Transforming Retail: From On-Premises .NET to AWS-Native
Era Mío is a Uruguayan omnichannel retailer operating in the secondhand market—one of the fastest-growing segments in global retail, valued at US$393 billion and expanding three times faster than traditional retail. With brick-and-mortar stores in Uruguay and an e-commerce platform (eramio.com.uy), Era Mío allows its customers to both sell used items and purchase curated secondhand products—from clothing to accessories and more.
Despite being an SME, Era Mío’s operational complexity rivals that of much larger retailers. Every item in inventory is unique—the company does not use SKUs. Each item must be tracked individually from consignment receipt through pricing, multi-store visibility, omnichannel synchronization, supplier settlements with commission calculations, and regulatory-compliant electronic invoicing.
The challenge
Era Mío’s original technology stack was built on .NET and SQL Server, running on limited on-premises infrastructure. As the business grew—adding new branches and launching an e-commerce channel—the legacy platform became a bottleneck:
- There is no real-time inventory synchronization between physical stores and the e-commerce platform. For a secondhand retailer where every item is one-of-a-kind, selling an item in-store while it’s still listed online results in a failed order that can never be fulfilled—there is no option to “restock.”
- SQL Server licensing costs are disproportionate to the company’s revenue, consuming budget that could be reinvested in growth.
- On-premises infrastructure that required a capital investment for each new branch and could not scale to handle seasonal spikes in demand.
- Lack of business analytics capabilities —the company had no structured way to analyze sales patterns, inventory turnover, or supplier performance.
The solution
BigCheese partnered with Era Mío to design, build, and operate a comprehensive cloud-native retail platform on AWS—migrating from .NET/SQL Server to a containerized Java architecture that supports the entire lifecycle of the secondhand retail business.
Application Modernization
The core ERP system was re-architected from .NET/SQL Server to a Java web application running on Amazon ECS with Fargate, eliminating the need for server management and providing automatic scaling. The migration from SQL Server to Amazon RDS MySQL eliminated licensing costs entirely—a critical savings for an SME.
Multi-branch architecture
Each branch operates as an independent deployment with its own application cluster and configuration, sharing a centralized database with logically separate schemas. This provides complete operational isolation—a deployment issue at one branch does not affect the other—while optimizing database costs.
Omnichannel Integration with E-commerce
A two-way integration via REST API connects the ERP system to the e-commerce platform in real time:
- When an item is sold at any physical store, the e-commerce platform is immediately notified so that the listing can be removed
- When a customer makes a purchase online, the ERP system is notified to mark the item as sold and generate the electronic invoice
- A resilient retry mechanism handles temporary API outages, ensuring eventual consistency
- A secondary validation at checkout confirms the item’s availability before the purchase is finalized
This two-tier synchronization model effectively eliminates the double-selling of unique items—an existential risk for any secondhand retailer operating across multiple channels.
Automated batch processing
AWS Batch, with dedicated computing environments for each branch, handles scheduled operational tasks: supplier settlement calculations (with commission prorating by item for bulk sales), inventory synchronizations, and report generation—replacing manual processes that previously consumed significant staff time.
Business Intelligence
Amazon QuickSight connects to a dedicated read replica of the database, ensuring that analytical queries do not impact transactional performance. This provides Era Mío’s management team with self-service dashboards offering visibility into sales performance, inventory metrics, supplier analytics, and distribution data. For the first time, the business can make data-driven decisions across all its channels.
Automated Deployments
Automated deployments via AWS CodePipeline ensure that both branches receive updates simultaneously, with built-in manual approval gates and the ability to roll back automatically in less than 5 minutes.
Results
| Metrics | Previously (On-Premise) | Next (AWS) |
|---|---|---|
| Database Licensing | SQL Server Licensing Costs | $0 (MySQL open-source) |
| E-commerce Synchronization | Hours/days of delay, risk of double-selling | Real time ( |
| Payments to Suppliers | 5–7 business days, prone to errors | |
| Deployment of a New Branch | Hardware purchase + setup (weeks) | CloudFormation Template (hours) |
| Business Analytics | Does not exist | Self-Service Dashboards in QuickSight |
| Rollback of deployments | Manual, hours | Automated, ~5 minutes |
| Infrastructure Management | Manual patching, capacity planning | Fully managed (ECS Fargate, RDS) |
Why It Matters
The story of Era Mío shows that cloud-native modernization on AWS isn’t just for large companies. By migrating from a legacy .NET/SQL Server stack to a containerized, fully managed architecture, a growing SMB retailer was able to:
- Zero licensing costs through the adoption of open-source technology
- Omnichannel operations with real-time inventory synchronization — critical for the secondhand one-of-a-kind item model
- Operational scalability — adding new branches is now a matter of replicating infrastructure, not a hardware project
- Data-Driven Decision Making Through Integrated Business Intelligence
As the global resale market continues to accelerate and consumers increasingly expect seamless omnichannel experiences, Era Mío’s platform on AWS positions the company to scale its circular economy model across Latin America.