The renewal is the only day the count can go down

Mid term true ups adjust the Enterprise Agreement in one direction: up. Every reduction the estate earns between signatures, leavers, role changes, dormant seats, retired workloads, waits for the renewal or gets carried for another three years. In our reviews, signed counts ran 8 to 17 percent above what active usage justified.

Prepared by Redress Compliance · August 15, 2026 · Microsoft advisory. EA renewal engagements and license position reviews, 2024 to 2026.

• Build the usage truth first: active usage per SKU per user over the trailing six months, dormant and duplicate accounts flagged, service level activity inside the bundles to test whether the tier fits the role.
• Harvest the dormant layer before the quote arrives, because the 10 to 20 percent of seats with no meaningful activity is the cheapest reduction in the whole negotiation: it costs nothing but reconciliation.
• Counter the anchor with a built count, not a discount ask. The quote starts from your commitment; the counter starts from your consumption, and the 15 to 30 percent gap between them is the negotiation.
• Cost the other doors even if you intend to stay: an MCA E and CSP scenario, priced and timelined, is what makes the EA number move, and with the EA retirement wave underway it may be the only door offered.
• Write the protections while the pen is out: price caps on the term, ramp schedules for growth you have not hired yet, and reduction rights at anniversaries where the program allows them.
• Start twelve to eighteen months out, because every one of the levers above is a quarter of work, and a renewal run on the vendor's timeline prices like one.

Every Enterprise Agreement contains a quiet asymmetry that most buyers discover only after living inside one. Growth is continuous and contraction is scheduled. The true up collects every addition the year produced, on time, every year; the reductions the same year produced, the leavers, the divested unit, the project that ended, the pilot that failed, collect nowhere. They wait. The contract has no mechanism for them until the term ends, and so the committed count becomes a high water mark of three years of headcount, worn like a uniform long after the tide went out.

This is why the renewal quote looks the way it does. Microsoft's proposed baseline is not an estimate of what you need; it is a restatement of what you committed, priced at the new rates. In more than half of the renewals we reviewed, that suggested floor sat 15 to 30 percent above measured consumption. Nothing about that is sharp practice. It is simply the incumbent's privilege: the seller anchors on the number the contract already contains, and the buyer who arrives without a number of their own negotiates against their own high water mark.

The estates that moved the outcome all did the same unglamorous thing: they replaced the committed count with a measured one before the first meeting. Active usage per user per SKU, the dormant tenth to fifth of the estate identified by name, the bundles opened to see whether the role ever touched the services that justify the tier. That document, not the discount conversation, is the negotiation. Once the counter count exists, the 12 to 25 percent movement we observed stops being a concession the vendor grants and becomes arithmetic the vendor can check.

Time is the other half of the outcome, and it is less about effort than about credibility. A buyer ninety days from expiry cannot re tier roles, cannot cost the MCA E or CSP doors, cannot survive a deadlock into a second quarter, and every party in the room knows it. A buyer twelve to eighteen months out can do all three, and the quote reflects that before anyone says it aloud. The renewal is the only day the meter runs backward; the preparation window is what decides whether it actually does.

What remains is sequencing. The true up keeps the mid term meter honest so the renewal starts from clean data; the renewal resets the count the true ups can only raise; and the wider structure, the program doors, the pressure points, the levers, lives in the EA pillar and the negotiation guide. Run the calendar, not the quote.

Watch the briefing · 4:21How to Prepare for Your Microsoft EA Renewal in 2027The EA retirement wave, the MCA E and CSP doors, the Multiple Equivalent Offers pattern, and the eighteen month clock.

Across the EA renewals and license position reviews we ran, the same two gaps decided the economics:

The pattern behind the numbers was consistent. The overhang was invisible internally because nobody owned the count between renewals. The dormant seats, 10 to 20 percent of the Microsoft 365 estate, surfaced only when someone finally pulled activity data. And the buyers who captured the movement were not better negotiators; they were earlier ones.

The buyer side move is to treat the renewal as a count rebuild with a price attached, not a price negotiation with a count attached. The wider library sits in the Microsoft practice.

When should we start preparing for a Microsoft EA renewal?

Twelve to eighteen months before expiry. The usage truth takes a quarter to build, the internal alignment on what to cut takes another, and the negotiation itself rewards a buyer who can credibly walk toward MCA E or CSP, which takes months of groundwork. A renewal started ninety days out inherits the incumbent count by default.

Can we reduce seat counts at an EA renewal?

Yes, and it is the only scheduled date you can. Mid term true ups adjust the count upward and never downward, so every reduction the estate has earned since signature, leavers, role changes, dormant seats, retired workloads, is collected at the renewal or carried for another three years.

Why does the renewal quote start above what we need?

Because the proposed floor is built from your committed count, not your consumption. In our renewal reviews Microsoft's suggested baseline overshot real active usage by 15 to 30 percent in more than half of cases. The quote is an opening position that assumes the count carries forward; the counter is a count built from usage data.

What data do we need before a renewal negotiation?

Active usage per SKU per user over the trailing six months, dormant and duplicate accounts, service usage inside bundles to test tier fit, and the license position from the last true up. In the estates we reviewed, 10 to 20 percent of committed Microsoft 365 seats showed no meaningful activity, and signed counts ran 8 to 17 percent above what usage justified.

What movement is realistic at an EA renewal?

Disciplined buyers, with usage evidence, a credible alternative path, and time, held or cut the renewal economics by 12 to 25 percent against the first quote: some of it count, some of it rate, some of it term protections. Buyers who renewed the incumbent count on the incumbent timeline held close to zero.

What happens if we simply renew the current count?

The overhang compounds. The 8 to 17 percent of seats that usage no longer justifies gets repriced at the new rates and committed for three more years, and every mid term true up stacks on top of it. Renewing the incumbent count is not a neutral act; it is the most expensive decision on the calendar, made by default.

Should we consider MCA E or CSP instead of renewing the EA?

Evaluate them regardless of intent. Microsoft is retiring the EA for a growing share of customers, so the renewal may arrive as a forced migration, and even where the EA remains available, a costed MCA E or CSP scenario is the leverage that makes the EA quote move. The door you would actually walk through is the one that disciplines the incumbent's pricing.