On this page
AWS#
AWS reps are trained to defend the EDP starting point with switching-cost framing and strategic-value anchoring. Here is how to push through.
“EDP discounts at your spend level top out at 12%.” False. Reply: "Our benchmark data shows 3-year EDP deals at our commitment volume closing at 18 to 22% with a PPA layered on EC2 and egress. Please escalate to the strategic-account team and come back with a proposal that reflects competitive benchmark reality." The 12% anchor is an AE-level conversation; deal desk authority extends well beyond it.
“We don't do Private Pricing Agreements.” Categorically false. PPAs are standard for strategic accounts. Counter: "We are aware PPAs are the mechanism AWS uses for strategic-retention deals with competitive pressure. We have a live Azure proposal scoped by Deloitte and expect a PPA layered on EC2, S3, and data transfer to close. Escalate to your strategic-account team." Hold firm, PPAs materialize when the AE escalates and the deal-desk team evaluates churn risk.
“MACC is fixed at 100%.” False on commitments above $5M. Counter: "100% MACC is an unacceptable risk given our product-launch variability. We will sign at 85% MACC with quarterly flexibility or we will reduce the commitment size. Please revise." 85% MACC with quarterly true-up is approved routinely on strategic deals.
“Egress pricing is standard and non-negotiable.” True of published rates, false of strategic contract pricing. Counter: "Our data-heavy workloads make egress the largest variable cost. A strategic commitment without egress protection is not commercially viable. Please return with tiered egress pricing or a dedicated PPA on data transfer." Egress PPAs reducing effective rates by 40 to 70% exist on strategic AWS deals.
“This EDP expires at quarter-end.” Sometimes true, often a close technique. Get the expiration in writing. If the terms are not right, let it expire, a Q2 EDP missed becomes a better-discounted Q4 EDP.
— amazon-web-services-aws-discount-negotiation
Microsoft Azure#
Microsoft reps defend EA pricing with portfolio value framing and "you're already getting EA discount" anchoring. Here is how to push through.
“Your MACC discount at this spend is 12%, that's our standard.” Reply: "Our benchmark data shows 3-year MACC at our commitment level closing at 18 to 22% with AWS competitive pressure. Please escalate to the strategic-account team and return with a proposal that reflects benchmark reality." The 12% anchor is AE authority; deal desk can materially exceed it.
“Azure Hybrid Benefit only applies to eligible licenses.” True, but the eligibility scope is larger than reps volunteer. Counter: "We need a full AHB eligibility audit across Windows Server and SQL Server SKUs. Please engage Fast Track or your FastTrack team to run the audit and apply AHB at the SKU level. We are not signing the MACC without AHB fully scoped." AHB audits are approved routinely and often reveal millions in previously-unapplied savings.
“EA bundle pricing is portfolio-level only.” Categorical refusal tactic. Counter: "We will not sign portfolio-level pricing. Please provide line-item pricing for Azure MACC, M365 E5, Dynamics 365, and Power Platform. We will benchmark each component independently before signing the EA." Microsoft deal desk accepts line-item requests on strategic accounts, though the AE will push back initially.
“True-up is charged at list price.” Standard language, negotiable terms. Counter: "True-up at list destroys the economics of MACC. We will sign at committed rate for true-up within 20% of commitment. Overage above 20% can be at tiered rates, but the first 20% must be at MACC discount depth." Approved regularly on deals above $5M MACC.
“This EA pricing expires at fiscal year-end.” Sometimes true, often a close technique. Get the expiration in writing. If terms aren't right, let it expire, a June deal missed becomes a better-discounted December deal.
— microsoft-azure-discount-negotiation
Google Cloud#
GCP reps, unlike AWS and Azure, often accept aggressive first counters because their commercial mandate is share capture. The "no" conversations usually center on contract structure rather than headline discount.
“Sustained-use discount is automatic, there's no additional concession.” Reply: "Sustained-use is a baseline, not a negotiation. Our benchmark data shows custom contracts with negotiated platform discount ADDITIONAL to sustained-use. Please return with a custom-contract discount on top of sustained-use baseline." Routine approval on commitments above $1M annual.
“BigQuery Editions slot pricing is fixed.” False on strategic accounts. Counter: "We will model 500-slot BigQuery Editions commitment against on-demand spend on our actual query patterns. If slot pricing does not deliver 35%+ savings, we will stay on-demand and negotiate on-demand rate discount instead." Google routinely concedes on BigQuery slot pricing when the alternative is losing the analytics workload.
“Egress pricing is standard.” Same response as AWS/Azure. Counter: "Our data-heavy workloads make egress the largest variable cost. A strategic commitment without egress protection is not commercially viable. Please return with tiered egress or dedicated Interconnect pricing." Google is typically more accommodating on egress than AWS or Azure.
“Vertex AI and Gemini commitment is not part of the compute negotiation.” Counter: "We need a bundled strategic commitment across compute, data, and AI services. Please engage your AI/ML specialist organization and structure a combined commitment with Vertex AI, Gemini, and TPU discount depth alongside Compute Engine CUDs." Google's AI commercial teams have substantial discount authority they do not proactively share.
“We can't accept reduction rights on the commitment.” Counter: "We will not commit to fixed multi-year spend without mid-term flexibility. We need 15% reduction rights at year-2 without penalty or we reduce the commitment size. Please revise." Approved regularly on strategic accounts.
— google-cloud-platform-gcp-discount-negotiation
Databricks#
Databricks' enterprise reps are trained on specific objection-handling scripts. Here's how to move through them.
"DBU commitment is the only way to get discount." True at surface level, but the depth of discount on commitment is highly negotiable. Counter: "We understand. The question is what commitment structure captures best-in-market economics. Multi-workload commitment, Unity Catalog bundling, and serverless pre-negotiation all improve the discount tier. Please submit a proposal with all three modeled against our workload."
"Serverless pricing is premium and not discountable on commitment." Revenue protection. Counter: "We're planning material serverless adoption during term. Without pre-negotiated serverless commitment pricing, effective 3-year cost is materially higher than this proposal implies. Please include serverless DBU commitment pricing explicitly with rollover and fungibility to non-serverless."
"Cross-workload DBU fungibility isn't how the platform works." This is the most important objection to push through. Counter: "We will not commit to DBUs in a workload-specific silo that punishes mix shifts. We are committing to Databricks as a platform, not individual SKUs. Fungibility across SQL, ML, Workflows, and Model Serving is a hard requirement." Databricks will concede this on strategic accounts; many customers simply do not ask.
"We can't cap per-DBU pricing, that's not in our standard agreement." Counter: "Every major SaaS and data platform contract at our company has CPI-capped pricing. If Databricks is unwilling, we'll reduce commitment duration to 12 months and re-evaluate annually, with Snowflake Snowpark and Microsoft Fabric included in the re-evaluation." The short-term alternative plus the competitive threat usually unlocks the cap.
"Unity Catalog is priced separately and we can discuss at renewal." Strategic delay. Counter: "Unity Catalog adoption is a year-one deployment commitment. Bundle at signing with phased deployment milestones and deactivation rights if we don't reach adoption targets. Otherwise we will evaluate alternative governance tooling (Atlan, Collibra, Alation) separately." Databricks wants Unity Catalog adoption; the bundling concession exists.
— databricks-discount-negotiation