Effective Coverage %

Effective Coverage % is CloudQuell’s commitment-efficiency metric: a single number that combines how much of your eligible compute runs on commitments (coverage) with how much of those commitments you actually use (utilization). It’s the one number a CFO and a platform team can both hold the org to — because a good-looking half can’t hide a leaking one.

It’s live on the Commitments view, computed from your Cost and Usage Report (CUR).

Two different “coverages”

The coverage inside this formula is commitment coverage — the share of eligible compute on a commitment. It is not the same as Attribution Coverage, which measures how much of your spend carries the tags you rely on. Both use the word “coverage”; they measure different things.

The formula

Effective Coverage % = Coverage × Utilization
  Coverage    = $ eligible compute on commitments / $ eligible compute total
  Utilization = $ commitment used / $ commitment purchased

• Coverage — the share of your eligible compute spend that runs on a Reserved Instance or Savings Plan instead of on-demand. Measured at on-demand-equivalent (list) rates on both sides, so committed and on-demand spend compare like for like — the same basis AWS uses for its own coverage reports.
• Utilization — of the commitments you bought, how much you actually consumed. Computed per commitment as used ÷ purchased capacity, then dollar-weighted across your portfolio so one large under-utilized commitment moves the number more than ten small ones.

Composition is done at full precision and rounded once, over a rolling 30-day window.

What counts as eligible compute

Coverage is measured only over commitment-eligible services — the compute and instance services AWS lets you cover with a Savings Plan or Reserved Instance:

• Savings Plan–eligible compute — EC2, Fargate (ECS/EKS), and Lambda.
• Reserved Instance–eligible instances — RDS, Redshift, ElastiCache, and OpenSearch.

Spend on services that can’t be committed never enters the denominator, so it can’t distort the number. Spot usage is excluded (it can’t be covered by a commitment and is already discounted). This eligibility list is versioned with the methodology.

How utilization is measured

Utilization is derived from the used vs unused commitment quantity on each commitment (the hours or committed dollars you consumed versus what you reserved), weighted by each commitment’s cost. CloudQuell deliberately does not infer utilization from the used/unused split AWS stamps on commitment fee line items: AWS marks an entire RI fee “unused” if the RI has any unused quantity, which would understate utilization for any partially-used reservation. Measuring quantity per commitment avoids that and matches what the provider’s own utilization report shows.

Edge cases

• No commitments — coverage is 0% (nothing is committed) and Effective Coverage % reads as a “no commitment program” state rather than a bare zero.
• Data still settling — when your commitments are actively covering usage but a period’s commitment data hasn’t fully posted yet, CloudQuell shows coverage and holds Effective Coverage % back rather than display a misleading value. It fills in once the data settles.
• Credits, refunds, taxes — excluded. The metric reflects usage and commitments only.

Worked example

A commitment program can leak money two ways — under-committing (too much spend stays on-demand) or under-utilizing (you bought commitments you don’t consume). Effective Coverage % catches both:

Org	Coverage	Utilization	Effective Coverage %	What it means
Org A	95%	70%	66.5%	Over-committed but disciplined — most spend is on commitments, but a third go unused. Shorten terms or shift to Savings Plans.
Org B	40%	95%	38%	Disciplined utilization, under-committed — 60% of compute is still on-demand. Buy more.
Org C	75%	90%	67.5%	Healthy on both axes, with a small tuning opportunity.

Org A and Org C land at roughly the same Effective Coverage % from opposite starting points: the metric tells you the level you’re operating at; the breakdown tells you where to act.

Reading the number

Rough heuristics — the right target depends on your workload mix:

• Below 30% — no active commitment program; most compute is at on-demand rates.
• 30–55% — commitment management is in flight but immature.
• 55–75% — an active practice with a regular review cadence.
• 75%+ — a mature program. The remaining gap is usually structural (bursty workloads that shouldn’t be committed).

Steady-state workloads (an always-on database, for example) can hit 90%+; bursty workloads should aim lower so utilization stays healthy.

Where to find it

On the Commitments view, Effective Coverage % is the top-line KPI, above its coverage and utilization breakdown. Alongside it:

• A 12-month trend chart, aggregated org-wide and broken down per integration.
• Right-sizing guidance on underutilized commitments — a low utilization number usually means over-purchase or a size/region mismatch, not a reason to avoid commitments.
• On the Savings view, each commitment recommendation shows its projected impact on the number (for example, “18% → 24%”).

Effective Coverage % is available on every plan, including Free.

Methodology version history

• v1.0 — Coverage at on-demand-equivalent (list) rates over the eligible-service set; utilization from per-commitment used/purchased quantity, dollar-weighted across the portfolio.