1. Environments that never sleep
A week has 168 hours. A team working 8 hours a day, 5 days a week, uses 40 of them. A DEV or STAGE environment that runs all the time is paid for the other 128 hours too. Scale non-production services to zero when idle, stop development machines at night and at weekends, and keep only what really needs to be always on.
2. Oversized machines and databases
Resources are often sized for a launch day that has passed, or for a guess made under pressure. Compare what each machine and database is given with what it actually uses over a few weeks. Your cloud provider's sizing recommendations are a good starting list; change one thing at a time and watch the metrics.
3. Things nobody deleted
Deleting a virtual machine does not always delete its disks, snapshots or reserved IP addresses, and an unattached static IP is still billed. Search regularly for disks not attached to anything, old snapshots and images, idle load balancers and reserved addresses in no use. Labels with an owner make it much easier to ask before deleting.
4. Logs and storage that never expire
Debug logs at full volume and buckets without lifecycle rules grow quietly every day. Set retention periods that match what you really need, lower the log level outside of investigations, and move old objects to cheaper storage classes or delete them automatically.
5. Data that travels too far
Moving data between regions or out to the internet is billed, and it adds up with chatty services, cross-region backups and large downloads. Keep services that talk a lot in the same region, serve public files through a cache, and check the network lines of the bill before assuming compute is the problem.
6. Paying full price for the steady part
Once your usage is right-sized, part of it is stable month after month. That baseline is a good candidate for committed use discounts or savings plans, which trade a commitment for a lower price. Commit to what you are sure of, not to your peak, and review it as you grow.
Make it a habit
Cleanups help once; habits help every month. Label every resource with a team and an environment, set budgets with alerts per project, and hold a 30-minute monthly review where each team looks at its own spend. Costs stay visible, and nobody gets surprised by the bill.
For the cleanup itself on Google Cloud, reCloud finds the unused VMs, disks, databases, IPs and snapshots in one read-only scan, so the monthly review starts from a list instead of a hunt.
If you want help building that habit, explore our FinOps Consulting for Cloud Cost Optimisation: we set up the dashboards, the savings plan and the routine with you.