Building the Loyalty Analytics Platform That Powers Targeted Campaigns Across AirAsia's Global Network

Data warehouse and ETL foundation: structured, secured, and built to run without engineering in the loop

The first architectural decision was the data layer: Snowflake as the warehouse, with security modules built across every component for storage of sensitive customer data, so user-based access was structural rather than a setting that could be misconfigured.

that foundation,

1. Talend handled ETL using SNL and configurable data-driven jobs, turning the pipeline from a custom engineering exercise into a replicable configuration.
2. The metadata models Ideas2IT developed were the key to the reporting layer: instead of handcrafted queries per report, campaign managers could compose ad-hoc reports against a model that already understood the loyalty data structure.
3. Initial reports ran through Jasper. Drill-down capability came via Tableau, giving campaign managers the depth to investigate anomalies in loyalty data without raising a ticket.

DELIVERABLES

• Snowflake data warehouse
• Talend ETL pipeline (SNL, configurable data-driven jobs)
• Metadata models for ad-hoc reporting
• Jasper canned report layer
• Tableau drill-down reports
• Role-based data access controls

Loyalty Point Calculation Engine: rules-based allocation, workflow approvals, and campaign feedback loop

The second layer was the rules engine governing how loyalty points were calculated, allocated, and approved. Ideas2IT built the Loyalty Point Calculation Engine on Drools, a production-grade rules engine that separated campaign logic from application code.

1. This meant campaign rules, tier thresholds, and point allocation logic could be updated by configuring rules rather than releasing code.
2. The workflow layer handled approvals and notifications at each step of the loyalty lifecycle, giving campaign operations teams visibility and control without touching engineering.
3. The final component was the feedback loop: campaign performance results fed back into the analytics engine, so each campaign cycle produced data that tuned the next one. Redis handled session and cache management to keep response times consistent across the platform under concurrent campaign load.

DELIVERABLES

• Drools Loyalty Point Calculation Engine
• Workflow approval and notification layer
• Campaign performance feedback pipeline
• Redis cache layer
• ReactJS + Redux campaign management UI
• Role-based campaign allocation and override rules