.png)
This is an illustrative example, but the decision is common.
Buying now could lock in more commitment than you need. Waiting could also mean paying On-Demand rates for stable usage that could have received a discount.
It is:
What eligible usage will remain after the planned changes, and how much commitment will still be active when those changes happen?
Use four inputs before buying more
1. Start with eligible usage, not your total AWS bill
A lower AWS bill can come from storage, data transfer, deleted resources, rightsizing, or other costs unrelated to the Savings Plan you are considering.
Narrow the question:
AWS applies active Savings Plans to eligible usage after applicable Reserved Instances. Sharing settings can also affect where the benefit applies across accounts. Importantly, an hourly Savings Plans commitment can only be used within that hour. Unused commitment does not carry into another hour.
That means an 8% monthly decline alone cannot tell you that an $8/hour or $12/hour purchase is safe.
2. Add the changes your teams already know about
Historical billing data cannot see your roadmap.If Engineering plans to retire an application, identify:
how much eligible usage it is expected to remove,
when that change should happen, and
who can confirm the plan.
A migration also does not always remove Savings Plans-eligible demand. The workload may move to another configuration that the same plan can still cover.
AWS confirms that Savings Plans recommendations use historical usage from a selected 7, 30, or 60-day lookback and do not forecast future usage. They also do not include queued or scheduled purchases and are generated as immediate-purchase recommendations. See AWS’s Savings Plans recommendation calculation documentation.
3. Separate today’s commitment position from the position after expiration
Suppose $15/hour of your current $55/hour commitment expires in 45 days.Do not automatically conclude that you need another $15/hour now.
Instead, compare two positions:
Today: What commitments are active, and what eligible demand remains after they apply?
After expiration: What commitments will remain, and what demand do you expect after the planned workload changes?
AWS Purchase Analyzer lets you exclude selected Savings Plans expiring within the next 90 days. But that control does not move the analysis 45 days into the future. The tool still analyzes historical usage and models an immediate purchase.
So use expiration exclusion to understand another portfolio scenario, not as proof that a replacement should start today.
4. Give uncertainty a specific review trigger
Sometimes holding is reasonable, but “wait and see” is not a complete strategy.If rightsizing finishes in three weeks, set the next commitment review after the new usage is visible. If an application retirement is still uncertain, identify the decision date and owner.
There is also a cost to waiting. Stable uncovered usage may remain at On-Demand pricing while you wait for better information.
You therefore need to compare two risks: committing too early versus delaying a useful discount.
Use Purchase Analyzer as a test, not a forecast
AWS Savings Plans Purchase Analyzer lets you compare recommended, target-coverage, and custom commitment scenarios. You can change the term, payment option, lookback period, account scope, and selected expiring plans.A practical test is:
Run the AWS recommendation using your current inventory.
Run a smaller custom commitment using the same scope and lookback.
Compare estimated cost, savings, utilization, and coverage.
Separately adjust your forward assumptions for planned workload changes.
Coverage and utilization are useful signals, but they should not become automatic buying targets. AWS Well-Architected guidance specifically advises against fixed coverage or utilization targets because a higher percentage does not necessarily mean better savings.
Worked example: buying now versus after expiration
Return to the opening scenario.Current commitment is $55/hour, including $15/hour that expires in 45 days.
| New purchase | Before expiration | After expiration |
|---|---|---|
| $0/hour | $55/hour | $40/hour |
| $8/hour | $63/hour | $48/hour |
| $12/hour | $67/hour | $52/hour |
It does not prove that $8/hour is the right purchase.
Buying $8/hour today raises the portfolio from $55/hour to $63/hour for the next 45 days. After the $15/hour expiration, the same purchase leaves a $48/hour commitment.
Now add a clearly hypothetical post-expiration model.
Assume:
the remaining eligible usage would consume exactly $48/hour at Savings Plans rates,
the modeled Savings Plans discount is a uniform 30%,
the equivalent On-Demand cost is therefore $68.57/hour,
there are no additional discounts, purchases, or sharing changes.
| Post-expiration scenario | Total commitment | Remaining On-Demand cost | Total modeled cost |
|---|---|---|---|
| No new purchase | $40/hour | $11.43/hour | $51.43/hour |
| Add $8/hour | $48/hour | $0 | $48/hour |
| Add $12/hour | $52/hour | $0 | $52/hour |
Adding $12/hour would cost $0.57/hour more than the no-purchase case because the modeled $48/hour demand would not consume the full $52/hour commitment.
But timing changes the result.
If the same $48/hour demand already exists before expiration, the current $55/hour commitment already exceeds it. Buying another $8/hour today would increase commitment without reducing modeled On-Demand charges.
That is why a good post-expiration opportunity does not automatically justify an immediate purchase.
Buy, hold, or reassess?
Use the three decisions as actions rather than fixed labels.| Decision | What it means |
|---|---|
| Buy | Remaining eligible hourly demand can support the additional commitment, including a realistic downside case. |
| Hold | A specific upcoming event will materially improve the decision. Set the review date now. |
| Reassess | Your old workload baseline no longer describes the environment. Rebuild it before deciding whether to buy or keep holding. |
What usage will remain after the planned changes?
What commitment will still be active at that point?
If either answer is unclear, define the next review trigger instead of approving a commitment based only on historical usage.
What if you buy too early?
Do not assume you can simply reduce or cancel a Savings Plan later because demand falls.AWS has a limited return process intended primarily for recent purchase mistakes. A Savings Plan with an hourly commitment of $100 or less may qualify for return when it was purchased within the previous seven days and within the same UTC calendar month, subject to AWS return limits and other eligibility requirements.
See the full AWS Savings Plans return policy before relying on this option.
It is not long-term protection against a workload that declines months after purchase.
Where Usage.ai fits
At Usage.ai, we focus on the commitment layer. We identify eligible AWS commitment opportunities and, once you approve a recommendation, automatically initiate the commitment purchase. The commitment is then managed through our Flex Commitment Program. Learn more about Flex Commitment eligibilityWith our Flex Insured Commitments, eligible teams can access up to 57% savings associated with a three-year AWS commitment with none of the commitment risk.If a commitment becomes more expensive than the equivalent On-Demand usage, Usage.ai provides cashback protection to help cover the difference.
The two layers work together: AWS cost governance helps you understand and control spending, while Usage.ai can automate eligible commitment management and help optimize the pricing of stable AWS usage.
Establishing ownership, budgets, and a reliable usage baseline first makes those commitment decisions more informed.
Frequently asked questions
Should I buy more AWS Savings Plans if usage is falling?
Not automatically. Determine whether the decline affects Savings Plans-eligible usage, then account for existing commitments, hourly demand, planned workload changes, and upcoming expirations.
Can I trust AWS Savings Plans recommendations when usage is falling?
Use them as historical evidence, not as a forecast. AWS recommendations use historical usage from the selected lookback period and do not forecast future consumption.
Does an expiring Savings Plan mean I should wait before buying?
Not necessarily. Compare the commitment and demand you have today with what should remain after expiration. Hold only when waiting for a specific event will materially improve the decision.
What is the difference between hold and reassess?
Hold means pausing a new purchase until a defined event or review date provides better information.
Reassess means the old baseline is no longer reliable, so you rebuild the assumptions before choosing whether to buy or continue holding.
Can I cancel a Savings Plan if usage keeps falling?
Do not assume you can. AWS provides a limited return process for qualifying recent purchases, including the seven-day and same-UTC-calendar-month requirements, but other restrictions also apply. It should not be treated as protection against long-term usage decline.
