The AWS EDP Negotiation Playbook

The EDP is a commitment to spend, not a discount on services

The Enterprise Discount Program reads as a service discount: 3 to 15 percent off your AWS bill in exchange for a multi year commitment to spend. Structurally, the commitment is the primary contractual obligation and the discount is the consideration AWS provides in return. If you spend less than the commitment, AWS bills you for the shortfall at the end of the term. The commitment is a take or pay obligation, not a target.

This matters because the question most enterprises ask is "what discount can I get". The right question is "what commitment level can I sustain". A 12 percent discount on a commitment 30 percent above your sustainable consumption is not a savings; it is an exposure. The EDP analysis must start with consumption modeling and end with commitment selection. The discount falls out of the math.

The PSA is where the actual terms live

An EDP signature involves three documents: the EDP itself (a short commercial summary), the AWS Customer Agreement (the standard service terms), and the Private Pricing Addendum (the negotiated commercial provisions). The first two are largely standard. The third is where every negotiated value lives. Marketplace eligibility, RI and Savings Plan treatment, true down rights, eligible services list, growth allowance, renewal mechanics, all sit in the PSA.

Customers who negotiate the EDP front page and accept the PSA as written sign the same EDP every other customer signs. Customers who negotiate the PSA section by section sign an agreement tailored to their consumption profile and risk tolerance. The difference is rarely visible in the discount column. It is consistently visible in the year three financial outcome.

The discount tiers and how they scale

AWS prices EDP discount as a function of three variables: total commitment value, term length, and the negotiating moment. Commitment values above $5M typically attract 5 to 8 percent discounts. Above $20M, 8 to 12 percent. Above $50M, 12 to 15 percent. Term length compounds: three year terms exceed one year terms by 1 to 3 percentage points. The negotiating moment, end of AWS quarter or end of fiscal year, can add another 1 to 2 points.

The tiers are not published. They are inferred from market behavior across many engagements. AWS account teams know them; customers usually do not. The customer side mistake is to accept the first proposal as a function of size, when the same commitment value at a different negotiating moment, with a different term length, in a different competitive context, would attract a meaningfully different rate.

Marketplace EDP inclusion is now table stakes

In 2024, AWS expanded Marketplace EDP eligibility to allow Marketplace purchases to count toward EDP commitment. The change reshaped the math. A customer with significant third party software procured through AWS Marketplace (Snowflake, Datadog, MongoDB, dozens of others) can now apply that spend against the commitment. The implication is that the same business consumption now requires a smaller AWS native commitment to satisfy the same EDP.

Marketplace inclusion is no longer optional. It is the default for any new EDP negotiated in 2025 or later. Customers signing EDPs without Marketplace inclusion are negotiating against a 2023 baseline against an AWS counterparty operating on a 2025 baseline. The customer carries the disadvantage of the asymmetry.

RIs, Savings Plans, and the dual counting question

The most consequential math question in any EDP is whether Reserved Instances and Savings Plans count toward the commitment at list price or at discounted price. AWS accounting treats them as paid spend; the question is what value the spend represents. A $1M Savings Plan purchase counted at list might satisfy $1.4M of EDP commit; counted at discount it satisfies only $1M. Across a three year term that delta can move the effective EDP requirement by 20 to 40 percent.

The answer is contractual, not standard. Some PSAs treat RI and SP at list, some at discount, some with hybrid treatment by service. The customer side mistake is to accept the answer without seeing the contractual language. The first thing every EDP negotiation should establish is the dual counting treatment in writing, before any commitment level is discussed.

The flexibility provisions that determine outcome

Five PSA provisions separate an EDP that compounds value from one that compounds liability. True down rights, allowing the customer to reduce committed value at annual checkpoints. Growth allowance, capping the upside commitment if consumption exceeds the original commit. Eligible services list, specifying which AWS services count toward commit and which do not. Marketplace inclusion, discussed above. Renewal mechanics, requiring negotiation rather than auto-extension at term end.

None of these appear in the discount column. All of them appear in the PSA appendix. Customers who negotiate the discount but not the appendix lose the structural value of the deal. Customers who negotiate the appendix first and the discount second routinely produce better outcomes on both metrics.

AWS counter moves and how to handle them

AWS account teams have a small set of repeatable moves they deploy when a customer signals serious negotiation intent. The first is the migration acceleration framing, in which an existing migration commitment is leveraged to argue for a higher EDP commit than the consumption baseline supports. The second is the bundled architecture review, in which AWS solution architects propose technical changes that conveniently increase the commit AWS is asking for. The third is the deadline manufactured around a calendar quarter end that compresses customer side preparation.

None of these are illegitimate. All of them are negotiation. The playbook includes the standard responses we deploy: migration handling that separates technical commitment from commercial commitment, architecture review that benefits from independent technical advisory rather than AWS solution architect framing, and deadline management that uses the compression as leverage rather than against the customer. Customers who have read the responses in advance handle the moves; customers who encounter them for the first time often do not.

The mid term renegotiation moment

The least understood feature of EDP is the mid term renegotiation right. Most EDPs include language permitting renegotiation if the customer's consumption profile changes materially during the term. AWS rarely volunteers this provision; customers who know about it can use it to extract additional value when their cloud roadmap accelerates or contracts. The trigger language varies, the threshold for "material" varies, but the right exists in most contracts.

The implication is that an EDP signed in 2024 is not a fixed obligation through 2027. It is an obligation subject to renegotiation when consumption shifts. Customers who track their consumption against the commit monthly identify renegotiation moments early and capture the value. Customers who treat the EDP as set-and-forget miss the moments and pay the full commit.