, and orphaned volumes. PointFive's DeepWaste Detection analyzed workload behavior at a depth that uncovered savings opportunities invisible to traditional tooling -- and delivered those findings within the 48-hour guarantee. For enterprises running complex, multi-service AWS environments, the gap between what basic tools report and what actually exists can represent millions in recoverable spend.