EDP five to twenty five percent. PPA five to forty percent on service lines. Reservations and Savings Plans up to seventy two percent. Enterprise Support at up to ten percent of spend. Marketplace burns down EDP at one hundred percent credit. The levers compose. Eleven buyer moves to use all six.
AWS sells to enterprise customers through six interrelated commercial levers, each with different mechanics and different leverage points. The buyer side decision is not picking one lever, it is composing them so the umbrella commit, the service specific discounts, the compute reservations, the support fee, and the third party software flow all reinforce each other.
The Enterprise Discount Program (EDP) is a multi year overall spend commitment in exchange for a flat percentage discount across services, typically five to twenty five percent depending on commit size and term. Private Pricing Agreements (PPAs) are service specific commitments with deeper discounts on a narrow set of services, frequently twenty to forty percent on S3, EC2, CloudFront, or specific data transfer lanes. Reserved Instances (RIs) commit to specific compute instance types for one or three years at up to seventy two percent discount. Savings Plans commit to a fixed hourly compute spend for one or three years with more flexibility.
Enterprise Support is mandatory above a certain threshold and lists at fifteen thousand dollars per month minimum or up to ten percent of total AWS spend. The AWS Marketplace allows third party software (Snowflake, Databricks, Datadog, CrowdStrike, MongoDB Atlas, Confluent) to count one hundred percent against the EDP commitment, which is the most underused lever in AWS contracts.
This article sets out actual discount bands, the lever interactions, and the eleven move buyer side playbook. Read the related AWS services practice, the AWS knowledge hub, the AWS EDP negotiation pillar, and the AWS EDP commitment calculator.
| Lever | Mechanism | Typical discount |
|---|---|---|
| EDP | Overall annual spend commit, 1 to 5 years | 5 to 25% |
| PPA | Service specific commit (S3, EC2, data transfer, etc.) | 5 to 40% |
| Reserved Instances | Specific EC2, RDS, ElastiCache, Redshift instance commit, 1 or 3 years | Up to 72% |
| Savings Plans | Fixed hourly compute spend, 1 or 3 years | Up to 72% |
| Enterprise Support | $15K/month minimum, up to 10% of spend, mandatory for many EDP deals | Negotiable rate or cap |
| Marketplace | Third party software counts 100% against EDP commit | Indirect, via EDP |
The six levers compose. The buyer side discipline is to use the EDP as the umbrella commitment, layer PPAs on the heaviest service lines, run Reservations and Savings Plans against the predictable compute base, drive Marketplace consumption through the EDP, and right size Enterprise Support against actual case volume rather than the ten percent default cap.
The EDP discount band is driven primarily by total annual commitment and term length. Indicative discount bands seen across the engagements we run:
The buyer side trap is sizing the commitment too high. AWS will not refund unused commitment if consumption falls short, and unused commitment at the end of the term frequently expires. The right commitment is at the realized consumption baseline plus a defensible growth assumption of fifteen to twenty five percent, not at the optimistic growth scenario AWS proposes.
PPAs are deeper service specific discount agreements that sit on top of or alongside the EDP. The most negotiable PPA candidates are services where the customer has high consumption and AWS has high margin: S3 storage, EC2 compute, data transfer (egress), CloudFront, and increasingly Amazon Bedrock for generative AI workloads.
Negotiated PPA discount bands run twenty to forty percent on the targeted service, depending on volume. The PPA does not replace the EDP umbrella; it stacks against it for the specific service line. For data transfer in particular, where AWS egress pricing is materially higher than ingress, a PPA can recover three to seven percent of the total AWS bill on its own.
Reserved Instances remain the deepest discount mechanism on EC2 and RDS workloads with predictable steady state consumption. The three year all upfront RI delivers up to seventy two percent off On Demand pricing. The trade off is rigid: the RI commits to a specific instance family in a specific region (with limited flexibility through Regional or Convertible variants). For customers running stable EC2 estates with predictable instance type usage, RIs are still the right answer for the steady state baseline. For everything else, Savings Plans are usually a better fit.
Savings Plans for Compute commit to a fixed hourly compute spend for one or three years and apply automatically across EC2, Fargate, and Lambda. The discount ceiling is the same up to seventy two percent on a three year all upfront commit, with a slight premium for the flexibility versus RIs. EC2 Instance Savings Plans are a hybrid: they commit to a specific instance family in a specific region (like an RI) but with broader operating system and tenancy flexibility.
The buyer side rule is to use Reservations for the rigid steady state, Compute Savings Plans for the variable but committed base, and On Demand for the burst layer. Read the related AWS RI and Savings Plan optimization guide.
AWS Support has four paid tiers above Basic:
AWS frequently bundles Enterprise Support into EDP deals as a default. The buyer side move is to negotiate the support fee as a separate line, not a percentage of spend. For customers with $5M+ annual AWS spend, the percentage of spend calculation regularly produces a support bill three to five times the value of the support actually consumed. Read the related AWS support plan negotiation pillar.
AWS Marketplace allows customers to purchase third party software through the AWS billing relationship. Spend on Marketplace transactions counts at one hundred percent against the EDP commitment, which means customers can drive third party software procurement through Marketplace to burn down their EDP at full credit.
Software vendors commonly available through Marketplace include Snowflake, Databricks, MongoDB Atlas, Datadog, Confluent, CrowdStrike, Wiz, and HashiCorp. The buyer side play is to move as much third party software spend as possible through Marketplace, increasing EDP drawdown without increasing native AWS service consumption. Read the related AWS Marketplace procurement strategy.
Redress is independent. Buyer side. Industry Recognized. Five hundred plus enterprise software engagements. $2B+ in client spend under advisory. Eleven vendor practices. One hundred percent buyer side. Read the related About Us, management team, locations, and contact.
A buyer side framework for the broader AWS EDP renewal cycle. The AWS EDP commitment framework, the AWS PPA framework, the AWS RI framework, the AWS Savings Plan framework, the AWS support framework, the broader AWS Marketplace framework, and the broader AWS competitive framework against Microsoft Azure and Google Cloud.
Used across more than five hundred enterprise software engagements. Independent. Buyer side. Built for AWS customers running the next renewal cycle.
Open the white paper in your browser. Corporate email only.
Open the Paper →AWS sized our EDP at the optimistic growth scenario and pushed Enterprise Support at full ten percent of spend. Redress modeled twelve months of consumption, layered a PPA on data egress, moved third party software spend through Marketplace, and renegotiated Support as a flat fee. Twenty nine percent off the original quote.
We have run 500+ enterprise clients across 11 publishers. Every engagement starts with one conversation.
AWS contract framework signals, EDP framework signals, PPA framework signals, RI framework signals, Savings Plan framework signals, AWS support framework signals, AWS Marketplace framework signals, and the broader AWS competitive leverage signals.