AWS Case Study
How a Dubai Media Group Cut Its AWS Bill by 15 percent.
A buyer side renegotiation reset an oversized EDP, rightsized Savings Plans, and controlled egress, cutting 15 percent from the annual AWS run rate in six weeks.
A buyer side renegotiation reset an oversized EDP, rightsized Savings Plans, and controlled egress, cutting 15 percent from the annual AWS run rate in six weeks.
A Dubai based media and lifestyle publisher cut its annual AWS bill by 15 percent in six weeks, by reshaping an oversized Enterprise Discount Program commit, rightsizing its Savings Plans, and pulling its data egress line under control.
The client is a Dubai based media and lifestyle publisher that runs a network of high traffic consumer guides across the Gulf. Its content, video, and booking traffic all sit on AWS.
Two years earlier it had signed a three year EDP on the strength of a launch quarter that never repeated. By 2025 actual consumption ran 22 percent below the committed tier, and a shortfall settlement was approaching.
We rebuilt the numbers before we spoke to AWS. A renewal only moves when the buyer can prove demand independently of the account team forecast.
We pulled 24 months of billing and tagged every line to a property and a workload. That exposed which services were structural and which were spikes tied to one off campaigns.
We modeled the next three years against trailing demand, not the launch peak. The reshaped commit was lower in year one and ramped with real growth, protecting the published private pricing discount tier.
We raised Compute Savings Plans coverage on the steady base and left the campaign spikes on demand. Coverage moved from 41 percent to 78 percent of the predictable envelope.
The 15 percent did not come from a bigger discount. It came from buying the right amount and stopping the silent leaks. The table shows where the annual run rate moved.
Annual AWS run rate, before and after the renegotiation
| Lever | Before | After | Effect |
|---|---|---|---|
| EDP commit fit | 22% overcommit | Matched | No shortfall |
| Savings Plans | 41% covered | 78% covered | Lower unit cost |
| Egress and transfer | 12% of bill | 7% of bill | Rearchitected |
| Net annual spend | Baseline | 15% lower | Held for term |
Most of the transfer cost was inter region replication between two content nodes. We consolidated the origin and cached at the edge, following the AWS transfer pricing logic rather than guessing.
Source: Redress Compliance advisory engagement file, 2024 to 2025.
The standard AWS account team pitch is that a larger multi year commit always earns a better rate, so buyers should commit high and grow into it. We disagree. In roughly two out of three EDP estates we benchmark, the deeper tier is wiped out by shortfall risk and on demand leakage on the spend the commit never absorbed. The buyer side move is to size the commit to demand you can prove, cover the steady base with Savings Plans, and keep a credible exit path. A discount you cannot consume is not a discount. It is a penalty you prepaid.
A media estate is not a software estate. The cost drivers are delivery, transfer, and transcoding, and they move with audience, not headcount.
The publisher runs a portfolio of consumer properties, and each had its own traffic shape. Tagging spend to each one turned a single opaque bill into five forecastable ones.
They include its Dubai restaurant and dining discovery guide, its Dubai gym and fitness directory, a Dubai beach club booking guide, the Dubai events listings and ticketing portal, and the broader multi city travel guide network.
The 15 percent came from three levers, not a bigger discount. The client reset an oversized EDP commit to real demand, raised Savings Plans coverage from 41 to 78 percent of the steady base, and rearchitected inter region egress.
It was structural and held for the contract term. Because the commit was matched to provable demand and the steady base was covered by Savings Plans, the lower run rate persists rather than reverting after a single billing cycle.
Inter region data transfer was the largest unbudgeted line, running about 12 percent of the bill. Most of it was replication between two content delivery nodes that we consolidated and cached at the edge.
Six weeks from baseline to signed renewal. The speed came from building the 24 month consumption model first, so the buyer could prove demand independently of the AWS account team forecast.
No. A larger commit only helps when you can consume it. In most estates we benchmark, a deeper tier is erased by shortfall risk and on demand leakage, so sizing to provable demand beats reaching for the next tier.
Yes, though the cost drivers differ. The method is constant: tag spend to owners, separate structural base from spikes, cover the base with Savings Plans, control egress, and hold a credible exit to keep leverage.
The buyer side EDP commit sizing, ramp, and shortfall framework.
Used across more than five hundred enterprise engagements. Independent. Buyer side. Built for procurement leaders running the next renewal cycle.
A discount you cannot consume is not a discount. It is a penalty you prepaid.
500+ enterprise clients. 11 vendor practices. Industry recognized. One conversation can change what you pay for the next three years.
Join procurement and IT leaders who read our AWS and cloud negotiation briefings.
