The Challenge
An HR technology company had built its original product on a legacy stack that was becoming an existential threat. Load times of 8–12 seconds were losing enterprise deals to faster competitors, the monolithic architecture made feature delivery painfully slow, and the engineering team was spending 60% of their time fighting production incidents rather than building features.
Our Approach
We ran a two-week technical assessment to understand the architecture, codebase quality, and delivery process. The diagnosis was clear: the monolith needed to be decomposed into services, the frontend needed a full rebuild on React, and the deployment pipeline needed to be automated from scratch.
Rather than a big bang rewrite — which carries enormous execution risk — we used a strangler fig pattern, replacing components incrementally while keeping the product live. Critical user flows were rebuilt first, with the legacy system serving as a fallback until each module was proven in production.
The Results
- Mean page load time reduced from 9.4s to 1.1s (89% improvement)
- Release frequency increased from monthly to daily deployments
- Production incident rate reduced by 84% in the six months post-launch
- £2M enterprise contract closed within three months of rebuild completion — the deal had previously been blocked by performance concerns