Cloud computing promised to reduce infrastructure costs. For many enterprises, it has done the opposite. Without financial governance, cloud spend grows unchecked — driven by over-provisioned resources, forgotten test environments, unoptimised storage, and a fundamental lack of visibility into who is spending what and why.
FinOps — Financial Operations for the cloud — is the discipline that bridges the gap between engineering velocity and financial accountability. At ENGIE Africa, implementing FinOps practices across our AWS, Azure, and GCP environments reduced our monthly cloud spend by 28% within six months of go-live, while simultaneously improving team autonomy and deployment frequency.
Why Cloud Bills Keep Surprising Finance Teams
The root cause is almost always the same: cloud decisions are made by engineers who optimise for performance and delivery speed, while the financial consequences land on a finance team that had no visibility into those decisions until the invoice arrives. In traditional IT, procurement was a gate that created financial accountability. In cloud, provisioning takes seconds and requires no approval.
Three specific patterns account for the majority of cloud waste:
- Idle and over-provisioned resources — instances running at 5% utilisation, right-sized for peak load that never materialises.
- Untagged resources — spend that cannot be attributed to a team, project, or cost centre, making accountability impossible.
- On-demand pricing for stable workloads — paying on-demand rates for workloads that run 24/7 and could be covered by reserved instances at 40–60% discount.
The FinOps Maturity Model
The FinOps Foundation defines three maturity levels. Most large enterprises I work with are somewhere between Crawl and Walk:
Crawl
Visibility
Basic cost dashboards. Some tagging. Finance sees the bill. Engineering doesn't.
Walk — Most enterprises
Accountability
Cost allocated by team. Budgets set. Regular reviews. Reserved instance coverage growing.
Run
Optimisation
Real-time cost signals in CI/CD. Automated rightsizing. Unit economics tracked per product.
Five FinOps Practices That Deliver Immediate ROI
1. Enforce Mandatory Tagging Policies
No resource should exist without tags that identify the owning team, environment (prod/staging/dev), project, and cost centre. Implement tag policies at the account or management group level that prevent untagged resources from being created. Every percentage point improvement in tag coverage improves cost attribution and accountability proportionally.
2. Reserved Instances and Savings Plans
For workloads with predictable, stable usage — production databases, core application servers, baseline compute — reserved instances and savings plans offer 40–65% discounts versus on-demand pricing. The analysis is straightforward: pull 90 days of utilisation data, identify stable workloads, calculate the break-even period (typically 7–9 months for 1-year reservations), and commit. This is the single highest-return FinOps action available to most enterprises.
At ENGIE Africa, converting 60% of our stable workloads from on-demand to reserved instances alone generated €380K in annualised savings.
3. Automated Shutdown of Non-Production Environments
Development, staging, and test environments typically run 24/7 despite being used for 8–10 hours per day. Automated shutdown schedules — stop at 8pm, start at 7am, off all weekend — reduce non-production compute costs by 65–70% with zero impact on developer productivity when implemented with appropriate self-service restart mechanisms.
4. Rightsizing Reviews — Quarterly
Cloud instances are typically provisioned for anticipated peak load, which often never materialises. Quarterly rightsizing reviews — using native tools (AWS Compute Optimizer, Azure Advisor, GCP Recommender) or third-party platforms — identify instances running well below capacity. A consistent rightsizing programme typically identifies 15–25% cost reduction opportunities across the compute estate.
5. Unit Economics and Cost per Business Metric
The most mature FinOps organisations move beyond total cloud spend to unit economics: cost per transaction, cost per active user, cost per GWh of energy managed (in our case at ENGIE). This reframes the conversation from "is our cloud bill too high?" to "is our cloud investment delivering proportional business value?" — a fundamentally different and more productive question.
Multi-Cloud FinOps: The Added Complexity
Managing FinOps across AWS, Azure, and GCP simultaneously adds significant complexity. Each platform has different pricing models, different discount mechanisms, and different native tooling. Organisations at this scale typically need a unified FinOps platform — Apptio Cloudability, CloudHealth, or similar — that provides a single pane of glass across providers. The investment is justified once cloud spend exceeds approximately €500K per month across multiple providers.
Typical Savings by Organisation Type
Where to Start
If your organisation has not yet started a FinOps programme, begin with two actions: enable cost allocation tags and establish a monthly cloud cost review meeting that includes both engineering leads and finance. Visibility and shared accountability are the foundation on which all other FinOps practices are built. From there, the optimisation opportunities become self-evident.