This case study explores how we contributed to optimizing the infrastructure of a leading Asian Buy Now, Pay Later (BNPL) provider. Operating in a market where debit card usage remains limited and traditional banks take a conservative approach to innovation, the company needed a resilient IT foundation capable of supporting a wide range of payment options. Our work centered on maximizing system availability and meeting stringent regulatory requirements, both of which are critical in a market that places exceptional value on reliability and user trust.
Client Profile
The client is a leading FinTech company in the Asia-Pacific region that offers Buy Now, Pay Later (BNPL) services. Their platform allows consumers to shop online without a credit card, relying only on a mobile number and email address for authentication. With a network of more than 700,000 merchants, the company stands out connecting digital commerce with Japan's convenience store ecosystem, enabling customers to settle monthly bills in a familiar, accessible way.
Challenges
The core challenge was to build a unified, high-scale notification platform capable of processing more than 50 million transactional messages each month. As a mission-critical part of its infrastructure, the system needed to handle SMS, push notifications, email, and voice communications with zero tolerance for downtime.
The project had to overcome three major hurdles:
Provider Complexity and Reliability
Integrating multiple third-party providers, each with its own APIs, queuing models, and performance characteristics, introduced significant operational overhead. For high-priority messages such as One-Time Passwords (OTPs), any provider outage could lock users out of their accounts, immediately affecting customer support volumes and eroding trust.
Regulatory Compliance and Auditability
Japan's tightly regulated financial environment required verifiable proof of delivery and strict adherence to communication time windows, such as restrictions on late-night debt-related notifications. The system needed a comprehensive audit trail for every action to satisfy both legal obligations and customer service expectations.
Cost and Traffic Optimization
To maintain healthy margins, the platform had to route traffic dynamically based on vendor-specific quotas, prioritizing certain providers until negotiated thresholds were met, then shifting to more cost-efficient alternatives.
Ultimately, our Client needed a solution that would hide this complexity behind a clean internal API while ensuring idempotency, low latency, and flawless automatic failover.
Solutions
To address these challenges, we designed and implemented a high-performance, event-driven notification platform. The solution centered on decoupling Client’s internal services from the complexities of external communication providers through a unified abstraction layer.
Core Architectural Components
We introduced a translation layer that normalizes disparate third-party APIs into a single internal protocol. This enables engineering teams to trigger notifications through a clean, consistent API without needing to account for provider specific nuances.
To ensure uninterrupted availability, we built a custom routing engine that continuously monitors provider health. When latency spikes or failures occur, the engine activates a circuit breaker to automatically reroute traffic; for example, switching SMS providers in real time. It also distributes traffic according to monthly volume quotas, helping the company meet negotiated pricing tiers with vendors.
Handling more than 50 million messages per month required an asynchronous architecture backed by message queues. Delivery reports are received through provider-specific webhooks, published to a queue, and processed by a dedicated service, ensuring tracking data is captured without slowing down the primary message flow.
All transactions and delivery reports are stored in a searchable database, giving customer support and compliance teams full visibility into message history, an essential capability for dispute resolution and meeting Japan's stringent financial regulations.
Reliability & Performance Optimization
To protect message integrity, we implemented strict idempotency controls to prevent duplicates, which is critical for avoiding issues such as double billing or repeated notifications.
The system was optimized for two distinct traffic patterns:
- Synchronous API: Low-latency paths for mission-critical messages such as OTPs.
- Asynchronous API: High-throughput queues for routine transactional updates.
The platform is fully instrumented with DataDog, enabling real-time visibility into provider performance, system health, and cost distribution.
Results
The project delivered a highly reliable platform that maintains strong delivery performance and the throughput required to support continuous user growth without service disruptions. With a more flexible infrastructure in place, our Client can now add or remove providers with ease, enabling the company to optimize operational costs and respond to business needs with far greater agility.
Are you looking for a way to modernize your business? Discover our services for financial platforms.
