FinOps guide: controlling cloud spend without sacrificing performance

Moving to the cloud was supposed to save money. For many South African businesses, the initial promise held - until the monthly invoice started climbing in ways that no one fully understood. Unused resources, oversized instances, forgotten test environments, and untagged workloads accumulate quietly until someone in finance asks the uncomfortable question: “Why is our cloud bill 40 % higher than last quarter?”

FinOps - short for cloud financial operations - is the discipline of bringing financial accountability to cloud spending. It is not about cutting costs at any price. It is about ensuring every rand spent on cloud delivers business value.

Why cloud costs are hard to control

Traditional IT spending is predictable. You buy a server, depreciate it over four years, and the cost is fixed. Cloud spending is variable by design - which is its great advantage and its great risk.

The drivers of cloud cost growth:

Pay-as-you-go pricing means costs scale with consumption, and consumption is not always intentional
Ease of provisioning makes it trivially easy to spin up resources and forget about them
Complex pricing models with on-demand, reserved, spot, and committed-use tiers across hundreds of services
Decentralised spending where multiple teams deploy resources without centralised oversight
Data transfer charges that are often invisible until the invoice arrives
Currency risk - cloud services are priced in USD, and the Rand’s volatility amplifies cost fluctuations

Without deliberate management, cloud costs follow a predictable upward trajectory.

The three pillars of FinOps

A mature FinOps practice rests on three pillars: visibility, optimisation, and governance.

Pillar 1: Visibility - know what you are spending

You cannot optimise what you cannot see. The first step is building a clear, granular picture of your cloud costs.

Tagging strategy

Tags are the foundation of cloud cost visibility. Every resource should be tagged with metadata that identifies:

Cost centre or business unit - who owns the cost
Environment - production, staging, development, testing
Application or service - what the resource supports
Owner - the team or individual responsible

Enforce tagging through cloud policy engines (Azure Policy, AWS Service Control Policies, GCP Organisation Policies). Untagged resources should be flagged automatically and treated as non-compliant.

Cost allocation

Map cloud costs to the business units, projects, and applications that consume them. This transforms a single opaque invoice into an intelligible breakdown that business leaders can act on.

Use native tools - Azure Cost Management, AWS Cost Explorer, GCP Billing Reports - or third-party platforms like CloudHealth, Apptio, or Spot by NetApp for cross-cloud visibility.

Showback and chargeback

Showback means reporting costs to the teams that incur them without charging them directly. It builds awareness. Chargeback means actually billing internal teams for their consumption. Both create accountability.

Start with showback. Once the data is trusted and the culture is ready, move to chargeback for teams with significant spend.

Anomaly detection

Configure alerts for unusual spending patterns - a development workload that suddenly costs three times more than last month, or a new resource type appearing in your account. Catching anomalies early prevents small issues from becoming large invoices.

Pillar 2: Optimisation - spend less for the same (or better) outcomes

Visibility tells you where the money goes. Optimisation reduces it without degrading performance or availability.

Rightsizing

The most common source of cloud waste is oversized resources. A virtual machine provisioned with 16 vCPUs and 64 GB RAM that consistently uses 2 vCPUs and 8 GB is wasting 75 % of its cost.

Use provider-native advisors (Azure Advisor, AWS Compute Optimizer, GCP Recommender) to identify rightsizing opportunities. Review recommendations monthly and action them through a controlled change process.

Reserved instances and savings plans

If you have predictable, steady-state workloads - and most businesses do for a significant portion of their estate - committing to one- or three-year reserved instances (or the equivalent savings plans) delivers 30–60 % cost reductions over on-demand pricing.

Guidance:

Analyse at least three months of usage data before committing
Start with one-year terms to limit risk
Reserve at the instance family level (flexible reservations) where available
Review reservation coverage and utilisation monthly

Spot and preemptible instances

For fault-tolerant workloads - batch processing, data analytics, CI/CD pipelines, rendering - spot instances (AWS) or spot VMs (Azure) offer 60–90 % discounts over on-demand pricing. The trade-off is that the provider can reclaim these instances with short notice.

Architect workloads to handle interruptions gracefully - use queues, checkpointing, and auto-scaling groups that mix spot and on-demand capacity.

Storage optimisation

Storage costs accumulate silently. Implement lifecycle policies that automatically transition data to cheaper storage tiers as it ages:

Hot storage for frequently accessed data
Cool/infrequent access for data accessed less than once a month
Archive/glacier for long-term retention where retrieval time is not critical

Delete orphaned snapshots, unattached disks, and old backups that no longer serve a purpose.

Idle and unused resource cleanup

Schedule automated scripts or use cloud-native tools to identify and remove:

Unattached public IP addresses
Idle load balancers
Stopped (but still billing) virtual machines
Empty resource groups and unused storage accounts
Development and test environments running outside business hours

A well-configured cloud architecture includes auto-shutdown schedules for non-production environments - a simple change that can reduce development cloud costs by 60 % or more.

Data transfer optimisation

Cross-region and cross-cloud data transfer is expensive. Minimise it by:

Placing compute resources in the same region as the data they process
Using CDNs to cache content close to users
Compressing data before transfer where practical
Reviewing architecture to reduce unnecessary east-west traffic

Pillar 3: Governance - keep spend under control as you scale

Optimisation is an ongoing activity, not a one-time project. Governance ensures the savings you achieve today are not eroded tomorrow.

Budgets and alerts

Set monthly budgets at the subscription, resource group, or project level. Configure alerts at 50 %, 75 %, 90 %, and 100 % of budget. Route alerts to the team responsible, not just to IT.

Spending policies

Enforce guardrails that prevent wasteful spending before it occurs:

Restrict the VM sizes and types available in non-production environments
Require approval for high-cost resource types (GPU instances, large databases)
Block deployment of resources without mandatory tags
Limit the regions where resources can be deployed

FinOps reviews

Conduct monthly FinOps reviews with stakeholders from engineering, finance, and business units. The agenda should cover:

Actual spend vs. budget for each cost centre
Top optimisation recommendations and their status
Anomalies and their root causes
Forecast for the coming month
Reservation coverage and utilisation

This meeting is the heartbeat of your FinOps practice. Without it, optimisation becomes sporadic and costs drift upward.

The cultural dimension

FinOps is as much a cultural change as a technical one. Engineers need to think about cost as a design constraint, not just an operational afterthought. Finance needs to understand that cloud costs are variable by nature and that some variability is healthy.

Practical steps:

Give engineering teams visibility into the cost of their deployments - dashboards, weekly cost summaries, even cost annotations in pull requests
Celebrate cost savings as publicly as you celebrate feature launches
Include cloud cost efficiency in team KPIs and performance reviews
Ensure business technology consultants and architects consider cost implications during design reviews

FinOps works when everyone understands that cloud spending is a shared responsibility, not just finance’s problem or IT’s problem.

Getting started: a 90-day plan

Days 1–30: Visibility

Implement a tagging strategy and enforce it through policy
Set up cost allocation across business units and projects
Configure budget alerts at the subscription level
Deploy a cost dashboard accessible to all stakeholders

Days 31–60: Quick wins

Identify and terminate idle and unused resources
Implement auto-shutdown for non-production environments
Right-size the top 20 most expensive compute instances
Review storage tiers and implement lifecycle policies

Days 61–90: Sustained practices

Analyse reservation opportunities and make initial commitments
Establish a monthly FinOps review cadence
Begin showback reporting to business unit leaders
Document and share governance policies

This 90-day plan typically delivers 15–30 % cost reduction while improving visibility and accountability.

Managing the Rand-Dollar risk

South African businesses face an additional challenge: cloud services are invoiced in USD (or converted from USD at the provider’s exchange rate). A 10 % weakening of the Rand translates directly into a 10 % cost increase with no change in consumption.

Strategies to manage this:

Use reserved instances denominated in Rand where available (some Azure Enterprise Agreements support local currency billing)
Budget in USD and reconcile to Rand monthly
Build a currency buffer into your cloud budget (typically 5–10 %)
Consider a forward contract with your bank for large, predictable cloud expenditures

Build your FinOps capability

Cloud cost management is not a project with an end date. It is an ongoing discipline that matures over time. Start with visibility, capture the quick wins, and build governance practices that sustain the results.

Your infrastructure and cloud teams are natural owners of this discipline, but success requires buy-in from finance and business leadership.

Get in touch to discuss how we can help you build a FinOps practice that keeps your cloud spend aligned with your business outcomes.