CASE STUDY

Rakuten Advertising

How Multi-Cloud Architecture Enabled 132% Revenue Growth

Affiliate marketing platform transformed from on-premise infrastructure to multi-cloud platform, enabling rapid market expansion and new product launches.

132%
Revenue Growth
Year-over-year
2
New Portals
Launched in 6 months
3 weeks
Market Entry
Down from 18 months

Business Context

The Challenge

Rakuten Advertising operated a major affiliate marketing platform serving publishers and advertisers globally. After acquiring Slice Intelligence, they wanted to expand their affiliate platform capabilities but were constrained by aging on-premise infrastructure.

The business challenge was clear: expand into new geographic markets and launch 2 new affiliate portals without the 18-month infrastructure timeline they'd experienced with previous expansions.

On-premise infrastructure required physical data center presence for geographic expansion, creating a fundamental constraint on go-to-market velocity.

Technical Challenge

On-Premise Bottleneck

  • Legacy monolith in data centers
  • 18-month lead time for new markets
  • Hardware failures causing outages
  • No infrastructure standardization

Business Impact

  • Lost competitive deals
  • Product launches delayed
  • Customer SLAs at risk
  • Operations scaled linearly

Market Constraints

  • Geographic expansion blocked
  • Multi-cloud requirements unmet
  • Enterprise deals lost
  • Platform evolution stalled

Strategic Approach

The Insight

Infrastructure location is a go-to-market decision, not just a technology choice.

To expand internationally and launch new products quickly, we needed infrastructure that could be deployed in new regions in days, not months. Multi-cloud wasn't about avoiding vendor lock-in - it was about matching infrastructure to customer requirements.

Key Decisions:

  • • Google Kubernetes Engine (GKE) as primary platform
  • • Multi-cluster architecture with service mesh
  • • Infrastructure-as-Code for repeatable deployments
  • • Regional isolation with failover capabilities

Implementation

Phase 1: Foundation

Months 1-2
  • • Designed 7-cluster GKE architecture
  • • Implemented Istio service mesh
  • • Built Terraform modules
  • • Created migration runbooks

Phase 2: Migration

Months 3-5
  • • Migrated core platform from on-prem
  • • Moved workloads from EKS to GKE
  • • Zero downtime maintained
  • • Multi-region traffic routing

Phase 3: Expansion

Months 6-8
  • • Launched 2 new affiliate portals
  • • Deployed to 3 new regions
  • • Automated deployment pipelines
  • • Self-service portal for teams

Business Impact

Immediate Results

  • • 2-hour release cycles (from weeks)
  • • Zero ops involvement in deployments
  • • Geographic expansion: 18 months → 3 weeks
  • • Infrastructure standardized across products

Business Outcomes

  • • 2 new affiliate portals launched in 6 months
  • • Entered 3 new international markets
  • • 132% revenue growth year-over-year
  • • 30% ops team reduction while scaling 3x

Strategic Value

  • • Multi-cloud capability as differentiator
  • • Product teams could launch in weeks
  • • Sales could commit to customer requirements
  • • Platform became business enabler