Free tool
Cloud waste calculator
Idle Azure VMs, forgotten disks, dev boxes running 24/7, oversized Fabric capacity — most UK SMEs pay for cloud resources they barely use. List what you pay for, mark how much of the time it is genuinely needed, and see exactly how much you could claw back. Everything runs in your browser; nothing is uploaded.
Where the money is going
Add a resource to see the maths.
Working: wasted per month = monthly cost × (1 − % of time genuinely needed). Presets convert weekly hours to a percentage of the 168-hour week (e.g. 8h × 5 days = 40 ÷ 168 = 23.8%). Estimates only — check your actual billing before acting.
Quick answer
Cloud waste is the gap between what you pay for and what you use. Wasted monthly cost = monthly cost × (1 − percentage of time the resource is genuinely needed). A dev VM costing £85/month that is only needed 8 hours a day on weekdays is needed 40 of 168 hours (23.8%), so £64.76 of every month is waste. Unused resources score 0%, so 100% of their cost is waste.
How the maths works
- Weekly hours to percentage: there are 168 hours in a week. Needed 8h × 5 days = 40 ÷ 168 = 23.8% utilisation.
- Waste per resource: monthly cost × (1 − utilisation). An £85 dev VM at 23.8% wastes £85 × 0.762 = £64.76 a month, £777 a year.
- Orphans are the easy wins: unattached disks, idle IP addresses and old snapshots are needed 0% of the time, so every penny is waste — delete or snapshot them.
- Annual figure: monthly waste × 12. Small leaks look harmless monthly and ugly annually.
What counts as cloud waste?
Anything you pay for but do not use enough: dev and test environments running nights and weekends, VMs nobody has logged into for months, unattached managed disks, storage accounts kept "just in case", oversized databases or capacities provisioned for a peak day that never comes, and idle integration or gateway resources. If a resource is needed 24/7 for production it is not waste — the point is matching spend to actual need.
How do I find idle Azure resources?
In the Azure portal, start with Cost Management + Billing to see spend by resource, then check Advisor recommendations (it flags idle VMs and unattached disks), look at the last-used dates on storage accounts, and review utilisation graphs on VMs and databases. Anything with a flat-line CPU graph or a "not used in 90 days" flag goes on the list above before you decide to delete, schedule or downsize it.
Does auto-shutdown actually save money?
Yes for compute. A stopped (deallocated) VM stops accruing compute charges — you only pay for its disks and any attached IPs. Auto-shutdown on a dev box that is needed 8h × 5d cuts its compute cost by roughly three quarters. It does not help for things that bill for storage or provisioned capacity regardless of state, which is why orphaned disks and old blobs should be deleted, not shut down.
Is this only for Azure?
No. The maths is the same for AWS, Google Cloud, Fabric capacities, Power Platform premium connectors, hosted databases or any monthly-billed resource where usage has a time dimension. Enter what you pay and how much of the time it is genuinely needed; the calculator does not care whose logo is on the invoice.
Related tools
Digital Adaption helps UK SMEs get their data estate under control: right-sized Azure and Fabric setups, automated start/stop schedules and reporting that shows what everything really costs. Get in touch. Want this calculator wired to your actual billing data with your own thresholds and alerts? We build custom cost dashboards and internal tools. Contact us for a tailored quote.