A fintech ran a $6.0M-a-year estate split across Google Cloud and Oracle Cloud, on pay-as-you-go rates and launch-day sizing. We took it to $3.9M, a 35 percent reduction, with committed use discounts, storage tiering and OCI shape rightsizing, on a clean baseline.
A fast-growing fintech ran its transaction platform and analytics on Google Cloud, with a regulated data and database tier on Oracle Cloud, spending about $6.0M a year across the two. Growth had outpaced cost discipline: GCP instances sized for launch-day peaks, no committed use discounts despite a large steady core, terabytes of analytics data sitting in standard storage long after it went cold, and OCI flexible shapes provisioned with far more OCPUs and memory than the workloads used. There was no single view of the combined bill, and finance forecasted each cloud separately and trusted neither.
$6.0M annual multicloud spend · GCP and OCI managed in isolation · no committed use discounts · cold analytics data on standard storage · oversized OCI flexible shapes · no unified forecast.
We worked the engagement in our standard order, See, Cut, Lock, Run, and treated the two clouds as one estate so commitments and storage decisions were made against the whole picture rather than each cloud in isolation.
We brought GCP and OCI spend into a single FOCUS-normalized view and fixed allocation, GCP labels and OCI compartments and tags, so every dollar mapped to a product line and an owner. For the first time the fintech could see its true combined cloud cost and where it concentrated, which is the foundation our OCI cost assessment method is built on.
On Google Cloud we moved cold analytics data out of standard storage into Nearline and Coldline tiers matched to access frequency, cutting the storage line substantially with no impact on the rare reads. On Oracle Cloud we right-sized the flexible shapes down to real OCPU and memory utilization, following the method in how to rightsize OCI compute shapes. Together with idle cleanup, this removed roughly half of the total saving before any rate change.
Only after the GCP fleet was right-sized did we buy committed use discounts on the steady core, leaving the variable layer on-demand. On the OCI side we sized the Universal Credits commitment to proven steady-state usage rather than peak, using the approach in how to negotiate an Oracle Cloud commitment. Buying the rate last, on a smaller true footprint, is what made the discount land on real usage instead of waste.
We set budgets and anomaly alerts on both clouds and built a single rolling forecast across the estate, so the savings could not drift back and finance finally had one number it could defend. Tagging and compartment policy now block untagged deployments on either cloud.
Annual multicloud spend fell from $6.0M to $3.9M, a 35 percent reduction, recovering about $2.1M a year. Roughly half came from storage tiering and OCI shape rightsizing, and half from committed use discounts and a right-sized OCI commitment applied on the clean baseline. Finance gained one forecast across both clouds.
The multicloud angle was the whole game. Managed separately, each cloud looked roughly fine; managed as one estate, the duplicated waste and the missing commitments were obvious. Had the fintech bought committed use discounts before right-sizing, as many teams do, the discount would have locked onto oversized GCP instances and over-provisioned OCI shapes, and most of the 35 percent would never have materialized. Right-size and tier first, then commit, is what made the saving both large and durable.
We will map your spend across every cloud, find the waste, model the right commitment and storage mix, and tell you the number. On the performance model, you pay only from realized savings. No savings, no fee.
Book a cloud cost audit →Figures reflect a real engagement outcome. Client identity withheld for confidentiality.