The Challenge
A national telecommunications operator serving 8 million subscribers was losing market share to digital-native competitors. Their BSS/OSS landscape had grown organically over 15 years into a fragmented ecosystem of 12 platforms: separate systems for mobile billing, fixed-line provisioning, CRM, order management, inventory, and customer care.
The consequences were severe:
- 72-hour average order-to-activate time for new services, compared to minutes for digital competitors
- Customer data scattered across 7 databases, making a unified customer view impossible
- €15 million annual maintenance cost for legacy platforms, with 80% of IT budget consumed by keeping the lights on
- Inability to launch 5G services due to rigid, monolithic architecture that could not support dynamic service orchestration
Our Approach
Architecture Assessment (5 weeks)
Our assessment focused on three dimensions:
Business Capability Mapping: We mapped 180+ business capabilities to their supporting systems, revealing extreme duplication. For example, “customer identity management” was implemented differently across 5 platforms.
Integration Topology Analysis: We documented 340+ point-to-point integrations between systems — a classic integration spaghetti pattern. Most integrations were file-based batch transfers, introducing latency and data inconsistency.
Technology Health Assessment: Each platform was evaluated against seven fitness criteria: scalability, maintainability, security posture, vendor viability, total cost of ownership, API readiness, and cloud compatibility. Six platforms scored “critical” — immediate replacement recommended.
Target Architecture
We designed a unified BSS platform architecture based on TM Forum Open Digital Architecture (ODA) principles:
API-First Integration Layer: An API gateway and event mesh replacing all point-to-point integrations. All capabilities exposed through standardized TMF Open APIs, enabling rapid partner ecosystem integration.
Unified Customer Platform: A single customer domain service replacing seven fragmented customer databases. Built on a CQRS pattern — write-optimized for transactional operations, read-optimized for real-time customer 360 views.
Digital-First Order Management: Event-driven order orchestration engine supporting both traditional and 5G network slicing-based service orders. Configurable workflow engine enabling business users to define new order types without engineering involvement.
Cloud-Native on Azure: Deployed on Azure Kubernetes Service with Terraform-managed infrastructure. Chose Azure for existing enterprise agreement and strong hybrid connectivity to on-premises network infrastructure.
Migration Strategy
Rather than a platform-by-platform migration, we organized by business capability domains:
- Customer Domain (Months 1-4): Unified customer identity, profile, and preference management
- Product & Catalog Domain (Months 3-6): Configurable product catalog supporting 5G service bundling
- Order & Fulfillment Domain (Months 5-9): Digital-first order management with real-time provisioning
- Billing & Revenue Domain (Months 8-12): Convergent billing across mobile, fixed, and 5G services
- Care & Engagement Domain (Months 10-14): Omnichannel customer care with AI-assisted resolution
Each domain migration followed a four-step pattern: Expose (API facade over legacy), Extract (migrate logic to new platform), Verify (parallel run with automated comparison), Cutover (traffic switch with automated rollback capability).
Results
- Order-to-activate reduced from 72 hours to under 15 minutes for standard service orders, with complex orders completing in under 2 hours
- 12 legacy platforms consolidated into a unified cloud-native BSS with a single API layer
- 22% reduction in customer churn attributed to improved service experience and faster issue resolution
- €8.5 million annual operational cost savings from platform consolidation and cloud economics
- 5G service launch capability: First 5G network slicing-based enterprise services launched within 3 months of platform completion
Architecture Highlights
The unified event mesh was the lynchpin of the entire transformation. By routing all inter-domain communication through Apache Kafka, we achieved:
- Real-time data consistency across all domains
- Complete event audit trail for regulatory compliance
- Ability to replay events for testing and debugging
- Foundation for future AI/ML-driven automation
The CQRS pattern for the customer domain was particularly impactful: customer care agents went from waiting 45 seconds for a customer view (aggregated across 7 systems) to sub-second response times with a complete history.