For many organisations, the cloud began as a way to avoid big capital outlays and lengthy procurement processes. Spin up what you need, when you need it, and only pay for what you use. Somewhere along the way, the bill started to hurt.
By 2025, a surprising number of SMEs and mid-market firms are spending more on cloud than they ever did on traditional hosting – and they’re not always sure why. FinOps, short for “cloud financial operations”, is the response: a discipline that blends finance, engineering and product to ensure cloud spend is understood, optimised and aligned with value.
If you’ve ever been shocked by a monthly invoice from AWS, Azure or GCP, FinOps is for you.
The first step is visibility. Many cloud accounts begin as an experiment under a single credit card and grow organically. Different teams use different subscriptions or projects. Tags are inconsistent or non-existent. The result is a bill that looks like an alphabet soup of services, regions and mysterious line items.
You can’t manage what you can’t see. Before 2026, make it a priority to consolidate accounts where appropriate, enforce tagging standards (e.g. environment, application, owner, cost centre), and get basic reporting in place. Most major providers offer native cost explorer tools; third-party platforms add more analysis.
Next, identify the “low-hanging fruit” of optimisation.
Common issues include:
- Overprovisioned instances: Servers sized for peak loads that rarely occur.
- Idle resources: Test environments left running out of hours, unused load balancers, orphaned volumes.
- Inefficient storage tiers: Frequently accessed data sitting on expensive, long-term storage, or vice versa.
- Lack of commitment discounts: On-demand pricing used everywhere instead of reserved or savings plans for predictable workloads.
You don’t need a six-month project to tackle these. A focused sprint with engineering and finance can often uncover 10–30% savings simply by rightsizing, shutting down, and switching pricing models.
The important shift is cultural: cloud costs are no longer “just an ops thing”. Product owners should understand the unit costs of their features. Engineers should see cost as a dimension of quality. Finance should treat cloud as a dynamic, controllable spend, not a fixed overhead.
FinOps introduces lightweight processes to support this. For example:
- Regular cost reviews at squad or product level, with clear dashboards.
- Design reviews that consider cost implications (e.g. polling vs event-driven architectures).
- Budget guardrails for experimentation, so teams can explore without creating runaway spend.
It’s also about aligning cost with value. A high cloud bill is not automatically bad if it supports high-margin revenue or critical capabilities. Conversely, a cheap setup that slows teams down or causes outages is a false economy. The goal is not minimum spend; it’s sustainable, intentional spend.
For SMEs without huge teams, don’t be intimidated by FinOps language. You don’t need a FinOps department. You need someone with access to cost data, a basic understanding of cloud services, and the authority to ask “why” when they see strange patterns.
Practical actions before 2026:
- Nominate an owner. Give someone explicit responsibility for cloud cost visibility and coordination.
- Set a tagging policy. Even a simple one will massively improve reporting. Enforce it in CI/CD pipelines if you can.
- Run a one-off audit. Use native tools or a consultant to identify quick wins. Implement them, document savings, and share the story internally.
- Agree thresholds. For example, any new service expected to cost more than £X per month requires a short cost review.
Cloud providers will continue to add services, features and pricing options. Without FinOps discipline, your bill will drift upwards by default. With it, the cloud can go back to what it was supposed to be: a flexible utility, not a mystery tax.
