60% Reduction in Monthly Cloud Costs | DollarDash
ABOUT THE CLIENT
DollarDash is an overseas fund transfer company focused on simplifying global money movement. The platform enables individuals to make international transactions efficiently, helping bridge the gap between disparate financial systems and global currencies. Through streamlined cross-border payments, DollarDash aims to make global transactions more accessible, reliable, and frictionless for users navigating international finance.
*All names and logos have been changed to respect the NDA
OVERVIEW
Our team inherited an AWS environment consisting of our original bootstrapped infrastructure and components handed over from previous teams. Within a single quarter, we reduced the monthly cloud spend from $8,100 to $3,300. This 60% reduction generated over $57,000 in annual savings.
No features were removed. No SLAs were compromised. Production remained stable throughout.
The engagement focused on the infrastructure lifecycle post-build, aligning existing systems with the operational requirements of the product.
Reduction in monthly cloud costs
Saved per month
Annual savings
BUSINESS
REQUIREMENT
The client’s primary requirement was to reduce cloud costs that had grown higher than expected, without impacting production stability or SLAs.
Key Requirements
The business goals included:
- Reducing monthly AWS spend
- Ensuring zero production downtime during optimization
- Aligning infrastructure with current usage patterns rather than historical assumptions
SOLUTION
GeekyAnts approached the problem as a system-alignment exercise rather than a pure cost-cutting effort.
1. Audited infrastructure usage and spend across all environments
2. Removed unused, idle, and legacy resources safely
3. Right-sized compute, databases, and storage based on real usage
4. Reworked non-production environments to operate on demand
5. Introduced governance and maintenance processes to prevent cost drift
CHALLENGES
IN EXECUTION
& SOLUTIONS
To address the challenge of escalating operational expenses, we conducted an audit of inherited infrastructure assumptions, realigning them with actual usage patterns to eliminate waste. This process targeted over-provisioned non-production environments, which were right-sized and converted to on-demand models, ensuring resources remain active only when necessary.
A systematic cleanup was initiated to identify and remove legacy and orphaned resources, utilizing a shared audit and approval tracking system to maintain transparency. To mitigate the risk of breaking production during these transitions, all modifications were executed through Terraform-backed rollouts, ensuring every change was evidence-based and reversible.
Escalating Expenses
1
Infrastructure Assumptions
2
Over-provisioned Environments
3
Production Risk
4
OUR APPROACH
We followed a phased, low-risk approach with clear milestones to ensure cost reduction without compromising system stability.
- Cleanup and hygiene
- Evidence-based right-sizing
- Environment strategy restructuring
- Maintenance mode and process correction
- Post-Implementation Cost Control
Cleanup and hygiene
We began with the safest set of actions: removing unused and orphaned resources that had no active dependencies or clear ownership.
Actions included:
- Removing idle load balancers, unattached Elastic IPs, and unused networking resources
- Cleaning up old snapshots and backups
- Adding S3 lifecycle policies for storage tiering and expiration
- Applying ECR lifecycle policies to remove unused container images
- Reviewing logging policies
Evidence-based right-sizing
Using CloudWatch metrics, we analyzed CPU, memory, network, and database usage across environments to identify over-provisioned resources.
Changes included:
- Right-sizing non-production environments
- Reducing ECS task replicas where traffic patterns no longer justify them
- Removing unnecessary database read replicas
- Disabling excessive monitoring where it no longer added value
Environment strategy restructuring
Production changes were limited to resources with prolonged low utilization and predictable traffic.
Examples:
- Reducing app replicas for admin-facing systems
- Right-sizing analytics-related database replicas
All changes were incremental and closely monitored.
Maintenance mode and process correction
We challenged the assumption that all environments must run 24/7.
Key changes:
- Removed always-on QA and Pre-Prod environments
- Standardized infrastructure using Terraform for fast recreation
- Retained a single staging/dev environment when actively needed
This aligned infrastructure availability with actual usage patterns.
Post-Implementation Cost Control
To prevent cost drift, the system was placed into maintenance mode.
Process changes included:
- On-demand environments as the default
- Explicit justification for always-on non-prod resources
- Periodic cost reviews
- Checklists for growing data and storage
- CloudWatch alarm setups
PROJECT
RESULTS
The optimization project transitioned DollarDash’s AWS infrastructure from a high-overhead legacy configuration to a high-velocity system. Reducing monthly cloud spend by 60% unlocked over $57,000 in annual capital for reinvestment. This was achieved without compromising production stability, demonstrating that infrastructure maturity and fiscal efficiency are mutually inclusive.
Reduction in monthly cloud costs
Saved per month
Annual savings
OTHER CASE STUDIES
3x Faster AI Feature Iteration for Smart Pantry
How GeekyAnts helped Smart Pantry achieve 3x faster iteration on AI-driven features with a scalable, personalized meal recommendation platform.
50% Reduction in manual validation cycles| Nexus
Discover how GeekyAnts reduced manual validation by 50% and accelerated testing by 30% using AI agents, RAG models, and BPM automation.
Chronicle Stream: Video Journalism Platform Modernization
Chronicle Stream Modernization: Achieved 20x faster search and 1s publish times. See how GeekyAnts built a scalable video journalism platform with React and Node.js.