Definition of FinOps
FinOps, a contraction of finance and DevOps, is a cloud financial management discipline. It organises collaboration between technical teams, finance and the business so that every cloud expense is visible, assigned to an owner and justified by the value it produces.
The definition of FinOps adopted by the FinOps Foundation, the body that maintains its reference framework, stresses its nature: an operational and cultural practice. Born for the cloud, the discipline now extends, according to this definition, to all technology spending. FinOps is neither a tool nor a one-off savings plan.
What FinOps is for
The cloud changes the nature of IT spending. In a data centre, hardware is bought in advance, after a budget has been approved. In the cloud, an engineer commits spending by deploying a resource, without a purchase order. The resulting bill is variable, detailed line by line and hard to link to a team or a product.
FinOps responds to this situation with four contributions.
- Visibility: knowing who consumes what, and why.
- Accountability: entrusting each team with tracking its own consumption.
- Optimisation: eliminating waste and paying the right rate.
- Forecasting: anticipating spending instead of discovering it at the end of the month.
Inform, optimise, operate: the three phases
The FinOps framework describes a cycle that teams go through continuously, scope by scope.
Inform. The first phase makes costs readable. It is based on allocation: each resource is linked to a team, a product or an environment, through an account structure and rigorous tagging. Costs are then reported back to teams, for information (showback) or as internal recharging (chargeback). Budgets, forecasts and unit costs are built on this basis.
Optimise. The second phase acts on two distinct levers. The first is usage: switching off what is not used, sizing resources to the actual load, choosing suitable storage classes, reviewing a costly architecture. The second is rate: committing to a volume of consumption in exchange for a discount, for the stable part of the activity.
Operate. The third phase makes these practices last: governance rules, automation, objectives per team, regular reviews, detection of spending anomalies.
The framework also speaks of maturity, in three figurative levels: crawl, walk, run. It recommends progressing in stages rather than aiming straight away for a complete set-up.
The players in FinOps
- The FinOps team. Often small, it provides the data, the methods and the coordination. It generally centralises the negotiation of rate commitments.
- Engineering and product teams. They decide on the architecture and are therefore accountable for consumption. They are the ones who carry out usage optimisations.
- Finance. It integrates the cloud into the budget, forecasts and cost accounting.
- Procurement. It negotiates contracts with cloud providers.
- Business owners. They make the trade-offs between cost, lead time and quality of service.
- Senior management. It sets the objectives and supports the approach.
The division of roles follows a simple rule: usage is decided as close to the teams as possible, rates are negotiated centrally. An internal developer platform helps apply tagging and sizing rules as soon as resources are created.
FinOps and cost reduction: not to be confused
FinOps does not aim for the lowest bill. It aims for the best value per euro spent. Spending that grows at the same pace as activity can be healthy, whereas stable spending can hide waste. This is why the discipline reasons in terms of unit cost: cost per transaction, per customer, per processing job.
Three other confusions are common.
- With traditional IT cost control, built on annual budgets and hardware investments. FinOps works on variable spending, tracked as it happens.
- With a tool. Cost consoles and specialised software equip the approach. They create neither allocation nor team accountability.
- With digital sobriety. Reducing wasted resources often reduces the environmental footprint, but the two approaches do not have the same objective or the same indicators.
A concrete example: a bill that rises without explanation
Take a company whose cloud spending rises every month without anyone being able to say why. The inform phase reveals that a large share of resources is linked to no team. A mandatory tagging rule, applied at deployment, corrects this.
Once costs are allocated, two findings emerge. Test environments run at night and at weekends without being used, and several databases are sized well above their load. Scheduled shutdown and resizing fall to the teams concerned. The stable part of consumption is then covered by a rate commitment, negotiated centrally.
The operate phase makes it all last: a monthly review per team, an alert in the event of abnormal spending and a unit cost indicator tracked by product owners.
To put this approach in place, see our page dedicated to FinOps and cloud cost control.