The published 10 percent support rate settled at 3 to 5 once it was coupled to the commitment
AWS support is a sliding percentage of consumption presented as a service tier, which means it is a variable cost most buyers model as a fixed one. The percentage is the lever, and it moves further inside the commitment envelope than it ever does on its own.
Prepared by Redress Compliance · August 15, 2026 · AWS advisory. Based on 30 to 40 benchmarked AWS support negotiations, 2024 to 2025.
Executive summary
Support is a percentage, not a fee. Enterprise Support opens near 10 percent of monthly consumption and tapers, so a growing bill grows the support line silently.
At scale the blended rate lands far below the headline. Accounts above roughly one million dollars in monthly consumption settled in the 3 to 5 percent band.
Coupling to the commitment is the largest single lever, cutting the effective rate 20 to 35 percent against a standalone support agreement.
Most estates buy the tier they need for their riskiest account and apply it everywhere. Production needs Enterprise. Sandbox does not.
Seven of ten buyers paid for Technical Account Manager scope they never invoked, worth 8 to 15 percent of the line.
The five levers, and what each one moves
| Lever | What it does | Typical movement |
|---|---|---|
| Commitment coupling | Places support inside the commit envelope so the negotiated discount applies | 20 to 35 percent off the effective rate |
| Tier break enforcement | Writes the lower band rate into the contract rather than leaving it to the calculator | Toward the 3 to 5 percent band at scale |
| Account scoping | Enterprise on production, lower tiers on development, test and sandbox | Removes spend that never needed the premium |
| Scope trim | Cuts Technical Account Manager scope that is never invoked | 8 to 15 percent of the support line |
| Savings Plans | Lowers consumption, which lowers the base the percentage applies to | Compounds with every other lever |
The levers are additive, which is why they should be pulled together. Commitment coupling sets the discount, tier breaks set the rate, account scoping removes spend that does not need the premium tier, and Savings Plans shrink the base underneath all of it. Negotiated separately, each looks modest. Stacked, they moved the support line 20 to 35 percent in benchmarked deals. Sequence matters too: lock the commitment and its discount on the proven baseline first, then apply Enterprise Support to that envelope, so you capture the discount on support without over committing to reach it.
Right sizing the tier per account
- Enterprise Support earns its premium on complex production estates with real severity one exposure and active Technical Account Manager use. That is a smaller set of accounts than most estates apply it to.
- Business Support on production gives 24 by 7 access without the Technical Account Manager premium, which covers many workloads that carry no severity one risk.
- Developer Support on non production is business hours coverage for development and test, where an out of hours response was never going to be requested.
- A mixed model puts Enterprise only where the risk lives, and is almost always cheaper than one tier applied estate wide.
- Third party managed services are worth pricing for estates that need hands on operations rather than escalation paths.
- Flag every non production account on Enterprise first, since that single audit usually finds the fastest reduction available, alongside the wider controls in the vendor management playbook.
The AWS EDP negotiation brief
Commitment sizing, the discount thresholds, the support coupling, and the buyer side moves across the AWS estate.
Get the brief →A percentage of spend dressed up as a service tier
The account team position is that Enterprise Support is a fixed percentage of spend and not open to negotiation. In roughly 8 of 10 estates we benchmarked, the line moved 20 to 35 percent once it was coupled to the commitment and the tier breaks were written into the contract. The published rate is where the conversation starts.
What makes support unusual among cloud costs is that it is charged as a percentage of something else. Nothing about the support workload changes when consumption doubles, but the invoice does, automatically and without anyone approving it. That is a reasonable pricing model from the vendor's side and a difficult one for a buyer to govern, because the cost grows through a mechanism that never produces a decision point. Most estates only discover the size of the line when someone graphs it against total spend.
The sliding scale is the second half of the problem. Rates step down at higher spend bands, so an estate that has grown into a lower band has already earned a better rate. Earning it and receiving it are different things. Writing the break into the contract rather than trusting a pricing page is the difference, and it is the single cheapest clause in the agreement to get right. Accounts above roughly one million dollars in monthly consumption settled at 3 to 5 percent in our file, which is half or less of where the conversation opened.
Then there is the part nobody audits. Seven of ten buyers paid for Technical Account Manager scope they never invoked, worth 8 to 15 percent of the line, and the same pattern appears in tier selection: Enterprise applied across sandbox and test accounts that will never raise a severity one case. Severity one entitlement is the real reason to hold Enterprise Support, and most accounts never raise one. Price the scope you use rather than the maximum offered, put Enterprise only where the risk is, and let Savings Plans shrink the base underneath. The commitment mechanics sit in the EDP negotiation brief and the renewal sizing in the renewal strategy.
- Support band, tier mix and scope modeled per account with dollar figures
- Every risky clause flagged with the exact quote, the page, and the replacement language
- A ranked savings queue your team can work through
What the support file shows
Across roughly 30 to 40 AWS support negotiations benchmarked in 2024 and 2025, the blended rate settled well below the published floor once support stopped being negotiated on its own:
For accounts above roughly one million dollars in monthly consumption, against a headline that opens near 10 percent.
Paid for by 7 of 10 buyers and never invoked, which is the cleanest line to trim at renewal.
The patterns: support modeled as a fixed cost, tier breaks left to the calculator, and one tier applied across accounts with completely different risk profiles.
The buyer side move is to negotiate support inside the commitment, never beside it. The wider library sits in the AWS practice.
Your first five moves
- Map every account to a support tier and flag any non production account sitting on Enterprise.
- Forecast support as a percentage of projected spend, not as a fixed fee, so the line is visible before it grows.
- Couple the support line to the commitment and confirm in writing that it counts toward the committed number.
- Enforce the tier break rate in the contract, not the calculator, and trim Technical Account Manager scope you do not invoke.
- Layer Savings Plans on the steady state to shrink the base, then benchmark the blended rate against the 3 to 5 percent band. The AWS practice runs the model with you.
Frequently asked questions
How is AWS support actually priced?
As a sliding percentage of your actual monthly consumption rather than a flat fee, subject to a minimum monthly charge. Enterprise Support opens near 10 percent of monthly usage and tapers as spend rises, which means a growing estate grows the support line automatically even when the support workload has not changed at all.
Is the support percentage negotiable?
Yes, despite the account team position that it is fixed. In roughly 8 of 10 estates we benchmarked, the support line moved 20 to 35 percent once it was coupled to the commitment and the tier breaks were written into the contract rather than left to the calculator. The list rate is an opening position.
What rate should a large estate expect?
Accounts above roughly one million dollars in monthly consumption settled in the 3 to 5 percent band in benchmarked deals, not the headline 10 percent. The sliding scale steps down at higher spend bands, so the practical task is enforcing the break you have already earned.
Why does coupling support to the commitment matter so much?
Because it converts support from a standalone tax into a negotiated line inside the commitment envelope. Support spend can count toward the commitment and inherits the negotiated discount, and coupling it cut the effective rate by 20 to 35 percent against a separate support agreement in our engagements.
Does every account need Enterprise Support?
No, and this is where most of the waste sits. Enterprise earns its premium on complex production estates with real severity one exposure and active use of the Technical Account Manager. Sandbox, development and test accounts rarely justify it, and a mixed model puts Enterprise only where the risk actually lives.
What is the Technical Account Manager scope worth?
Only what you invoke. Seven of ten buyers paid for scope they never used, worth 8 to 15 percent of the support line. Price the scope you actually consume rather than the maximum offered, and trim the rest at renewal rather than carrying it another year.
How do Savings Plans affect the support bill?
They shrink the base it is calculated on. Because support is a percentage of consumption, lowering the consumption rate through Savings Plans on steady state workloads lowers the support line at the same time, which is the one lever that improves both numbers with a single decision.
Where should the tier break be enforced?
In the contract, not the calculator. Write the lower band rate into the agreement so the break applies when your spend reaches it, rather than relying on a pricing page to award it automatically. This is the difference between a rate you have earned and a rate you actually receive.
Negotiating AWS 9: Credits, Marketplace and Support
Credits that can manufacture your own shortfall, marketplace as insurance rather than saving, support billed on gross spend, and the arbitrage that retires commitment while staying out of the support fee base.