An ASX-listed property group was spending $880,000 per year on cloud infrastructure for a platform serving 40,000 users. The board wanted answers.
An ASX-listed property services group had built their tenant and landlord management platform over seven years. By the time we were engaged, they were spending $880,000 per year on AWS infrastructure for a platform serving 40,000 active users — a cost drawing board-level scrutiny.
Our fractional CTO’s infrastructure and architecture review identified three major cost drivers: a microservices architecture implemented without understanding inter-service communication costs; an ElasticSearch cluster consuming $180,000 per year for functionality replaceable with PostgreSQL full-text search; and development environments running 24/7 at production scale.
We designed and oversaw an 8-month re-architecture programme: consolidating 23 microservices into a modular monolith, replacing ElasticSearch with native PostgreSQL, implementing environment lifecycle management, and migrating from EC2 to containerised workloads on ECS Fargate. Every change used feature flags and blue-green deployments for zero-downtime delivery.
Annual infrastructure spend dropped from $880,000 to $500,000 — saving $380,000 per year. Page load times improved from 4.2 seconds to 0.8 seconds on property search. The development team gained three weeks per quarter previously spent on infrastructure maintenance. The CFO presented the results as one of the highest-ROI technology initiatives in the group’s history.
Cloud cost overruns are almost always an architectural problem, not a spending problem. Fix the architecture and the costs follow.
Book a free call with a CTO who has 27 years of hands-on experience — and a track record of measurable outcomes.