A practical guide to why Microsoft 365 costs rise, the levers that bring them down, and how to realize the savings, written for Switzerland and Europe.
Microsoft 365 cost optimization is the process of reducing what you pay for Microsoft 365 across usage, rate, licensing and contract, without losing capability. It combines reclaiming unused or over-tiered licences with negotiating the right plan and terms at renewal, typically cutting spend by around 30%.
Updated
List prices climb year over year, and the 2026 round raised some Microsoft 365 plans by up to 43%.
New bundles (E7, Copilot, Agent 365) and forced SKU moves push customers toward higher tiers.
Inactive users, double-licensed accounts and E5 seats using only E3 features quietly inflate spend.
Without a plan before the renewal, you renew last year's estate at next year's prices.
Every Microsoft 365 saving comes from one of these four levers.
Right-size to what teams actually use and reclaim inactive seats.
Best unit price via commitments, reservations and savings plans.
Match plan and tier to real demand, with no feature waste.
Optimize terms, discounts and timing at signing and renewal.
A finding is only money once it can be executed. The third column is what decides whether a saving lands this year or next.
| Lever | What to inspect | When it can take effect |
|---|---|---|
| Usage | Sign-in and service activity per user, licences assigned to disabled or departed accounts, shared and resource mailboxes carrying a paid licence. | Any time. Seat reductions still wait for the cancellation window. |
| Licensing | Users on a tier whose distinguishing features they never touch, double-licensed accounts, users missing a required base licence, add-ons that duplicate something a suite already includes. | At renewal, or at the next term boundary for the affected subscription. |
| Rate | Term length and commitment mix, Azure reservations and savings plans against actual run-rate, currency and billing terms. | At purchase and at renewal. A commitment locks the rate for its term. |
| Contract | Programme fit, discount structure, co-terming of subscriptions that renew on different dates, and the timing of the negotiation itself. | Only at signing or renewal. Nothing here moves mid-term. |
Two dates constrain everything above. Under the New Commerce Experience a seat count can only be reduced within seven days of a purchase, a renewal or a mid-term addition, so quantity is fixed for the rest of the term. And Microsoft raised Microsoft 365 list prices on 1 July 2026 for new and renewing customers alike, which means an estate renewed without a clean-up renews last year's quantity at this year's price.
A baseline is three numbers, not one. Most estates can produce only the first, which is why the discussion stalls at the discount instead of the quantity.
What the contract obliges you to pay for. Comes off the invoice and the subscription list.
What is actually attached to a user object. The gap to purchased is pure waste and needs no debate.
What the assigned user demonstrably does with it. The gap to assigned is where tier decisions live, and where the argument happens, so the evidence has to be per user and per service.
The order matters. Purchased minus assigned is waste nobody defends, so it gets removed first and funds the rest of the work. Assigned minus used is where tier changes live, and every one of those needs per-user, per-service evidence before anyone will sign it off. Optimizing the rate before the quantity is settled just locks in a commitment for licences you were about to remove.
The most common Microsoft 365 savings a licence analysis surfaces:
The renewal is the only moment when the contract and the licensing levers are both open. The work is scheduled backwards from it, not started when someone notices the invoice.
Connect the tenant and the billing data, and produce purchased, assigned and used per user and per service. Nothing after this step is arguable without it, and gathering it late is what forces teams to renew on last year's numbers.
Reclaim licences on disabled and departed accounts, resolve double licensing, fix mislicensed shared mailboxes. None of this needs a negotiation or a business decision, and it lowers the quantity the renewal is priced on.
Take the downgrade and consolidation candidates to the business with per-user evidence. This is the step that needs time, because it needs agreement, and a tier change proposed in the final weeks gets declined for lack of confidence rather than lack of merit.
Programme fit, term mix, co-terming and discount, all against the estate as it will be, not as it was. A discount on licences you are about to remove is not a saving.
Under the New Commerce Experience the seat count can be reduced within seven days of the renewal, and not again until the next one. Anything missed here is carried for a full term.
Joiners, movers and leavers reintroduce waste continuously, and Microsoft keeps changing bundles and prices. Without a recurring cycle the estate drifts back to where it started, and the next renewal begins from scratch.
A finding becomes a saving only once someone owns the decision and the result is measured. Every recommendation runs through four stages.
Recorded with a reason. A recommendation deliberately set aside is a decision, and writing down why stops it being re-proposed every quarter and re-argued from zero.
The preparation: who is affected, what has to be communicated, which term boundary it waits for. Most findings fail here rather than on merit.
The formal sign-off, from the person who carries the consequence. Without a named approver a licence change becomes an audit finding later.
The measured saving, not the projected one. This is the number that survives being challenged by finance, and the only one worth reporting.
The reason this matters for cost, and not only for tidiness: an unmeasured saving gets contested at the next budget round and quietly reversed, and a licence removed without a named approver turns into an audit finding. A number that survives being challenged is worth more than a larger number that does not.
Demand-based procurement timed to your renewal, on average ~30% savings.
Explore→Continuous management so savings hold across the year, not just at renewal.
Explore→Audit-ready licence intelligence that finds these levers automatically.
Explore→Flexible licensing with optimization built in, all from one partner.
Explore→Let's find your levers together.