Microsoft's standard price hold protects only the SKUs already on your order form, which is why buyers who accepted it still absorbed 15 to 23% effective increases at renewal
The default construction locks price for purchased products for the remaining term and nothing else: not renewal-term rates, not add-on seats at new list, not true-up rate, not the extension years. Four specific drafting gaps carried the July 2026 list move (5 to 33% across the M365 catalog) and the November 2025 removal of EA volume tiers (4 to 18%) straight through to the invoice. The clause you table has to name the base, the term, the additions, and the extension, or it does none of the work you think it does.
Prepared by Redress Compliance · August 22, 2026 · Microsoft advisory. EA and MCA-E renewal engagements, 2024 to 2026.
Executive summary
The phrase "pricing for products you have already purchased is locked for the remainder of your term" is the entire standard protection, and it left a 25,000-seat E5 estate paying roughly $18M against a $15M prior baseline.
The $57 to $60 move alone added $900,000 a year, and removal of the volume discount tiers added roughly another $2M, neither of which the purchased-products lock touches at renewal.
Loophole one is the SKU boundary: anything not already on an order form is priced at the current price list, so a clause written around E3 and E5 leaves a catalog that moved 5 to 33% completely exposed.
The crude field workaround is a token order of a single license per anticipated service, but the durable fix is clause language that extends the hold to the full published price list as of the effective date, not to an enumerated SKU schedule.
Loophole two is the extension term: negotiated concessions routinely attach only to the initial 36 months and lapse on a 24-month extension, so buyers who extend to buy time re-enter renewal with the discount stack gone.
Strong wording binds price and discount level to the initial term plus any extension exercised under the agreement, with the anniversary rate carried at zero uplift.
Loophole three is scope: with EA volume levels removed on 1 November 2025 and every discount now individually negotiated, a hold on list price protects nothing if the discount percentage is not held alongside it.
The clause must fix net unit price in dollars per user per month, not a percentage off a list Microsoft controls, and it must state that percentage and dollar floors both survive.
Loophole four is new-SKU displacement: Copilot at a flat $30, E7 at $99 from 1 May 2026, and four new bundles on 1 July 2026 all arrived mid-term and outside any per-seat hold, alongside a consumption model formalized separately from seat licenses.
Ask for successor-product and substitution pricing at the held per-user rate for equivalent functionality, or accept that the hold expires the moment Microsoft renames the bundle.
What Microsoft's standard price hold actually does, clause by clause
A price hold has four operative components, and Microsoft's default paper covers exactly one of them. Component one is the base: what the frozen number attaches to, a specific set of SKUs on the order form or the whole published price list.
Component two is the term: how long the freeze runs, and critically whether it survives into the renewal negotiation or dies at the moment you need it most. Component three is additions: what happens when you buy seats or new services after signature, including the true-up rate.
Component four is extension: whether the held rates carry through any extension you exercise under the agreement. Microsoft's standard construction covers the base, narrowly defined as purchased products only, for the remaining term. That is it.
Everything else is left to the current price list in effect on the day you order, which is Microsoft's list, revised on Microsoft's schedule, which moved 5 to 33% across the M365 catalog on 1 July 2026 alone.
The purchased-products boundary is the one buyers underestimate.
If a service is not on your order form, it is not protected, and the fix Microsoft's own channel documents is a token order: buy a single license of anything you might deploy later, and the lock extends to future purchases of that service for the remainder of the term.
That is a real workaround and worth executing before signature, but it is a patch on a gap that should be closed in the clause.
The second gap is that Microsoft's ordering policy distinguishes price-locked SKUs from unlocked ones, and add-on products ordered mid-term are explicitly instructed to be priced from the current list.
The third is the true-up: reconciliation happens at the anniversary, and unless the unit price is bound to the held rate, the true-up prices at whatever list is live at that moment. The fourth is timing.
The hold expires exactly into the renewal negotiation, which is the single window where you have the least alternative and Microsoft knows it. Our EA and MCA-E redline playbook treats this as the highest-value clause on the paper for exactly that reason.
| Mechanic | Microsoft's default position | Exposure | Redline that closes it |
|---|---|---|---|
| Base (which SKUs) | Products on the order form only | Any new service prices at current list; July 2026 moves ran 5 to 33% | Held Rate applies to the entire published price list in effect on the effective date |
| Term (how long) | Remaining term of the current agreement | Hold expires into the renewal, where you have zero alternative | Held Rate governs renewal-term unit pricing for a stated number of years |
| Additions (new seats) | "Order anytime using the current price list" | 25,000-seat E5 estate: $57 to $60 adds $900k per year on the increment alone | Add-on seats of held products price at the Held Rate for the full term |
| True-up rate | Reconciled at anniversary list price | Growth is repriced annually against a moving list | True-up unit price equals the anniversary Held Rate, no reconciliation uplift |
| Extension years | Silent; concessions treated as expiring with the base term | Extension becomes a repricing event disguised as continuity | All pricing and discount concessions survive any extension exercised |
| Special discounts | Routed through CPS approval, discretionary and per-transaction | Concession does not attach to the contract and can be withdrawn | Discount stated as a net unit price in the agreement, not a CPS reference |
The table's real message is that Microsoft's default hold protects the position you are already in and none of the positions you will move into.
Every growth event, every new service, every anniversary reconciliation.
And every extension is a repricing opportunity, which is why the removal of EA volume tiers on 1 November 2025 (worth 4 to 18% off list at Levels A through D) mattered so much: it stripped out the one protection that applied automatically and left the negotiated clause as the only durable instrument.
Read the two 2026 changes together and the exposure is not theoretical. The tier removal alone moved a large estate 4 to 18%; the July list move added another 5 to 33% depending on SKU mix; combined effective increases land at 15 to 23%, and past 23% with modest Copilot adoption.
Buyers who accepted the standard hold absorbed both, because the hold protected the E3 and E5 lines on the order form while the growth, the new SKUs, the true-ups, and the extension all priced off the moving list.
The exact language to table: four drafted clauses
Table these as a block, not as isolated asks. Microsoft's contract team will attack the defined terms first, so define them yourself. Clause 1, Held Rate and base.
"Held Rate means the net unit price in United States dollars per user per month set out in Exhibit A, applicable to each product and service on Microsoft's published enterprise price list in effect on the Effective Date.
And not limited to the products listed on any order form." Fix the number in dollars per user per month.
Do not accept a percentage off list: a 20% discount against a list that moves 8.3% is a price increase, and the entire July 2026 catalog move ran through customers who had percentage-based protection. Clause 2, renewal-term pricing.
"The Held Rate shall govern unit pricing for the initial term and for the first renewal term of this Agreement.
And Customer may elect renewal at the Held Rate by written notice not less than sixty days prior to expiration." Without a renewal-term sentence the hold expires at the exact moment Microsoft has maximum leverage, which is the structural point our note on renewal timing and leverage develops.
Clause 3, additions and true-up. "Additional licenses of any product or service subject to the Held Rate, whether ordered as an addition, a true-up, or a new enrollment thereunder, shall be priced at the Held Rate.
No reconciliation, anniversary, or annual order shall be priced at Microsoft's then-current price list." That single sentence kills the "order anytime using the current price list" instruction and the anniversary reconciliation uplift in one move. Clause 4, extension survival.
"All pricing, discounts, and commercial concessions granted under this Agreement, including the Held Rate, shall survive and apply in full to any extension of the term exercised by either party." Microsoft's redline will try to convert Held Rate into a SKU schedule.
Add "for products purchased under this Agreement," insert "subject to Microsoft's then-current pricing policies," and route the discount through a CPS reference rather than the contract body.
Reject all four: a CPS-referenced discount is a per-transaction approval, not a contractual right, and it disappears when the approver does.
Move from EA to MCA without losing your terms
How to move from a Microsoft EA to the MCA without losing discounts or terms: the transition traps, the price protections to keep, and the timing.
Get the white paper →Why the tier removal made the price hold, not the discount, the durable instrument
Before November 2025, the price hold was housekeeping. Volume tiers did the real work: Levels A through D awarded roughly 4% at 5,000 seats and up to 18% at 50,000-plus, and they did it programmatically. You did not negotiate for them, you qualified for them.
That distinction mattered more than anyone treated it at the time, because a programmatic entitlement renews itself. If your seat count held, your discount held. The contract did not need to carry the number forward; the price list did.
A buyer who walked away with a weak price-hold clause in 2022 still landed roughly where they expected in 2025, because the tier table was doing the durability work in the background.
That mechanism is gone. With tiers removed for new agreements and renewals, every point of discount is now an individually negotiated concession attached to a specific order form for a specific term. Nothing in the program structure carries it forward.
The percentage you fight for in the renewal meeting is, absent drafted language, a one-time event that expires the day the term does.
This is the single most important shift in Microsoft enterprise commercial posture in a decade, and it is almost entirely invisible in the renewal proposal, because the proposal shows you a discount and says nothing about what happens to it in 36 months.
Forward guidance makes the asymmetry worse. Level A (500 to 2,399 users) is projected to compress from a historical 8% reduction to a 5% ceiling in 2026 agreements, with the B and C brackets constricting on the same pattern.
Read that alongside the July 2026 list move of 5 to 33% across the M365 catalog and the picture is unambiguous: the discount you negotiate this cycle is the high-water mark, not the baseline.
Anything that is not contractually fixed reverts to a list price that is rising and a discount ceiling that is falling. The combined effect (15 to 23% effective increase, past 23% with modest Copilot adoption) is what happens when both variables move against you and neither is locked.
So the negotiating capital allocation most buyers use is now backwards. Teams spend the entire cycle pushing the headline percentage from 14 to 17 and treat the price-hold language as boilerplate to be accepted at signature. That trade buys a single-year win.
Three points of additional discount on a 25,000-seat E5 estate is real money, but a hold that fails to cover renewal-term pricing, add-on seats, or the extension years surrenders far more than three points when the term rolls.
On the same worked example, a $57 to $60 list move alone adds $900,000 a year, and discount removal adds roughly $2 million. No headline percentage recovers that if the clause does not reach it.
Microsoft's stated rationale (alignment with MCA-E, which has no volume tiers) is a leverage narrative, not an accounting constraint. Nothing prevents a global price list from coexisting with customer-specific contractual pricing; MCA-E agreements carry negotiated commercial terms all the time.
What tier removal actually accomplished is the transfer of pricing durability from the program to the paper, and paper is where Microsoft controls the drafting. The company did not raise prices by 4 to 18% in a single stroke.
It removed the mechanism that made your discount self-renewing, then invited you to renegotiate the number annually against a rising list.
The practical conclusion for anyone sitting in this negotiation: the price-hold clause is now the buyer's substitute for the tier table. It is the only instrument that converts a negotiated number into a durable one.
Treat it with the same seriousness you once gave the seat-count threshold that qualified you for Level C, and price your concessions accordingly.
If you are still building the commercial case, the analysis of the price level you sit at versus the percentage off list is the right frame; the hold clause is what makes that level survive the renewal.
How Microsoft's negotiators will push back, and the counter
Expect four objections, in roughly this order, and expect the first one to be delivered as though the conversation is already over. None of them is a genuine constraint. Each is a test of whether you know what the standard construction actually covers.
| Microsoft's objection | What it really means | Your counter |
|---|---|---|
| "Price protection is standard and already in your terms" | The default covers purchased products for the remaining term only: no renewal rates, no add-ons at held pricing, no extension | Read the clause back verbatim, name the four gaps, and table your drafted amendment as the redline, not a request |
| "We cannot commit to renewal-term pricing, the price list is global" | The list is global; your contractual pricing is not, and MCA-E deals carry customer-specific terms routinely | Ask for held pricing as a customer-specific commercial term, not a list change. Offer 5-year term length or a Copilot seat commitment as the paid trade |
| "New SKUs are new products, not covered" | True under standard paper, which is why E7 at $99 and the four July 2026 Copilot SKUs land at full list mid-term | Draft the hold against a defined discount-off-list floor for the product family, not a SKU list. Place token orders on likely future services now |
| "The extension is a separate commercial event" | The extension years are where held pricing quietly lapses and reprices | Bind the hold to the term "including any extension or renewal exercised under this agreement," and place it in the amendment, never the SOW |
The placement point in the fourth row is the one that decides outcomes. Language that lives in a statement of work or an order form travels with that document and dies with it; language in the enrollment amendment survives the order cycle.
Microsoft's desk knows this and will offer to "capture the intent" in the SOW as a concession. That is not a concession, it is a scope reduction.
When the desk refuses renewal-term language outright, stop negotiating with the account manager. The account team is measured on this quarter's number and has no authority to price a commitment that lands three years out.
Escalate to the licensing executive or the regional commercial lead, put a dated decision point on it, and make clear the term length and any Copilot commitment are conditional on the clause landing.
Buyers who run the escalation early, with the drafted language already in hand, get renewal-term coverage far more often than those who raise it at signature. The broader clause set is covered in the EA and MCA-E redline playbook; the hold is the one that pays for the rest.
What a strong outcome looks like in numbers
Set three tiers before you draft, because Microsoft's team will negotiate against whichever one you name first. Acceptable is a held net unit price (not a held discount percentage) on every SKU on the order form.
Extended to add-on seats of those same SKUs for the full term including any extension years, with renewal-term uplift capped as a separate clause rather than folded into the hold.
That is the floor, and it is roughly what a well-run team gets without a fight. Good extends the held rate to the entire price list as of the effective date, not just what you bought, and fixes the true-up rate at the anniversary Held Rate rather than the then-current list.
That single word change matters: on a 25,000-seat E5 estate the $57 to $60 move alone is worth about $900,000 a year.
And a true-up priced at list rather than held rate quietly re-imports that increase on every seat you add. Best case adds successor-SKU pricing at held per-user equivalence (so an E5-to-E7 or Copilot-bundle repackaging lands at your rate, not the $99 or $32 sticker) plus a 0% extension uplift.
Against the full stack (tier removal at 4 to 18%, the July 2026 list move at 5 to 33%, and MCA-E migration), the same 25,000-seat estate is defending close to $3 million a year, roughly the $15M to $18M gap in the modeled case. Price the clause against that number, not against the headline 5.3%.
Evidence base and the patterns that repeat
Tier removal (4 to 18%) stacked on the July 2026 list move landed here even for buyers who thought pricing was locked.
Clauses drafted around E3 and E5 numbers left the rest of the estate fully exposed.
The recurring failure is not that buyers skipped the price hold. It is that they read it as covering the agreement when it covered the order form.
The most common post-signature discovery across renewal engagements is that the concession applied only to the initial 36 months, so the extension years reprice at whatever list has become, which after November 2025 and July 2026 is a materially different list.
The second pattern is the token order: buyers who placed a single-license order for services they expected to deploy later locked today's rate for those services across the remaining term, while buyers who waited paid the new price.
That is a two-minute administrative action worth six figures, and Microsoft's own guidance describes it, which tells you the gap is intentional rather than accidental.
Third, flat-priced items defeat tier-based thinking entirely: Copilot has been $30 per user per month regardless of customer size since launch, so a hold anchored to your volume level protects nothing on it, and the agent consumption model sits outside per-seat pricing altogether.
Fourth, new SKUs shipped mid-term (four Copilot variants on 1 July 2026, E7 at $99) are not covered by any hold on what you already own, which is exactly how a repackaging becomes a price increase.
Companion pieces in this cluster cover uplift caps, swap and substitution rights, and the MCA-E non-negotiables.
Read them alongside the broader Microsoft EA and MCA-E redline playbook and the sequence that actually moves price, because a clause tabled at the wrong point in the cycle gets edited rather than accepted.
- 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
Your first five moves
- Pull the executed agreement and find the price-protection sentence itself, not the summary slide: read where the term boundary sits, whether it says "products purchased" or "products ordered under this agreement," and diary the exact date protection lapses, because that date is the only hard deadline in the negotiation.
- Build a SKU exposure list that includes what you have not bought yet, covering anything the estate might plausibly deploy in the next 24 months (Copilot at $30, E7 at $99, Entra Suite, agent consumption), then place a token order (a single seat is documented as sufficient) on each candidate service before renewal so the lock reaches forward purchases rather than only the current install base.
- Table all four drafted clauses in the first paper exchange, not at signature: language introduced in week one is a negotiation position, language introduced in the final week is treated as a schedule change requiring desk approval you will not get, and the sequence that actually moves Microsoft's price depends on redlines being live before the quarter-end pressure builds.
- Price the extension scenario in writing before you need it, asking for the co-term and one-year extension rate in the same paper as the base deal, because with EA volume tiers gone since November 2025 (a 4 to 18% move on its own) an unpriced extension reprices at whatever list does in July.
- Name the escalation path and the date now, identifying the account executive, the licensing desk owner, and the regional escalation contact, and stating that unresolved clauses go to escalation on a fixed calendar date, which is the single most reliable way to stop the desk from stalling redlines into your renewal window.
Frequently asked questions
Does a Microsoft price hold cover new seats added mid-term?
Only for products already on an order form. Microsoft's ordering guidance directs that additional product SKUs are ordered at the current price list, so a service you never bought is priced at today's rate, not your held rate.
The fix is clause language extending the hold to the entire published price list in effect on the effective date, not to an enumerated SKU list.
Can I lock renewal-term pricing, or only the current term?
Renewal-term pricing is negotiable but is not in the default construction, which expires exactly into the renewal conversation. Buyers who get it typically trade term length or a named workload commitment for it.
Expect the first refusal to come from the desk rather than from policy, so escalate rather than accept the standard answer.
What happens to my discounts if I extend the EA by two years?
In many agreements they lapse. Negotiated concessions and exceptions frequently attach to the initial 36-month term only, so a 24-month extension is priced at then-current rates and you arrive at the next renewal with the prior discount stack gone.
Draft survival language naming any extension exercised under the agreement.
Why did prices rise even though I had price protection?
Two separate events. Microsoft removed EA volume discount tiers A to D on 1 November 2025, worth roughly 4 to 18%, and raised M365 list prices 5 to 33% across the catalog on 1 July 2026.
A purchased-products hold addressed neither at renewal, which is how the combined 15 to 23% effective increase reached the invoice.
Should I hold a percentage off list or a dollar unit price?
Hold the dollar unit price per user per month, and hold the percentage as a floor alongside it. Microsoft controls the list, so a percentage-only hold moves with every list increase. Fixing net USD converts the concession into something that does not erode.
Does a price hold cover Microsoft 365 Copilot or agent consumption?
Generally not. Copilot has been priced at a flat $30 per user per month regardless of customer size, and the agent consumption model is formalized separately from seat licenses, so per-seat holds reach neither.
If Copilot matters to your roadmap, price and hold it explicitly rather than assuming coverage.
Does a token single-license order really extend the price lock?
It is a documented field workaround: placing a small order, even one license, can bring a service inside the purchased-products lock for the remainder of the term. It is tactical rather than durable, and it works only if you can predict deployments 24 months out.
Prefer clause language that removes the need for it.
