The EA renewal arrives as one number, and it is eleven
Two things happened to the Microsoft Enterprise Agreement between 2020 and 2026: the Level D discount that justified the vehicle fell from 22 to 25 percent to 10 to 12, and Microsoft stopped renewing commercial enrollments below 2,400 licenses, making the EA one option among three rather than the default for a large estate. The renewal arrives as one number on one piece of paper because a single number can only be accepted or rejected; split it into its eleven line items and each one has to survive on its own evidence.
Prepared by Redress Compliance · August 8, 2026 · Microsoft advisory. Based on 60 to 80 EA renewals sat on the buyer side, January 2024 to December 2025.
Executive summary
The tier collapse changed what the vehicle is for.
The Level A to D gap narrowed from 22 to 25 percent in 2020 to 10 to 12 percent in 2026, so large customers face nearly the same per user pricing as mid market ones.
And the commercial renewal floor of 2,400 licenses removed the EA as an option for smaller enrollments entirely: entry starts at 500 users, Level A runs to 2,399, Level D from 15,000, and below 500 the route is CSP.
Three of the eleven line items are settled before you sit down, the price level by license count, eligibility by Microsoft, and the uplift by an unpublished percentage on your expiring rate. The other eight are yours to move.
The engagement file's three measurements price the preparation. Final discounts landed 6 to 9 percentage points above the number the account team first called achievable.
E5 right sizing measured against 90 days of actual usage data, not the deployment plan, returned 8 to 14 percent of the M365 envelope; and MACC proposals sized on Microsoft's forward forecast ran 22 to 38 percent above trailing twelve month Azure consumption.
The average outcome of working all eleven lines was a 20 to 30 percent EA run rate improvement against the publisher's first renewal package.
The structural traps are NCE, the true up ratchet, and the Copilot bundle.
The NCE conversion removes the per user pricing flexibility historically built into the EA unless it is negotiated back into the contract; seats added at true up cannot come off until the renewal anniversary, the one reset point in the term.
And the Copilot push targets broad coverage while the buyer side anchors the cohort to the usage report, not the rollout plan.
The MACC horizon has its own rule: three and four year terms hold leverage, six and seven year terms do not.
Co termination converts three quiet losses into one winnable negotiation.
The EA is a master agreement with the commercial detail in enrollments carrying their own terms and dates, and enrollments expiring on different dates hand the account team a rolling series of small negotiations, each too small to fight: co terminate the dates and you get one date, one number.
And every product family on the table at once.
The 2026 pricing reset lifts E3 and E5 rates at the next anniversary with limited flexibility against the reset itself, which moves all the recoverable value into the eight lines that remain negotiable.
The eleven line items, and where each is set
| Line item | Where the number lives | The status |
|---|---|---|
| Price level, eligibility, and uplift | The price sheet, the 2,400 floor, and the unpublished percentage | Settled before you sit down |
| Baseline seat count | The VLSC relationship summary against real headcount | Yours: the reconciliation |
| M365 SKU mix and Copilot cohort | E3, E5, F3 populations and the Copilot usage report | Yours: the measurement |
| Security and Power Platform attach | Defender, Entra, Intune, Sentinel, and Power BI, priced standalone | Yours: the unbundling |
| Billing cadence and the MACC | NCE term rules and the Cost Management trailing twelve months | Yours: the structure |
| The eight terms protections | Enrollment amendments, not the price sheet | Yours: the paper |
The single number is a design choice, and the split defeats it.
Microsoft prefers one number because one number can only be accepted or rejected, and rejecting it means rejecting the whole renewal.
Eleven numbers each have to survive on their own evidence, the seat count against headcount, the E5 population against 90 days of usage, the Copilot cohort against the usage report, and the MACC against the trailing consumption.
The 20 to 30 percent run rate improvement is the cumulative arithmetic of eight lines corrected at once.
The structural traps, and the counters
- The NCE conversion: the historical per user flexibility is removed unless negotiated back into the contract, a term worth more than most percentage points.
- The true up ratchet: seats added mid term cannot come off until the anniversary, so the baseline scrutiny outranks the discount scrutiny.
- The Copilot bundle: the publisher pushes broad coverage, and the buyer anchors to the cohort the usage report justifies, not the rollout plan.
- The MACC horizon: sized to trailing twelve month consumption, on three or four year terms that hold leverage, never six or seven that surrender it.
- The enrollment calendar: co terminated to one date, converting the rolling small negotiations into one with everything on the table.
The Microsoft EA renewal playbook
Forty pages: the price level math, the 2,400 seat floor, NCE conversion, the Copilot cohort test, MACC sizing, and the eight terms protections.
Get the white paper →The arithmetic, at published list
The guide's method is arithmetic a CFO can rebuild: every list price quoted is Microsoft's published figure, the E5 step prices against the measured feature usage rather than the deployment plan, the Copilot cohort prices against weekly active data.
And the MACC prices against the Cost Management trailing twelve months, because each of the three measurements in the file, the 6 to 9 points, the 8 to 14 percent, and the 22 to 38 percent, is the gap between a vendor forecast and a buyer measurement.
The five overspend categories the renewal resets run in the Microsoft overspending report, the EA against CSP crossover in the program comparison, the Copilot meter map in the Copilot licensing guide, and the negotiation sequence in the EA negotiation strategies.
- Percentile standing for your exact deal size and industry, from real closed transactions
- Scenario simulation before the call: test alternative terms and see the financial impact of each
- A negotiation playbook, talking points, and a two page executive brief on day one
What we saw across EA renewals, 2024 to 2025
Our Microsoft practice sat on the buyer side of roughly 60 to 80 EA renewals between January 2024 and December 2025, and three measurements from that file carry the guide:
Of the M365 envelope, measured against 90 days of usage rather than the deployment plan.
Where final discounts landed above the account team's first stated ceiling.
The strategic frame is the tier collapse: with the Level D advantage compressed to 10 to 12 percent and the 2,400 floor excluding smaller enrollments, the EA is now a commercial vehicle to be justified rather than a default to be renewed, one option among three whose case rests on scale, stability.
And the negotiated terms rather than the program discount that once carried it.
The 2026 cycle is a commercial, and running it that way, the eleven lines split, the three measurements taken, and the eight movable numbers each argued on its own evidence, is what separated the 20 to 30 percent improvements from the accepted packages.
Your first five moves
- Split the renewal into the eleven lines, because the single number can only be accepted or rejected.
- Measure the E5 population against 90 days of usage, the 8 to 14 percent of the envelope the plan hides.
- Size the MACC to trailing consumption on a three or four year term, against proposals running 22 to 38 percent high.
- Anchor Copilot to the usage report cohort, and negotiate the NCE flexibility back into the paper.
- Co terminate the enrollments to one date, one number, everything on the table. The Microsoft practice runs the renewal with you.
Frequently asked questions
What is the Microsoft Enterprise Agreement in 2026?
A three year volume commitment starting at 500 users or devices, with a baseline committed at signing, pricing set by volume level, and growth reconciled annually through the true up: Level A runs 500 to 2,399, Level D from 15,000.
And Microsoft stopped renewing commercial enrollments below 2,400 licenses.
With the Level D discount compressed to 10 to 12 percent, the EA is one option among three rather than the default.
How much have EA discounts fallen?
The Level A to D gap narrowed from 22 to 25 percent in 2020 to 10 to 12 percent in 2026, so large customers face nearly the same per user pricing as mid market ones, and the program discount that justified the vehicle no longer does.
The recoverable value moved to the negotiated layer: the eight movable line items, the terms protections, and the measurements that price them.
What are the eleven EA line items?
Price level, renewal eligibility, baseline seat count, the renewal uplift, the M365 SKU mix, billing cadence under NCE, the Copilot cohort, security attach, Power Platform attach, the Azure MACC, and the eight terms protections: three are settled before you sit down.
And the other eight are yours to move.
Unbundling all eleven produced an average 20 to 30 percent run rate improvement against the first package.
How should an Azure MACC be sized?
To trailing twelve month consumption from the Cost Management page, never Microsoft's forward forecast: proposals sized on the forecast ran 22 to 38 percent above trailing consumption in our file.
The horizon matters as much as the size, with three and four year terms holding leverage at each renewal while six and seven year terms surrender it for the duration.
How does the EA true up work?
Growth reconciles once a year, and the ratchet runs one way: seats added mid term cannot come off until the renewal anniversary, the single reset point in the term, which makes the baseline seat count at signing worth more scrutiny than the discount percentage.
The reconciliation of the VLSC relationship summary against real headcount is the line item most renewals never check.
Is the Microsoft EA still worth it?
At scale, with stability and high Copilot attach, often yes, but as a justified choice rather than a default: the tier collapse compressed the program discount, the 2,400 floor removed smaller enrollments, and the crossover against CSP sits around 3,000 to 6,000 knowledge worker seats.
The renewals that improved 20 to 30 percent did it by splitting the eleven lines and arguing each on its own evidence, not by defending the vehicle.
