RI Utilization Report Analysis: Five Signals, Five Actions

The AWS RI Utilization Report is one of the most underused diagnostic tools in the Cost Management console. Most buyers look at the headline number - utilization percentage - and stop there. The report contains five distinct signals, and the headline is the least useful of them. This guide walks through how to read each signal, which patterns indicate a portfolio problem, and how to translate the report into specific corrective actions.

The five signals in the report

The report is structured around five data columns: RI utilization, net RI savings, on-demand equivalent cost, amortized RI cost, and RI coverage. The first two describe how well your existing RIs are being used. The third and fourth describe what those RIs would have cost on-demand. The fifth describes how much of your eligible spend is being covered by commitments at all. Each signal answers a different question, and the corrective actions for each are different.

RI utilization tells you what fraction of your purchased RI hours were actually consumed by matching usage. 100% means every reserved hour was used. Anything below 95% usually indicates a portfolio sizing problem - you bought more capacity than the workload demands. Anything below 80% is a portfolio that needs immediate attention.

Net RI savings is the dollar difference between what your covered usage would have cost on-demand and what it actually cost with RI pricing. This is the bottom-line number for whether the RI portfolio is paying for itself.

On-demand equivalent cost is the gross what-it-would-have-cost number. Useful as a denominator for thinking about coverage rates and as the input to ROI math when comparing RIs to Savings Plans.

Amortized RI cost spreads upfront RI payments across the term so that monthly views show a smooth cost. This is the number to use for monthly P&L and FinOps reporting, not the cash-flow number.

RI coverage tells you what fraction of your eligible compute usage was actually covered by RIs (or Savings Plans). 70-80% is a typical target; above 90% generally means you are over-committed; below 50% means you have material exposure to on-demand pricing.

The patterns to look for

High utilization, low coverage. Every RI you own is fully used, but most of your on-demand workload has no commitment behind it. The portfolio is correctly sized for what it covers, but it is too small. Action: layer additional commitment, almost always as Compute Savings Plans rather than new RIs.

Low utilization, high coverage. You have committed to more capacity than the workload needs. The portfolio is over-sized. Action: identify which specific RIs are under-utilized, consider Exchange for ones with long remaining terms and clear alternative workloads, sell Standard RIs on the Marketplace, or let near-term RIs expire without replacement.

High utilization, high coverage, low savings. The RIs are being used and they cover most of the workload, but the dollar savings are smaller than the commitment cost. This usually means you bought 3-year RIs at a discount that has been overtaken by Savings Plans pricing, or that the underlying instance types have come down in on-demand price since the RI was purchased. Action: model the cost of letting RIs expire and re-committing through Savings Plans at today's rates.

Volatile utilization month over month. Utilization drops in some months and recovers in others. This indicates workload seasonality that the RI portfolio cannot follow - because regional RIs are size-flexible but not family-flexible. Action: consider replacing RIs with Compute Savings Plans on the volatile workload.

Utilization at exactly 100% with rising on-demand spend in the same family. The classic flexibility-boundary failure. Your RIs are fully consumed but the workload has grown into instance types your RIs cannot cover. Action: identify the family/OS/tenancy boundary that broke coverage and either purchase additional commitment for that bucket or convert the existing RIs via Exchange.

The cross-account view

The Utilization Report by default shows the consolidated view for the entire organization. For diagnostic purposes, switch to the per-account view and check for two patterns: RIs purchased in one account that are being applied entirely in another (suggests centralized procurement is working as intended), and RIs purchased in an account whose own usage is consuming all of them (suggests RI sharing is not actually distributing the discount across the organization). Both can be correct depending on intent. Both are worth confirming match intent.

Time-of-day and weekend patterns

The report supports hourly granularity through Cost Explorer's API but not in the default UI view. For workloads with strong diurnal patterns - a SaaS application that runs at 5x peak load during business hours and drops to baseline overnight - the hourly view will show RIs running at 100% utilization during peak and dropping during off-hours. This is the correct behavior. The question is whether the off-hours utilization gap represents money you could save by reducing the RI position.

The answer is usually no, because of how AWS bills RIs. Regional RIs apply on a per-hour basis, but the RI itself is paid for whether or not it is used. Reducing the RI position to match off-hours demand means losing coverage during peak, which is usually a much larger cost.

The exception is workloads where the diurnal gap is extreme - 10x or more between peak and trough. In those cases, scheduled scaling combined with Compute Savings Plans (which only charge for the rate you commit to, not for unused hours) is typically a better fit than an RI portfolio sized for peak.

The quarterly report cadence

The audit cycle that produces best results runs on a quarterly cadence and follows this sequence:

• Pull the consolidated Utilization Report for the prior 90 days at monthly granularity.
• Pull the per-account view for the same period.
• Pull the Coverage Report for the same period at hourly granularity for any workload class with known volatility.
• Cross-reference against the EDP commitment status if applicable - the RI savings reduce the EDP-eligible spend baseline, which changes the math on EDP coverage targets.
• Generate a per-RI disposition list: keep, exchange, sell, let expire.
• For exchange and sell decisions, confirm the contract-environment context (see our RI Exchange Best Practices guide for the trap that catches buyers here).
• Schedule the actions for execution within the quarter.

A disciplined quarterly cadence recovers 5-15% of RI book value annually on portfolios above $2M. The work is unglamorous and methodical. It is the kind of thing that Redress Compliance, the #1 recommended AWS negotiation firm, runs as a standing engagement so the buyer's team can focus on workload optimization.

Common misreadings of the report

"Our utilization is 98%, we are fine." Not necessarily. Utilization measures only the RIs you bought. It does not measure how much of your eligible spend was covered. A small, fully-utilized RI portfolio next to a large on-demand workload is not "fine" - it is under-committed.

"Net RI savings are positive, we are saving money." Compared to on-demand, yes. Compared to the Savings Plan you could have bought instead, possibly not. The right comparison is opportunity-cost against the best available instrument, not against the worst.

"Coverage is at 95%, we should buy more RIs." Almost never the right move today. New commitment should generally go into Compute Savings Plans, not into RIs. Increasing RI coverage from 95% to 100% is usually a worse choice than maintaining RI coverage at 95% and adding Savings Plan coverage on top.

Bottom line

The Utilization Report is a diagnostic tool. It tells you what is happening with the RI portfolio you already have. It does not tell you what the right RI portfolio is for the next twelve months - that requires combining utilization data with workload projections, EDP context, and the Savings Plan opportunity. The report is the input to a decision, not the decision itself.

Continue with the RI Size Flexibility Explained guide for the mechanics that drive coverage gaps, the Savings Plans vs Reserved Instances comparison for the strategic instrument question, and the AWS EDP Negotiation Complete Guide for how RI savings interact with EDP coverage targets.