A dairy plant has no pause button. Milk arrives, the line runs, shifts come and go around the clock. That one sentence shaped the migration of 1500 seats to Microsoft 365: there was no weekend in which you could “switch everything at once”.
The starting point
The group covers a dozen or so plants and offices that grew at their own pace over the years. Email and documents lived in local solutions, account standards differed between plants, and the common denominator was the absence of one place where you could see who has access to what. On top of that came very different digital skills across teams: one in the sales office, another on the production floor.
Why it was hard
Scale is only half the problem here. The other half is round-the-clock operation. In an office company you run a migration over the weekend and on Monday people find the new environment. In a plant that runs non-stop there is no weekend, and every hour someone cannot take an order or confirm a delivery is felt. The third factor was people: with 1500 of them, even a small confusion multiplies by 1500 and turns into an avalanche of tickets.
How we ran the rollout
We built the plan around each plant’s shift schedule rather than the delivery team’s calendar. We split the Microsoft 365 migration into a dozen or so waves and started with the least critical plant. That was a deliberate decision: the first wave was to prove the procedure on a smaller group before we touched the plants where a mistake would cost the most. Each wave used the lessons from the previous one, so the procedure kept getting shorter and more reliable.
In parallel we put identity in order in Entra ID and brought computers under management with Intune. At this number of seats and with natural turnover on production, it is the only way to answer the question of who has access and from which device. Multi-factor authentication went on all accounts.
We gave the most attention, though, to people rather than servers. At every wave there was a first-day on-call person: short instructions matched to how each team actually works, and a fast path for tickets. That way small problems closed within hours instead of piling up in a queue and growing into a sense that “the new email does not work”.
After the last wave we covered the whole environment with independent Microsoft 365 backup and daily recoverability checks.
What changed
1500 seats work in one environment, and no plant stopped production for the migration. Access is controlled centrally: it is clear who has an account, on which device they work and what happens when they leave the company. Instead of a dozen or so local worlds, the group has one environment that can be maintained and developed.
From the board’s perspective, what mattered most is what does not show in the statistics: a migration of this scale went through without a crisis, without stopping the line and without the week of paralysis everyone expected.