The first GCP cost optimization sprint takes two weeks and runs in a fixed order: remove pure waste, then rightsize, then tier storage, then size committed use discounts last. GCP differs from AWS and Azure in one helpful way, sustained use discounts apply automatically the longer a VM runs in a month, so you start with a discount you did not have to negotiate. Committed use discounts, the CUDs, layer deeper savings on top, but only against a forecast you can defend after the waste is gone. Done in order, a first sprint contributes to the 31 percent median reduction we see in the first 90 days, with almost no risk to production.
This is a GCP cost fundamentals exercise. You need only the billing export and the native Recommender to start. Here is the plan.
Before the sprint: turn on the billing export
Every measurement in the sprint reconciles against the BigQuery billing export, GCP's granular source of truth. The console is fine for a glance, but the export is what you trust for allocation and for the before and after. Enable it, confirm labels are flowing, and pull a clean baseline of the last full month by project, by service, and by label. Aligning the export to the FOCUS specification makes joining it with other clouds easier later.
Switch on the native Recommender too. It surfaces rightsizing, idle resource, and CUD recommendations, and like every native advisor it recommends but does not decide.
Week one: cut the waste that carries no risk
The first week removes spend you can delete without an architecture conversation.
- Orphaned storage. Unattached persistent disks, old snapshots no policy will restore, and Cloud Storage buckets nobody owns.
- Idle resources. Stopped VMs still paying for attached disks, reserved static IPs sitting unused, and idle load balancers.
- Unlabelled and unknown spend. Anything the billing export cannot attribute to a team. Label it or flag it for an owner.
- Non production left running. Dev and test projects running nights and weekends. Scheduling them off outside working hours is one of the highest return moves in the sprint.
This is where the visible number moves first and none of it risks a live service.
Week two: rightsize, tier, then commit
The second week works through changes that need a human to confirm against real behaviour.
- Rightsizing. Recommender flags oversized VMs; you confirm each against actual CPU and memory before resizing. Custom machine types let you size precisely instead of jumping between fixed shapes.
- Storage tiering. Move infrequently accessed Cloud Storage data to Nearline, Coldline, or Archive with lifecycle rules, matching the class to the access pattern.
- Committed use discounts, last. With waste gone and VMs right sized, the steady state floor is visible. Size CUDs to that floor on a defensible forecast. Choose between spend based CUDs, which cover a dollar commitment flexibly, and resource based CUDs, which lock specific resources for a deeper rate. Sustained use discounts keep applying automatically underneath.
A European SaaS company began its first sprint with a year of unreviewed growth. Week one cleared orphaned disks, idle IPs, and a non production estate running around the clock. Week two confirmed rightsizing to custom machine types, tiered cold Cloud Storage down, and only then placed spend based CUDs sized to the reduced steady state, with sustained use discounts already applying. The estate finished materially lighter. Figures are verified against billing data and anonymised.
How to make the savings stick
A sprint without a cadence leaks back. The day it closes, stand up the rhythm that holds the gains: a weekly anomaly and CUD utilization check and a monthly cost by project review. The full rhythm is in the GCP spend review cadence, and the recurring line items it catches are in common GCP billing surprises. The whole GCP picture lives in the GCP cost optimization guide.
Frequently asked questions
How long does a first GCP cost optimization sprint take?
What is the difference between sustained use and committed use discounts?
Should you buy committed use discounts in the first sprint?
Run your first sprint with us
We run the two week sprint with your team, then hand you the cadence that keeps the savings. Our guarantee: we reduce your cloud spend or we reimburse our service fee. Pricing is either a Fixed Fee scoped up front or Gainshare, a share of verified savings with no retainer and no risk to you.
Put a defensible number on your cloud spend.
No provider in the room, no published price list. Tell us your footprint and we will scope the savings against your billing data — we reduce your cloud spend or we reimburse our service fee.
The Cloud Spend Navigator: what changed in cloud pricing, commitments, and FinOps — no vendor spin.