Cloud Cost Optimization Checklist (FinOps Lite)

Set one accountable owner for cloud cost

Cloud waste often grows when everyone can deploy but no one owns the bill. Assign one accountable owner for cost governance and define who approves spending changes.

• Assign an executive sponsor, usually CFO, COO, or CEO.
• Assign a technical owner, usually IT Manager, Sysadmin, or MSP lead.
• Assign workload owners for each major application, client environment, or department.
• Define what needs approval: new subscriptions, premium SKUs, region changes, reserved capacity, and AI or analytics workloads.

Baseline the last 90 days of spend

A 90-day baseline gives you enough signal to separate normal usage from drift. Microsoft Cost Management can analyze, monitor, report, and optimize Microsoft Cloud costs across billing accounts, subscriptions, resource groups, and management groups.³

• Export total monthly spend for the last three complete months.
• Break spend down by subscription, resource group, service, region, and meter.
• Flag the top 10 services by cost.
• Flag the top 10 month-over-month increases.
• Document committed spend, pay-as-you-go spend, support costs, and marketplace purchases separately.

Create a usable tagging standard

Tags are key-value metadata used to identify Azure resources by settings that matter to the organization. Microsoft notes that tags can group billing data by categories such as cost center or runtime environment.⁸

Start with a short, mandatory tag set. Long tag programs fail when busy teams cannot follow them.

Key takeaway: If a resource has no owner, department, environment, or application, it should be treated as a cost-governance exception.

Enforce the tag standard with policy

Manual tagging is not enough. Microsoft guidance recommends Azure Policy to enforce tagging rules and avoid resources being deployed without expected tags.⁹

• Require core tags on new resource groups and resources.
• Use remediation tasks to correct existing resources where appropriate.
• Do not store sensitive information in tag values.
• Audit untagged resources weekly until compliance is stable.
• Block or flag high-cost resources without required tags.

Set budgets, forecasts, and anomaly alerts

Microsoft Cost Management supports budget alerts, anomaly alerts, scheduled alerts, and saved cost views.³ A FinOps Lite setup should include alerts that trigger a conversation before a bill becomes painful.

• Create monthly budgets at subscription, resource group, and major workload levels.
• Set thresholds at 50 percent, 75 percent, 90 percent, and 100 percent of budget.
• Create forecast alerts when projected spend will exceed budget.
• Review anomaly alerts for unexpected spikes or dips.
• Send alerts to both technical and financial owners.

Find idle and orphaned resources first

The safest savings usually come from resources that are not doing useful work. Azure Advisor can identify idle and underutilized resources, including virtual machines and virtual machine scale sets, and recommend shutdown or resize actions.⁵

• Stopped but still attached disks.
• Idle virtual machines and scale sets.
• Unused public IP addresses.
• Old snapshots and unattached managed disks.
• Unused load balancers, gateways, and test databases.
• Over-retained logs, diagnostics, and monitoring data.
• Resources in old project, migration, proof-of-concept, or sandbox resource groups.

Right-size before buying commitments

Right-sizing means matching resources to the workload they actually run, not the workload someone guessed months ago. Azure Advisor recommends resizing when a workload can fit on a more appropriate, less expensive SKU without expected performance impact.⁵

• Review CPU, memory, disk, and network utilization together.
• Separate user-facing workloads from batch or background workloads.
• Test right-sizing changes in a maintenance window.
• Record rollback steps before changing production systems.
• Delay commitment purchases until after right-sizing changes settle.

Schedule non-production environments

Pre-production environments do not always need production-level uptime. Microsoft Well-Architected guidance recommends tailoring each environment for its purpose and turning off resources when they are not actively used, such as during evenings or weekends.¹¹

• Stop dev and test VMs outside working hours.
• Use expiry dates for sandbox environments.
• Limit maximum scale settings in dev and test.
• Restrict premium SKUs in non-production.
• Use infrastructure as code for temporary environments that can be recreated when needed.

Review reservations and savings plans after cleanup

Once waste and rightsizing are addressed, evaluate commitment discounts. Microsoft states Azure savings plans can reduce eligible compute costs by up to 65 percent compared with pay-as-you-go pricing, while Azure reservations can reduce eligible resource costs by up to 72 percent, depending on workload, region, term, and usage pattern.^{7, 12}

• Use reservations for predictable, steady workloads.
• Use savings plans when compute usage is steady but instance family, region, or workload mix may change.
• Do not commit to workloads that may be retired, migrated, right-sized, or replaced soon.
• Track utilization monthly so commitments do not become hidden waste.

Clean up storage, backups, and logs carefully

Storage is often ignored because individual line items look small, but data volume, replication, backups, retention, file formats, and access tiers all affect cloud cost. Microsoft Well-Architected guidance recommends optimizing data costs according to data importance and access frequency.¹³

• Move infrequently accessed data to the right storage tier.
• Review snapshot age and retention policies.
• Remove duplicate, expired, or abandoned backup sets only after confirming recovery obligations.
• Set retention rules for logs, diagnostics, and security telemetry.
• Do not reduce retention below legal, insurance, or business continuity requirements.

For Microsoft 365 environments, pair this review with your M365 backup approach so storage savings do not create recovery gaps.

Review licensing and Microsoft subscriptions

Cloud cost optimization includes SaaS and licensing, not just infrastructure. Review inactive users, duplicate licenses, over-assigned premium plans, add-ons, support plans, and unused security or productivity features.

• Remove or reassign licenses for departed users.
• Match license tiers to actual role needs.
• Review Microsoft 365, Dynamics, Azure, security, backup, and third-party marketplace licenses together.
• Align license decisions with access control, data governance, and security requirements.

For a Microsoft 365-specific operating rhythm, use a weekly, monthly, and quarterly administration checklist alongside your cloud cost review.

Put spending guardrails in place

The Azure Well-Architected Framework treats cost optimization as one of five architectural pillars, alongside reliability, security, operational excellence, and performance efficiency.⁴ A cost guardrail should help teams move quickly without letting spend drift out of control.

• Allowed regions for common workloads.
• Allowed VM families and maximum SKUs by environment.
• Approval requirements for premium database tiers, GPU workloads, and AI services.
• Budget limits for sandbox and proof-of-concept work.
• Required backup, monitoring, and security baselines before production go-live.