Last updated: June 2026
Azure reservation recommendations project savings from your recent usage over a 7, 30, or 60-day lookback. They are an input, not an instruction: rightsize first, check the lookback and scope, then buy in tranches against the baseline you are confident in.
- Recommendations reflect recent usage, including waste. Clean the baseline before reading them.
- The 30 or 60-day lookback is usually safer than 7 days for a stable baseline.
- Scope (shared, subscription, resource group) decides where the discount applies.
- Azure reservations save up to about 72% versus pay-as-you-go on covered usage.
An Azure reservation recommendation is a suggestion, generated by Azure Advisor and the Cost Management reservations experience, that estimates which one or three year commitments would save you money based on your recent usage. Azure analyzes a 7, 30, or 60-day lookback of your consumption and projects the savings, and the recommendations can be genuinely useful, but only if you understand that they describe your past usage rather than your ideal future state. This article is part of our Azure cluster; start with the complete guide to Azure cost optimization, the pillar this piece links up to. Acting on commitments is a Lock step in our See, Cut, Lock, Run method, and the sequence is always rightsize first, then reserve the clean baseline.
Where do Azure reservation recommendations come from?
They come from Azure analyzing your recent usage and projecting commitments that would lower the bill. Azure Advisor and the reservations purchase experience in Microsoft Cost Management look at how much on-demand capacity you have consumed over a chosen lookback window and recommend the reservation quantity and term that would have saved the most across that period. Azure reservations themselves save up to about 72% versus pay-as-you-go on covered usage, with the deeper discounts on three-year terms, per Azure reservation pricing. The recommendation tells you what the discount would have captured on your recent behavior; whether you should act on it depends on whether that behavior is clean and stable.
Should you trust the recommendation?
Treat it as an input, not a decision. Because the recommendation is built from recent usage, it includes any idle, oversized, or soon-to-be-retired resources you were running during the lookback, and it will cheerfully recommend committing to all of it. If you reserve on that basis, you lock in a discount on waste for one or three years. The fix is the order of operations: remove the waste, then read the recommendation against the cleaned-up baseline, so you commit to capacity you will genuinely keep using. This is the same discipline behind Azure reserved capacity for SQL and Cosmos DB, where rightsizing the databases comes before reserving them.
How to act on Azure reservation recommendations, step by step
These five steps turn a raw recommendation into a commitment you will not regret.
- Rightsize before you read the recommendationEliminate idle and oversized resources first, so the recommendation is calculated on a clean baseline rather than on waste. The result is a recommendation that reflects what you will actually keep running.
- Choose the right lookback windowCompare the 7, 30, and 60-day lookbacks and prefer the longer window for a stable workload, because it smooths short-term spikes. The result is a recommendation anchored to your steady baseline, not to a temporary peak.
- Check the recommendation scopeConfirm whether the recommendation is shared, single-subscription, or resource-group scoped, since scope decides where the discount can apply. The result is coverage that lands on the subscriptions and resources you intend.
- Buy in tranches, not all at onceCover the confident floor first and add coverage in stages as utilization proves out, rather than buying the full recommendation in one purchase. The result is high reservation utilization and little stranded commitment.
- Review utilization and adjustTrack reservation utilization, exchange or adjust under-used reservations, and re-run the recommendations after any major architecture or usage change. The result is coverage that stays matched to a moving baseline over the term.
About to buy whatever Azure recommended?
Our Azure cost audit rightsizes the estate first, reads the recommendations against the clean baseline, and builds a coverage plan in tranches so you capture the discount without committing to waste. On the performance model, you pay only from realized savings. No savings, no fee.
Book an Azure cost audit →What lookback window should you use?
The 30 or 60-day lookback is usually the safer choice for a stable workload. A longer window averages out short-lived spikes and one-off batch runs, so the recommended quantity reflects the capacity you genuinely run rather than a transient peak. The 7-day lookback is sharper and reacts faster, which is useful only when your usage has recently and permanently stepped up and you want the recommendation to capture the new, higher baseline immediately. The wrong window over-commits in one direction or under-commits in the other, so match the window to how stable and how recently changed your usage is.
How do you avoid over-committing?
Cover the floor, buy in tranches, and review utilization. The safest pattern is to reserve only the baseline you are confident you will keep using, add coverage in stages as the next layer proves stable, and let pay-as-you-go absorb the variable top. Then watch reservation utilization through the term and use exchanges to adjust anything that drifts under-used. This staged, utilization-driven approach is how a reservation program captures most of the available discount while keeping stranded commitment near zero, and it scales across the whole Azure cost optimization program rather than being a one-time purchase.
The lookback windows, recommendation surfaces, and discount ceiling reflect Azure as of June 2026. Verify current recommendation behavior and reservation discounts in the linked Azure documentation before purchasing, because reservation tooling and rates change.
The Azure Cost Optimization Field Guide includes the reservation coverage worksheet and the tranche-buying schedule we apply on engagements. It is the downloadable companion to this article.
Frequently asked questions
Where do Azure reservation recommendations come from?
Azure generates reservation recommendations in Azure Advisor and the Cost Management reservations experience by analyzing your recent usage over a 7, 30, or 60-day lookback and projecting which commitments would save money. They are a starting point, not an instruction.
Should I trust Azure reservation recommendations?
Treat them as a useful input, not a decision. They reflect recent usage, including any waste, so rightsize first and verify the lookback and scope before buying. A recommendation built on idle or oversized resources will tell you to commit to waste.
What lookback window should I use for Azure reservations?
The 30 or 60-day lookback is usually safer than 7 days because it smooths short-term spikes and reflects a stable baseline. Use the shorter window only when usage has recently and permanently stepped up and you want the recommendation to reflect the new level.
How much can Azure reservations save?
Azure reservations save up to about 72% versus pay-as-you-go on covered usage, depending on the service, term, and instance, with deeper discounts on three-year terms. Always size to the baseline you are confident in, because reservations are use-it-or-lose-it commitments.
The short version
Azure reservation recommendations project savings from recent usage, waste included, so they are an input rather than an instruction. Rightsize first, pick the lookback that matches your stability, check the scope, buy in tranches against the confident floor, and review utilization through the term. When you want the estate cleaned and the coverage plan built correctly, that is exactly what our Azure cost optimization service delivers.
Cloud pricing and service behavior change frequently. Verify the specifics in this guide against the providers’ own current documentation and the FinOps Foundation: Azure pricing ↗, Azure documentation ↗ and FinOps Foundation Framework ↗. This article also reflects Cloud Cost Room’s hands-on, vendor-neutral engagement experience.