Three lessons for an organisation of this size

The calendar is not its own. Office 2021 and Office LTSC 2021 lose support on 13 October 2026 [19][20][21], the pricing grid of the Microsoft 365 suites changed on 1 July 2026 [13][14], EWS stops in Exchange Online on 1 April 2027 [17] and switching charges disappear on 12 January 2027 [3]. These dates fall whatever the path; the organisation aligns them with its anniversary date.

The contract has opened, and its texts are calibrated for this size. The commitments of 12 September 2025 open suites without Teams, a right to switch at renewal and data portability [1][2]; the free Teams export tool covers tenants of up to 500 users [27].

The field has never been documented at this size. The public cases are public administrations of 3,000 to 44,000 workstations, and all of them stumbled on applications and usage, never on the licence [5]. The dossier replaces this missing field experience with seven checks to run in-house: that is where the decision is made, and not in the prices.

The dated triggers

The Italian investigation and the Swiss programme are signals, not obligations to leave [38][39][5].

Three paths, what makes each one rational and what breaks it

What the sources establish

Two findings emerge from the sources. The first holds with a reservation: the licence is no longer what keeps an organisation with Microsoft. The commitments of 12 September 2025 [1][2], the Data Act [3] and the end-of-subscription rules [4] make it possible to change plan, take one's data away and leave on a fixed date. Two remnants of lock-in remain: early-termination penalties [11], and the F1 and F3 offers, which the commitments do not cover [2]. The second finding is established: what holds an organisation back is its applications and its usage. Bern runs a solution in parallel without deciding to leave [5][6], Zurich notes dependencies on its line-of-business applications [7], and three German public administrations have returned to Microsoft for reasons of usage [8][9][10].

These cases are public administrations of several thousand workstations. For a private SME of 50 to 500 seats, the transposition is an inference [inferred]: fewer line-of-business applications and old macros, but also fewer means to run two environments in parallel, and no operator of the Dataport kind. The lock-in changes in nature in the same way; its weight cannot be measured with the sources gathered.

One fact remains established: no public document gathered in this dossier describes an organisation of 50 to 500 seats that has left Microsoft 365, with duration, cost and application scope recorded. The NIS2 obligations that frame this choice are dealt with in chapter 7 [44].

What is found only in the organisation's documents

• What proportion of the active Excel files contains VBA code, and is that code compatible with the alternatives under consideration?
• Does the history of the shared mailboxes have evidential value, and would it be retained in a change of email system?
• Does the Teams history have sufficient evidential value, and do the alternative tools offer equivalent search and retention?
• To what extent do Intune, Conditional Access and Power Platform, alone or together, create a technical lock-in that goes beyond the licence?
• Does the organisation fall within a NIS2 entity class, and what obligations then weigh on the choice of its digital services provider?
• Does the licensing channel change the possibility of terminating, migrating or changing plan, and does the anniversary date lock the calendar?
• Can the existing perpetual Office licences be kept and reassigned during the transition, and on what terms?

These seven answers are the part of the decision that the dossier cannot supply: they lie in the organisation's documents, and they say whether the full exit is open, what goes beyond the licence, what NIS2 lets no one transfer, and what the real calendar of the contract is. The dossier does not rank them, because their weight depends on each estate.