Case study · Loyalty SaaS platform · In production
VecinoClub
Serverless loyalty platform for local businesses in Chile
The problem
Local businesses need to build relationships with returning customers beyond each individual sale.
Many have no way to thank loyal customers or know who they are and how often they return.
Imaquinaria, a manga and Japanese culture shop in Temuco, is VecinoClub’s first business and proving ground.
The platform in use
Real interface with sample data. Camila Rojas is a fictional person; balances, purchases, and rewards do not belong to real customers.
VecinoClub × Imaquinaria
An experience that follows every visit
From the first scan to the next reward: this is how VecinoClub appears on a customer’s phone.
Customer journey
Explore five steps to see the screens and decisions behind the experience.
Step 01 of 05
Sign up and scan
- The customer scans the business’s QR code to enter in a browser.
- The club and its benefits are available on their phone, with no app to install.
Mobile views
Phone screens
Swipe or use the arrows to cycle through each screen in light and dark themes.
The solution
Customer experience
- — Scan a QR code to get started, with no app to install
- — Check stamps and points on a phone
- — Review purchases and available benefits
- — Redeem rewards from the business
Business dashboard
- — Define benefits and program rules
- — Register purchases at the counter
- — Manage rewards and stamps
- — View customer and purchase statistics
Architecture
GitHub Actions — deploys frontend (S3 / CloudFront) and backend (sam deploy). Manual gate required before production.
Key technical decisions
Multi-tenant from day one
The system was designed for more than one business from day one: every query carries a business identifier, so adding another business means registering it, without changing the business logic. Imaquinaria remains the first club, and the Bushidō program is one configuration of the platform, not its limit.
DynamoDB single-table
ADR-0006More than ten entities share one table. PK/SK keys follow access patterns, and every query includes the business identifier.
Pipeline with manual gate
ADR-0023GitHub Actions deploys to staging automatically. Production requires manual approval via issue. Five bugs caught in staging before reaching the live system.
Data migrations as code
ADR-0019Idempotent Python runner integrated in the CI/CD pipeline. Every DynamoDB data change goes through a PR, is versioned in git, and is applied automatically on deploy — zero manual post-deploy steps.
Production — real metrics
CloudWatch data · Lambda + API Gateway + DynamoDB · last 30 days
Architecture Decision Records
Every non-obvious technical decision was recorded with context, alternatives considered, and consequences. The project has accumulated 23 ADRs. The four most representative ones are available as public documents:
Use case: Imaquinaria’s Bushidō
The Imaquinaria Bushidō
Level system with points multipliers — the multipliers are not arbitrary: they are calibrated so that reward redemption is attractive at exactly the average purchase volume of an Imaquinaria customer.