Why native tools aren't enough
Azure Cost Management and AWS Cost Explorer are useful, but they remain generic: they don't know your organization, don't tie costs to your applications or their owners, and don't track the savings achieved. As a result, everyone looks at a different number, and nobody feels responsible for their consumption.
What the dashboard contains
- Overview: month-to-date spend, annual forecast versus budget, savings achieved year to date, potential savings that can be reallocated.
- By application: monthly cost, owner from the CMDB, potential savings, CO₂ emissions.
- Recommendations: every action quantified, with its status, all the way to measuring the real saving.
- Reservations: coverage, utilization and expiry dates of Reserved Instances and savings plans.
- GreenOps: emissions by application and business unit, with their trend.
The views are defined with you: leadership sees the overall trajectory, each application owner sees their scope, and everyone reads the same numbers. An interactive mockup is shown further down this page.
Recommendations and alerting
The dashboard includes a recommendation engine more reliable than native tools, and monitoring with alerts on the forecast, anomalies, AI services and commitments. Both are detailed further down, with an interactive mockup.
How it is built
- Daily billing exports (Azure Cost Management, AWS Data Exports or Cost and Usage Report).
- Mapping to applications and owners through your CMDB, whatever the tool (ServiceNow or other).
- Usage metrics (Azure Monitor, CloudWatch) for the recommendation engine.
- Delivered in Power BI, with KQL queries where useful.
The monthly FinOps SteerCo
Every month, the dashboard feeds a short SteerCo with IT leadership, finance and the FinOps champions. The agenda is always the same:
- Monthly consumption, by business unit or region, with every variation explained: savings achieved on one side, new resources on the other.
- Yearly projection against budget, with the average cost per VM and its trend.
- Year-to-date savings, by lever (Stop/Start, rightsizing, cleaning, logs, data, reservations), and the remaining potential.
- FinOps KPIs against their targets, and recommendation tracking.
- A focus of the month: backup growth, AI services consumption, an upcoming migration.
- GreenOps: monthly emissions and carbon intensity.
The KPIs we track
| KPI | What it measures | Indicative target |
|---|---|---|
| Tagging | Share of costs tied to a CMDB application | ≥ 95% |
| RI / savings plans coverage | Share of eligible compute covered by a commitment | ≥ 80% |
| Rightsizing | Share of VMs correctly sized | ≥ 90% |
| Metric coverage | Share of VMs reporting usage metrics: without them, no recommendation | 100% |
| Non-prod Stop/Start | Share of non-prod VMs running less than half the time | ≥ 80% |
| VDI Stop/Start | Same indicator, on virtual desktops | ≥ 80% |
| Average cost per VM | Full monthly cost of a VM, disks included | Decreasing |
| FinOps scoring | FinOps maturity of each business unit, out of 100 | Increasing |
Each KPI reads overall and by business unit: the gap between units shows where to act.
Tracking recommendations all the way to the bill
A recommendation only has value if it is applied. We track each one with a status: done (and measured on the bill), in progress, to execute, awaiting approval, rejected or not assigned. A rejected recommendation is not a failure: its reason is recorded and used to refine the engine's rules. Unassigned or pending amounts are presented at the SteerCo, so the decision goes up to the right level.