Case Study · 02
Cloud Cost Optimization
Annual cloud spend reduced from ~IDR 5B to ~IDR 4B — roughly IDR 1B saved every year.
Spend was growing as a fixed cost, not as a function of value.
Cloud spend had become a background assumption — workloads provisioned generously, reservations unexamined, and the monthly bill treated as weather. Cost was growing roughly with the platform, and nobody could say which part of it produced business value.
Impact · FinOps
~IDR 1B saved every year.
Annual cloud expenditure reduced from approximately IDR 5 billion to IDR 4 billion — a ~20% reduction sustained through systematic FinOps, not one-off cuts.
Context matters: spend reduction came from measured usage, workload placement, commitment planning, and commercial review — not from starving environments of capacity.
- Rightsizing
- Workload consolidation
- Kubernetes clusterization
- Spot VMs
- Committed Use Discounts
- Resource segmentation
- Billing optimization
- Partner negotiation
FinOps as engineering work, not an exercise in spreadsheet discipline.
The program treated cloud cost as an engineering property of the platform: measured continuously, owned by workload teams, and optimized at the architecture level — not by cutting corners on capacity.
Levers applied
- Rightsized compute and memory across workloads against real utilization.
- Consolidated workloads and clusters to remove idle and duplicate capacity.
- Kubernetes clusterization — container density replaced per-VM provisioning.
- Adopted Spot VMs for fault-tolerant, interruptible workloads.
- Structured Committed Use Discounts around stable, predictable usage.
- Segmented resources and billing so cost had an owner per domain.
- Automated billing analysis instead of quarterly archaeology.
- Negotiated commercial terms with the cloud partner.
~IDR 1B a year, returned to the business.
The result was a sustained reduction of roughly 20% — from approximately IDR 5 billion to IDR 4 billion in annual cloud expenditure. Context matters: this was achieved while the platform continued to carry the same products, users, and transaction volumes.
Outcome
- ~IDR 1B in annual savings, sustained year over year.
- Cost visibility with an owner per product and environment.
- Optimization levers that compound: density, commitments, and automation.
- No capacity starvation — the platform kept its reliability posture.