Skip to content
Back to glossary
Cloud Cost Optimisation

What is FinOps?

FinOps is a discipline that brings financial accountability and collaboration between engineering, finance and product teams around variable cloud costs.

What FinOps does

FinOps (Cloud Financial Operations) is a discipline that structures cloud cost management as a shared responsibility between engineering, finance and product. Before public cloud, infrastructure costs were predictable: 3- or 5-year capex, linear depreciation, few variations. Cloud flipped that model: costs became variable, granular and instantaneous, and nobody knows exactly what the 50,000 CHF monthly AWS bill actually covers.

FinOps tackles that disorder by formalising three practices: visibility (who spends what, on which project), optimisation (reduce technical waste), and governance (budget framing, alerts, monthly reviews).

The three phases of the FinOps cycle

1. Inform. Tag every resource with its owner, environment, project. Build cost dashboards by team, product, customer. Give engineers visibility into their own spend; without this first step, no optimisation is sustainable.

2. Optimize. Identify oversized resources, orphan volumes, workloads idle 80% of the time, on-demand instances that should have been reserved or spot. The first wave of optimisations typically frees 25 to 40% of spend within a few weeks, with no major architectural change.

3. Operate. Embed cost reviews in the monthly operational cycle. Define per-team budgets with automatic alerts. Make month-end cloud a ritual as structured as month-end accounting.

In practice for a Swiss SME

FinOps is not reserved for companies with multiple millions of dollars of annual cloud spend. On Hidora engagements, a Swiss Romande SME of 100 people with 30,000 CHF monthly infrastructure spend can typically free 8,000 to 12,000 CHF in savings in the first quarter. The most common levers observed in the field:

  • Kubernetes right-sizing: adjust pod requests and limits based on actual consumption, not on pessimistic initial assumptions. Typical savings: 30 to 40% of reserved CPU.
  • Spot instances on tolerant workloads: CI/CD, batch jobs, staging environments. Compute-cost reduction of 60 to 80%.
  • Appropriate storage class: move from premium SSD to cold storage for backups and old logs. Factor 5 to 10 savings.
  • Reservations on stable baseline: cover 50 to 70% of continuous CPU with 1- or 3-year commitments.

FinOps and Swiss sovereignty

Swiss organisations that consolidate their infrastructure on local providers (Hikube, Infomaniak, Exoscale) get one variable less: costs are in CHF, with no foreign-exchange risk. For workloads still on hyperscalers, factoring USD/CHF volatility into the forecast has become a FinOps reflex since 2022.

When FinOps is still premature

For a startup in product-discovery phase or infrastructure below 5,000 CHF monthly, the operational cost of standing up FinOps exceeds the gains. A one-off quarterly audit is preferable to a structured programme at that scale.

Related Hidora services

  • Consulting: initial FinOps audit, identification of savings opportunities, costed action plan.
  • Managed Services: monthly cost review in the operational report, budget alerts.
  • Kubernetes and Observability: technical prerequisites to measure consumption per workload.