TL
The short answer

The right Azure commitment coverage target is your steady state usage floor, the capacity that runs every hour of every month, not a round number picked from a slide. Cover that floor with reservations and the Azure Savings Plan, leave the variable layer on demand, and your effective rate falls without stranding spend on capacity you stop using. Coverage strategy is risk adjusted, not discount maximised.

This sits in the Azure compute and reservations cluster. The mistake to avoid is chasing a high coverage percentage as if it were the goal; the goal is the lowest risk adjusted cost, and that usually means covering less than all of your usage.

Why a coverage percentage is the wrong starting point

A target like 80 percent coverage sounds disciplined but it is arbitrary unless it maps to your usage shape. Two estates with the same spend can have very different floors: one runs a stable baseline every hour, the other spikes hard during business hours and idles overnight. Committing the same share of both strands money on the spiky one. Start from the shape, not the percentage.

Find the steady state floor

The floor is the level of usage that is present continuously across the period, reconciled against your Azure cost exports. Plot hourly compute usage over a representative few months and look for the line everything sits above. That floor is the spend that is safe to commit, because it will be there for the full term regardless of how the variable layer moves.

Everything above the floor is variable, and variable usage belongs on demand or on the more flexible Azure Savings Plan, not on a deep reservation you might not fully use.

Blend reservations and the Azure Savings Plan

The two instruments trade discount for flexibility, and a sensible book uses both.

LayerInstrumentWhy
Stable, specific usage at the floorReservationsDeepest discount on capacity you know will run
Steady but shifting across families or regionsAzure Savings PlanTrades a little discount for flexibility as the mix changes
Variable, spiky, or uncertainOn demandNo commitment risk on capacity that may not recur
Worked example

A European SaaS company had committed to a flat coverage target across all compute and was carrying underused reservations on a workload that idled overnight. We rebuilt coverage from the hourly floor, moved the shifting layer to the Azure Savings Plan, and left the overnight variance on demand. Effective rate improved and the underused commitment disappeared. Figures are verified against billing data and anonymised.

Manage the downside

If you over commit, you pay for capacity you do not use, which can erase the discount entirely. Reservations can be exchanged or, within limits, refunded, but the cleaner answer is to size coverage to the floor so over commitment does not happen. Review coverage on a quarterly cadence against an updated forecast, and ladder terms so no single expiry forces a rushed decision.

Frequently asked questions

What is a good reservation coverage target on Azure?
There is no universal number. The right target is your steady state usage floor, the capacity that runs continuously, expressed as a share of total compute. For stable estates that is often high; for spiky ones it is lower. Size to the floor, not to a benchmark percentage.
Should you use reservations or the Azure Savings Plan?
Reservations give the deepest discount on specific, stable usage; the Azure Savings Plan trades a little discount for flexibility across instance families and regions. Cover the predictable core with reservations and the shifting layer with the savings plan.
What happens if you over commit on Azure?
You pay for capacity you do not use, which can wipe out the discount entirely. Reservations can be exchanged or, within limits, refunded, but the cleaner answer is to size coverage to the floor so over commitment does not happen in the first place.

Size coverage with us

We help enterprises size Azure reservation and Azure Savings Plan coverage to a defensible floor, so the discount is real and nothing is stranded. It sits inside the strategy laid out in the Azure cost optimization guide. Our guarantee: we reduce your cloud spend or we reimburse our service fee.

Independent · buyer-side

Put a defensible number on your cloud spend.

No provider in the room, no published price list. Tell us your footprint and we will scope the savings against your billing data — we reduce your cloud spend or we reimburse our service fee.

Buyer-side intelligence, monthly.

The Cloud Spend Navigator: what changed in cloud pricing, commitments, and FinOps — no vendor spin.