Cloud bills grow quietly: test environments left running, storage nobody deletes, oversized machines picked on day one. Most of that waste is visible in the billing data you already have. These four steps explain the bill, remove what is clearly unused and set a routine so it does not creep back.

Step 1 — Make the bill readable

Goal: See spend by service, environment and owner.

  1. Export the last three months of billing data from each cloud account.
  2. Group spend by service and by environment: production, staging, test.
  3. Tag resources with an owner and environment where tags are missing.
  4. List every account and subscription, including ones on personal cards.

You finish with: A spend breakdown with an owner for each major line.

Step 2 — Find the clear waste

Goal: Spot spend that delivers nothing today.

  1. Find machines and databases with low use over the period.
  2. List unattached disks, old snapshots and storage with no recent access.
  3. Check non-production environments that run around the clock.
  4. Flag services nobody can name a current use for.

You finish with: A waste list with an estimated monthly cost for each item, from your own billing data.

Step 3 — Remove and right-size safely

Goal: Cut the waste without breaking production.

  1. Confirm each item with its owner before changing it.
  2. Snapshot or back up before deleting anything that might be needed.
  3. Schedule non-production environments to stop outside working hours.
  4. Resize oversized resources one at a time and watch performance.

You finish with: Changes made, each with an owner sign-off and a rollback.

Step 4 — Keep it down

Goal: Stop the bill drifting back up.

  1. Set budget alerts per account and environment.
  2. Require an owner and environment tag on new resources.
  3. Review the top cost lines monthly in a short, standing meeting.
  4. Consider commitment discounts only for steady production use, after right-sizing.

You finish with: Alerts, tagging rules and a monthly review with an owner.

Common pitfalls

  • Buying commitment discounts before removing waste.
  • Deleting resources without asking their owner.
  • Cutting production capacity to hit a number.
  • A one-off clean-up with no routine afterwards.

Read as Markdown Fractional CTO