Solution
Cloud cost optimisation
A cloud bill can rise through real growth, pricing, licences, architecture, sizing, ownerless resources or retention never decided. The analysis attributes the cost first, then measures each change against performance and risk.
The gains from optimisation
The bill becomes readable
Every franc attached to a service and a team. That’s the prerequisite for any discussion, and it’s usually what is missing.
Decisions become trackable
Shutdown, retention, capacity and alerts get an owner and a measure after the change. The benefit gets confirmed or corrected, never assumed to persist.
Risk travels with the saving
Every reduction comes with what it risks breaking. A saving that causes an outage costs more than it returns.
In practice
- Full inventory and attribution: who consumes what, on whose behalf, since when
- Orphaned resources identified: detached disks, reserved addresses, old snapshots, forgotten test environments
- Sizing based on measured use rather than on the shape chosen on day one
- Automatic shutdown of non-production environments at night and at weekends
- Retention and storage tiers declared per data type, with archiving of what is no longer read
- A spend alert and a monthly review, so drift shows up within the month rather than the year
Systems involved
- Azure and Google Cloud
- Swiss providers and on-premise infrastructure
- The infrastructure-as-code tool already selected, such as Terraform, to keep changes reviewable
- BI tool for the monthly review
- Internal billing and chargeback systems
Service lineAzure and Google Cloud →
Where the bill goes
The same work, against each sector’s own constraints. Every card opens the full sector.
Energy and utilities
Cost of the measurement platforms
Time-series storage and forecast computation attributed to what they serve, with retention decided rather than inherited.
Logistics and supply chain
Cost of end-to-end visibility
Flow tracking and transport history costed by use, with test environments switched off outside the periods they serve.
How it runs
Attribution
Tagging and attribution until the unattributed share of the bill turns negligible.
Low-dependency decisions
Resources whose owner, use and dependencies have been checked are proposed for removal. Savings count only after the bill actually changes.
Sizing
Resources are brought back to measured use, one at a time, with the associated risk written beside each.
Guardrails
Scheduled shutdown, declared retention, spend alert. Without this step the bill climbs back.
Dependencies and next steps
Test the fit: Cloud cost optimisation
Describe the context, constraints and decision you need to make. The first conversation qualifies scope, boundaries and the next useful step.
Describe the situation