Cut cloud spend without slowing your engineers
Cloud cost programmes often become approval bureaucracy. Visibility, ownership and a few structural changes deliver most of the savings without the friction.
When the cloud bill grows faster than revenue, the reflex is to add approvals. Engineers wait for sign-off to spin up environments, velocity drops, and the bill comes down a little. There's a better trade.
Start with visibility and ownership
Most waste is invisible to the people who create it. Tag every resource by team and service, then put cost per service in front of each team weekly, next to their latency and error dashboards. Ownership changes behaviour faster than policy.
The structural wins
- Right-size and autoscale: most production fleets run at 15–25% utilisation.
- Schedule non-production: dev and staging don't need to run at 3am.
- Commit to the baseline: savings plans for the steady load, on-demand or spot for the peaks.
- Fix data transfer: cross-zone and egress charges hide in architecture decisions.
- Cache aggressively at the edge: the cheapest request is the one that never reaches origin.
The takeaway
The biggest savings usually come from autoscaling, scheduling and commitments, and none of them require taking a single permission away from engineers.
- FinOps
- AWS
- Cost optimisation
