The Microsoft EA, decoded

Roger: The headline everyone missed this spring was not a price increase. It was that Microsoft quietly stopped offering the Enterprise Agreement to companies under about twenty four hundred seats. If you are mid sized, the vehicle you have negotiated against for twenty years is being taken away, and most buyers found out at renewal.

Bill: And the ones who found out at renewal found out with no leverage and no time.

Bella: Today we take the whole Microsoft machine apart, the new top tier suite, the July price increases, and the thing that actually decides your bill, which is never the discount. This is Off List. Read the paper before you sign it.

Bella: Welcome to Off List. I am Bella. Roger the consultant is here, Bill from the manufacturing side, Laura from retail. Today, the single biggest line on most software budgets and the one moving fastest right now: Microsoft. And this is a genuinely unusual year, because three things are happening at once. The agreement structure is changing underneath people. The packaging changed on the first of July, including a brand new top tier suite. And the AI pricing is being wired permanently into the base. We are going to take all three in turn. Bill, you renew a large agreement. Set the stakes.

Bill: The stakes are that this is the year the ground moved, and a lot of people are negotiating on last year's map. I have renewed these agreements for most of my career and I have never seen this many structural changes land in a single cycle. The mistake I am watching people make is treating this like a normal renewal, argue the discount, sign, move on. This year the discount is almost the least important thing on the table, and I mean that literally, we will get to why.

Laura: And the pace matters for people like me, because retail runs lean and moves fast, and Microsoft is very good at using your lack of time as a pricing lever. The less runway you have, the more the new structure just happens to you.

Bella: Let us start with the structural change, because it is the one people know least and it matters most. Roger, what actually happened to the Enterprise Agreement?

Roger: So, two moves, and they compound. Move one. Microsoft has been pulling the Enterprise Agreement away from the middle of the market. If you are below roughly twenty four hundred seats, at your renewal you are increasingly told the EA is simply not available to you anymore, and you are pushed toward one of the newer vehicles, either the partner led program, the Cloud Solution Provider channel, or the direct enterprise version of the Microsoft Customer Agreement. And the direct enterprise agreement carries a floor, you need to be committing something like half a million dollars a year to Microsoft to even sit on it. So a whole band of mid sized companies is being funneled off the agreement they know, onto vehicles with different discount mechanics, different terms, and frankly less negotiated protection.

Bill: And the reason that stings is that the EA, for all its faults, was a known quantity. You understood the price protection, the true up cadence, the level structure. Moving a mid market company to the partner channel or the customer agreement is not just a paperwork change. It changes who holds your price, how increases flow through, and what you can push on. People are treating it as an administrative migration and it is a commercial demotion.

Bella: Move two, Roger.

Roger: Move two is quieter and, if anything, worse for the large enterprise. Last August, Microsoft removed the volume price levels on the online services. For decades the EA had levels, Level A started around five hundred seats, Level B at twenty four hundred, Level C at six thousand, Level D at fifteen thousand, and the bigger you were the better your baseline unit price. That was the whole logic of volume licensing, buy more, pay less per unit. They deleted that for the cloud products. Everybody now pays the same baseline regardless of size.

Laura: Which sounds neutral until you realize who it hits.

Roger: It hits the big buyers, precisely. If you were a large enterprise sitting at Level C or D, you had a structural discount baked in just for your size, and it is gone. The practical effect for the largest organizations is on the order of a low double digit increase, something like thirteen percent, for the exact same licenses, before you negotiate anything, before any list price change. Your size used to be leverage. Microsoft just switched it off.

Bill: And here is the part that makes me genuinely angry as a practitioner. They removed the reward for scale in the same window they are raising list prices and pushing the expensive AI bundles. Three increases stacked, and each one is announced separately so no one adds them up. The volume level removal, the July list increases, and the AI attach. Add them together and a large estate can be looking at a very different number for functionally the same thing.

Bella: So the first job in a Microsoft negotiation this year is not the discount. It is figuring out which vehicle you are even going to be on, and whether the size advantage you used to have still exists.

Roger: Exactly. And if you are in that mid market band losing the EA, the single most valuable thing you can do is start the vehicle conversation a year early, model the partner channel against the customer agreement against staying on the EA if you can qualify, and treat the vehicle choice as the negotiation. Because once you are funneled onto the vehicle they prefer, the discount conversation happens inside a box they built.

Bella: Now the packaging, because this is the flashy news and it landed on the first of July. Roger, walk people through what actually changed in the lineup.

Roger: So on the first of July, Microsoft did two things at once. They raised prices on the existing suites, and they introduced a new top tier. Start with the increases, because they are concrete. The enterprise flagship, E5, went from fifty seven dollars a user a month to sixty. E3 went up in the mid single digits. And the frontline tier, F3, the license for your task and shop floor workers, went up twenty five percent, from eight dollars to ten. That frontline jump is the one nobody is talking about and it is the biggest percentage increase in the whole lineup, and for a company like Bill's, with thousands of frontline seats, twenty five percent on that line is real money.

Bill: It is enormous for us. We have far more frontline workers than knowledge workers, most manufacturers do, and everyone's attention is on the expensive knowledge worker licenses while the frontline line quietly went up a quarter. I flagged it internally and half my own team had not registered it, because eight to ten dollars sounds trivial per seat. Multiply it by your frontline population and it is a number that would get a project cancelled anywhere else in the business.

Bella: And the new tier. E7.

Roger: Right. This is the first genuinely new enterprise suite since E5 launched back in twenty fifteen. They are calling it the Frontier Suite. It went generally available on the first of May, and it lists at ninety nine dollars a user a month. And what is inside it is the tell. E7 is E5, plus the Microsoft 365 Copilot, plus something called Agent 365, plus the Entra identity suite, bundled into one number.

Laura: So it is the everything bundle. And the ninety nine dollar price is the interesting part. Do the separate math.

Roger: The separate math is exactly why it is clever. E5 is sixty now. Copilot is thirty. The Entra suite is around twelve. That is already a hundred and two, before you add Agent 365, which on its own is around fifteen. So the components add up to something like a hundred and fifteen, a hundred and seventeen dollars a seat, and the bundle is ninety nine. On paper it looks like a genuine discount, buy it together and save fifteen or twenty dollars a seat.

Bill: And here is where the old rule saves you, the one from every episode. The bundle is only a saving if you were going to buy all of the components, for all of the people. And you were not. Nobody needs Copilot on their entire population. Nobody needs Agent 365 across every frontline worker. The ninety nine dollar bundle is a discount on a basket most of your people should not be carrying. So the per seat comparison, ninety nine versus a hundred and seventeen, is the wrong comparison. The right comparison is ninety nine versus what that specific person actually needs, which for most of your estate is E3 or E5 and nothing else.

Roger: That is the whole trap in one sentence. They priced the bundle just below the sum of the parts, so that a lazy comparison makes it look like free money, and a real comparison, against actual need per role, makes it look like what it is, which is standardizing your most expensive possible configuration across people who will never touch two thirds of it.

Laura: And the Agent 365 piece deserves a flag on its own, because it is the newest and the least proven. Some of its security features were still in preview when the suite went generally available in May. The flagship agent feature was research preview only, not production ready. So part of what you are being asked to standardize and pay for in that ninety nine dollars is capability that, on the day it shipped, you could not fully use in production. Paying today for a roadmap is the oldest move there is, and it is dressed up here as the future of work.

Bella: So the honest read on E7. It is a real product, it will be right for a narrow population of your heaviest users, and it is priced to get standardized across everyone through a per seat comparison that hides the fact that most of your people do not need the contents.

Roger: And if you take one number out of this segment, take the frontline one. Twenty five percent on F3. Because while everyone argues about the ninety nine dollar AI bundle for the executives, the biggest increase in the whole lineup is happening on the cheapest, most numerous seat in the building, and it is going through unexamined.

Bella: Let us stay on the AI attach, because this is where the multi year money is. Laura, you are piloting this. What is the pressure actually like?

Laura: The pressure is well designed, I will give them that. Copilot is thirty dollars a user a month, on top of a qualifying base license. So for an E3 shop, adding it can nearly double the seat. For E5 at sixty, it is another thirty on top, that is ninety a seat, before you have touched E7. And the way it gets sold is through the pilot. You run a pilot with a few hundred enthusiastic users, the pilot goes well because the people in a pilot are self selected to love it, and then the expansion conversation is, let us just add it to the base for everyone, and here is a better rate if you commit it across the estate or move to E7.

Bill: And once it is in the committed base, it does not come out. That is the mechanic that matters. Under the agreement you true up, you only ever add, you cannot true down mid term. So an AI line you commit across ten thousand people in a moment of enthusiasm is a line you carry for the whole term whether people use it or not. My rule now is a hard one. Copilot is licensed to a named, measured group, never to the whole population, and never folded into the committed base to unlock a bundle rate. We measure actual usage every quarter and we resize at renewal, which is the only moment we can.

Bella: Is it delivering, though? Set the licensing aside.

Roger: For the right person on the right task, genuinely yes, and I am not a skeptic on the capability. The problem is never whether it helps a finance analyst who lives in spreadsheets. The problem is buying ten thousand seats because five hundred people love it, and locking it into a multi year base where the price never comes back down. Value for some, priced for all, on a term that outlives the enthusiasm. And E7 is the most elegant version of that trap yet, because it hides the thirty dollar Copilot line and the fifteen dollar agent line inside one round bundle number, so you stop seeing the individual AI costs at all.

Bella: Talk about the agents, because this is genuinely new and it is where Microsoft is steering everyone.

Roger: This is the part buyers need to think hard about, because the whole E7 pitch, the reason they call it the Frontier Suite, is that it is built for a workplace where AI agents work alongside people. Agent 365 is a control plane for managing those agents. And the strategic question underneath it, which nobody is asking in the room, is how the agents get licensed over time. Right now you are buying agent management as part of a per seat bundle. But an agent is not a seat. It does not map to a person. And Microsoft has already signaled that consumption based pricing may come for this later. So there is a real risk you standardize onto E7 for the agent story, and then the agent economics shift under you to a consumption meter you did not model.

Laura: Which is the credits problem from a future episode, arriving early. The moment the thing you are buying stops being a person and starts being a unit of work, the per seat mental model breaks, and the vendor gets to define the new unit. My position on agents right now is simple. I will pay for the management plane if I am actually running agents at scale, and I am not yet, most people are not yet. I am not going to prepay for an agent future by standardizing my whole estate onto the most expensive suite today, on the promise that it will matter in two years.

Bill: Buy the estate you have. Take options on the estate you might become. The agent workplace might be real and it might be huge. That is an argument for negotiating good option pricing on E7 for when you need it, not for moving everyone onto it now because the demo was impressive.

Bella: Bill said at the top that the discount is almost the least important thing on the table this year, and I want him to defend that, because it is the opposite of what most people believe. Bill.

Bill: I will defend it hard, because it is the single most useful idea in this whole episode. Everyone walks into a Microsoft renewal focused on the unit discount. What percentage off did we get. And I am telling you, after thirty years, the unit discount matters least. Three things matter more, and each one is worth more than the discount, and each one is something you control and Microsoft would rather you ignore.

Bella: Go through them.

Bill: One. Your baseline. The agreement lets you add through the year and true up annually, and the true up is at your price, but the counting is on you. If you have deployed sloppily, if projects ended and seats never got reclaimed, you true up on a bloated baseline and you bake that bloat into the committed base for the rest of the term. A clean baseline before the true up is worth more than several points of discount, because it is not a percentage off a big number, it is deleting the big number. We reconcile actual deployment against entitlement ourselves, every year, before we report a single seat.

Laura: We found close to a fifth of our seats were assigned to people who had not signed in for months. That is not a discount you negotiate, that is a cost you delete, and it is entirely on your side of the table.

Bill: Two. The Azure commitment. Large agreements bundle a cloud consumption commitment, and committing gives you a better rate if you actually spend it. But the pattern we see everywhere is over commitment to unlock a headline discount. In the estates we have looked at, the cloud commitment is routinely sized fifteen to thirty percent above what the company actually consumes. So you buy a discount on spend you never use, and at the end of the term you are inventing workloads to burn a commitment, which is the opposite of saving money.

Roger: It is arson with a spreadsheet, which is a phrase we cannot stop using because it keeps being true. Size the commitment to a real forecast owned by the person who runs the cloud, not to the number that unlocks the prettiest discount slide.

Bill: Three. The editions and the roles. We covered E7 and the bundles. The biggest pool of savings in almost every Microsoft estate is people carrying a more expensive license than their role needs. Premium suites standardized across populations who use a sliver of them. Right sizing editions by role, before the renewal, is worth more than any discount percentage you will negotiate in the room. And Microsoft will never suggest it, because their ideal customer standardizes up and never looks again.

Bella: So the reframe is, the discount is a percentage off a basket, and your leverage is the basket, not the percentage.

Roger: And this is why the vendor loves fighting on discount. A big percentage off a bloated, over configured, over committed estate costs you more than a small percentage off the estate you actually need. The percentage is their battlefield. The basket, the baseline, the commitment, and the roles, those are yours, and they are worth multiples of the discount. Fix the structure first. Then negotiate the price.

Laura: And structure work is unglamorous and internal and it is where all the money is, which describes this entire profession.

Bella: Timing. When does the work start this year, given everything is moving?

Bill: Twelve months out, and I would say that even more firmly this year than usual, because you have an extra job before the normal jobs. Before you can negotiate anything, you have to know which vehicle you will be on, whether your size advantage survived the volume level removal, and what the July increases did to your specific mix. That is a quarter of analysis before you even build a target. So twelve months out you start, and if you are in the mid market band losing the EA, honestly eighteen.

Roger: And the calendar leverage is real if you are ready to use it. Microsoft's fiscal year ends at the end of June, and the field carries genuine pressure to close in that window. Which is also, not coincidentally, right when the July packaging changes land, so there is a real moment each spring where their year end pressure and a packaging deadline collide, and a prepared buyer can use that. But only a prepared buyer. Timing leverage is worthless if you are not ready to sign or walk when the moment arrives.

Bella: The reset. The sentences.

Roger: Adapted for Microsoft this year. One. We have modeled our vehicle options and our role based license needs, and we are renewing to actual usage, not to our current entitlement or to a standardized top tier. That sentence tells them you know the base is padded and you will not be bundled up by default. Two. Here is our clean baseline, here is our right sized cloud commitment, here is our edition mix by role, and here is the target. You are handing them the analysis their own desk will need. Three. We are prepared to align our decision to your fiscal window if we get close, and prepared to extend and revisit if we do not. That last one says you control your own clock, and their June pressure is now working for you instead of against you.

Laura: And for everyone who is not a giant enterprise, because the mid market is where this year's structural change hits hardest. The same logic scales down. Model your vehicle before you are forced onto one. Watch the frontline increase, not just the executive bundle. Keep Copilot out of the committed base. Right size by role. You do not need fifty thousand seats to be the prepared account. You need to have read the changes and known your own usage, and that alone moves your opening number.

Bill: Be the account that read the paper. It is still the cheapest leverage there is. It just has more paper to read this year than usual.

Bella: Quick round to close. One sentence. The single most important thing a buyer should do before a Microsoft renewal this year. Bill.

Bill: Model your vehicle before they choose it for you. If you are under twenty four hundred seats, the EA may already be gone, and the vehicle decision is the whole negotiation.

Laura: Watch the frontline line. The twenty five percent increase on F3 is the biggest in the lineup and it is going through unexamined because everyone is staring at the ninety nine dollar AI bundle.

Roger: Fix the structure before the discount. Clean baseline, right sized cloud commitment, editions by role. A big discount on a bloated estate costs more than a small discount on the right one. The percentage is their battlefield.

Bella: And mine. Do not let E7 standardize your most expensive configuration across people who will never use two thirds of it. Buy the estate you have. Next week, the other giant, SAP, the twenty twenty seven deadline, and the licensing metric that gets padded by a third before anyone checks. This is Off List. Read the paper before you sign it.