A simple definition
FinOps, a blend of "Finance" and "DevOps", is the set of practices used to manage cloud costs. Its principle: every euro spent in the cloud must be visible, tied to a team, and justified by the value it brings. It isn't cost cutting at any price: it is about spending right.
The reference framework is published by the FinOps Foundation, part of the Linux Foundation. It describes the discipline's principles, roles and practices.
The 6 FinOps principles
The FinOps Framework sets out six principles. Here they are, with what they mean in practice:
- Teams need to collaborate: finance, engineering and business work together, in near real time.
- Business value drives technology decisions: the goal isn't the lowest cost, but the best value for every euro.
- Everyone takes ownership of their usage: the team that creates a resource owns its cost.
- FinOps data is accessible, timely and accurate: hence a dashboard refreshed daily, tied to the CMDB.
- FinOps is enabled centrally: a FinOps team provides tooling, negotiates commitments and sets the rules.
- Take advantage of the cloud's variable cost model: what starts on demand can also be shut down.
The framework then organizes the practice into four domains: understand usage and cost, quantify business value, optimize usage and cost, and manage the FinOps practice. We detail what we deliver in each.
Why FinOps has become essential
- The cloud is pay-as-you-go, by the minute and by the GB. The bill changes every month, with no investment to approve.
- Purchasing is decentralized: every team can create resources in a few clicks.
- The bill drifts silently: at 5% a month, it grows by 80% in a year.
- Generative AI adds new cost lines that can run away within hours.
Who is involved?
FinOps is a team effort. Finance wants a reliable budget. IT wants to control its trajectory. Engineering teams control the resources. Leadership arbitrates. In practice, a central FinOps team provides tools and drives the effort, and champions in each business unit relay it to the teams.
The phases of FinOps
| Phase | Goal | Examples |
|---|---|---|
| 1. Inform | Understand your spending | Reading the bill, cost allocation by application and owner, dashboard, forecast |
| 2. Optimize | Find the savings | Usage-based recommendations, reservations, rightsizing, Stop/Start, re-architecture of costly services with application teams |
| 3. Operate | Get them applied and measure them | Remediation plan, SteerCo, savings measured on the bill |
| 4. Stay in control | Keep waste from coming back | Governance, alerting, FinOps by design |
The first three phases are those of the FinOps Foundation framework. We add the fourth because it is what makes savings last: without governance, every new project recreates the waste. See our approach.
The main levers
- Rate optimization, paying less for the same consumption: Reserved Instances and savings plans (about 50% on a prod VM), Azure Hybrid Benefit, negotiated commitments.
- Usage optimization, consuming only what is useful: rightsizing, automatic Stop and Start on demand (about 90% in non-prod), cleaning, storage tiering, log filtering.
- Governance: tagging, showback then chargeback, budget alerts, FinOps by design.
What return on investment?
When nothing is managed yet, about 35% of the cloud bill is recoverable on average. The first savings come in the first month, with reservations, cleaning and storage tiers, which require no application change.
That is why a FinOps engagement pays for itself very quickly: the engagement pays for itself in less than 3 months of savings. For SMEs, we are even paid on results, at 20% of the annual savings achieved. See our offers.
FinOps and GreenOps
GreenOps applies FinOps thinking to the cloud's environmental footprint. The two go hand in hand: a deleted VM costs nothing and emits nothing. With one exception worth knowing: a reservation lowers the bill without reducing physical consumption.
FinOps and artificial intelligence
AI plays two roles. It is a new cost line, to be watched with guardrails (budgets, quotas, alerts). It is also a tool: analyzing billing data, automating and producing recommendations is much faster than before.
Where to start?
- Make costs visible: export billing, tag, build a first dashboard.
- Go after the quick wins: reservations on the baseline, cleaning, Azure Hybrid Benefit.
- Launch a remediation plan: rightsizing, Stop/Start in non-prod, services that drift with volume.
- Set up governance: champions, monthly SteerCo, showback then chargeback.
- Make FinOps by design: standards and building blocks for new projects.