Cloud Migration
GfK
GfK, one of the world’s largest data analytics and consumer intelligence companies, relies on its Performance Pulse product to deliver real-time sales performance data to leading consumer electronics brands and retailers. By partnering with MakeCloud to migrate Performance Pulse from ageing on-premises infrastructure to a modern, fully serverless AWS architecture, GfK dramatically reduced operational overhead, unlocked elastic scalability, and achieved significant cost savings—while ensuring the platform can keep pace with an increasingly fast-moving market.
The Challenge
Performance Pulse had been running on legacy on-premises servers for several years. As demand for the product grew, GfK faced several mounting challenges:
- Scaling limitations — the on-premises infrastructure couldn't elastically scale to handle traffic spikes during peak retail periods such as Black Friday and the Christmas trading season.
- High operational overhead — a dedicated team was required to manage, patch, and maintain the physical servers, diverting engineering time from product development.
- Search performance — the existing search and analytics layer was struggling under increasing data volumes, leading to slow query response times for end users.
- Deployment bottlenecks — releases required coordinated maintenance windows and manual deployment steps, slowing the delivery of new features.
- Cost inefficiency — the infrastructure was provisioned for peak load, meaning GfK was paying for idle capacity during quieter periods.
GfK needed a partner with deep AWS expertise to re-architect the platform without disrupting the service that thousands of users relied on daily.
The Solution
MakeCloud designed and implemented a fully serverless architecture on AWS tailored to Performance Pulse’s workload characteristics:
- AWS Lambda – All application logic was decomposed into Lambda functions, removing the need for server management. Latency-sensitive endpoints used provisioned concurrency to minimise cold-start impact and maintain consistent response times during traffic spikes.
- Amazon API Gateway – A RESTful API layer was built with authentication, rate limiting, request validation, and usage plans. This gave GfK fine-grained control over API consumption, with built-in monitoring and throttling for reliability and governance.
- Amazon OpenSearch Service – The search and analytics backend was migrated to a managed OpenSearch cluster, delivering sub-second query performance across billions of data points. MakeCloud created custom index mappings and query optimisations specifically tuned for Performance Pulse’s aggregation-heavy analytics workload.
- Infrastructure as Code (IaC) – The entire stack was modelled in AWS CloudFormation, enabling consistent, repeatable deployments across environments and providing a complete, auditable history of infrastructure changes.
- CI/CD Pipeline – An automated deployment pipeline was implemented, allowing GfK’s engineering team to release updates multiple times per day with zero downtime.
The migration followed a phased approach. MakeCloud ran the legacy and new serverless systems in parallel, gradually shifting traffic using weighted routing. This allowed the team to validate performance, stability, and functional correctness at each step before fully cutting over to the new architecture.
“Migrating Performance Pulse to a fully serverless architecture on AWS has transformed how we operate. We can now scale effortlessly through peak trading periods, ship features whenever we need to, and focus our engineers on product innovation instead of infrastructure.”
The Results
- Query performance improved dramatically: OpenSearch now returns sub-second responses for complex aggregation queries that previously took several seconds on the legacy infrastructure.
- Automatic scaling: AWS Lambda and API Gateway seamlessly absorbed Black Friday traffic spikes without manual intervention or performance degradation.
- Deployment frequency increased: Release cycles shifted from monthly deployments with maintenance windows to multiple deployments per day.
- Operational overhead eliminated: With no servers to manage, GfK's engineering team can focus entirely on product development and data quality.
- Cost savings: The pay-per-request model of Lambda and API Gateway, combined with right-sized OpenSearch instances, significantly reduced infrastructure costs compared to the always-on on-premises setup.
Services Used
AWS Architecture & Consulting
Design AWS solutions that are scalable, resilient, and cost-effective.
AWS Cloud Migrations
Plan and execute seamless AWS migrations that minimise risk, reduce downtime, and set you up for long-term success.
AWS DevOps
Ship faster on AWS with reliable CI/CD, automation, and modern DevOps practices.
AWS Advanced Tier Services Partner
Ready to get started?
Book a free, no-obligation call with one of our AWS-certified engineers. We'll listen to your challenges, share honest advice, and only recommend next steps if we genuinely think we can help.