The holiday shutdown is your best IT window of the year. Plan it now.

Every IT change in a running plant negotiates with production. The switch upgrade waits for a weekend, the ERP patch waits for month-end to pass, the network segmentation project waits indefinitely because there is never a week when the line can tolerate surprises. Then, once a year, most plants hand IT the thing it can never otherwise have: days of scheduled darkness. The holiday shutdown, Christmas through New Year for many, summer changeover weeks for others, is when maintenance rebuilds machines, and it is the one window when the network, the servers and the systems that feed production can be worked on without costing a single unit of output.

Plants that use the window well treat it like maintenance treats it: scoped in October, parts ordered in November, executed to a plan in December. Plants that waste it decide in mid-December that “IT should do something during shutdown,” discover lead times, and spend the week doing nothing more ambitious than Windows updates.

This post is the October planning nudge. Here is what fits in a shutdown window, what does not and how to sequence the next 10 weeks.

What a shutdown window is uniquely good for

The best candidates are changes that are low-risk in themselves but intolerable to schedule against production:

• Network segmentation cutover. The office-floor separation covered in OT cybersecurity for manufacturers is mostly design and staging work that happens in advance; the risky moment is the cutover, when VLANs change and firewall rules go live. During shutdown, a misconfigured rule is a finding, not a line-down event
• The plant’s Windows layer. The HMIs and machine PCs that never reboot during production, the legacy boxes that get handled with gloves, can be imaged, patched where the vendor permits and swapped where they cannot. This is the week the 2012 PC beside the press finally gets replaced without a fight
• Server and infrastructure moves. ERP server migrations, storage swaps, UPS replacements, switch-stack upgrades in the plant IDFs. Anything with a “and then we power everything down” step
• Restore tests at full scale. Not the file-level spot check, the real one: restore the ERP to clean hardware and prove the plant could come back from zero. Shutdown is the only time the test can run against production-grade stakes without production-grade risk, and the evidence it produces carries insurance renewals and customer audits for a year
• Credential and access cleanup. Rotating the service accounts and shared passwords that touch production systems, the ones nobody rotates midyear because something might stop. Pair it with the vendor access review: shutdown is a natural moment to disable every vendor connection and re-enable only the ones that still have a reason to exist

What does not belong in the window

The shutdown is a cutover window, not a project window. Work that starts during shutdown fails during shutdown. The ERP upgrade, the new EDR platform, the segmentation design, all of that is built, tested and staged in November so that shutdown week is execution against a checklist. Two other honest exclusions: anything whose rollback takes longer than the window, and anything dependent on a vendor who is also closed that week. Confirm vendor holiday coverage in writing; the OEM’s support line matters precisely when their customer base is all doing the same work.

The 10-week countdown from here

• This month: scope and decide. Walk the deferred-work list with maintenance and pick what the window gets. Rank by “impossible any other week,” not by size. Agree the week’s single most important outcome, because weather, parts and surprises will force cuts
• Early November: order and schedule. Hardware lead times still bite; what is not ordered by mid-November is not arriving by Christmas. Book the vendor techs and your IT provider’s shutdown coverage now, since everyone’s calendar fills with the same week
• Late November to mid-December: stage and rehearse. Configurations built, images prepared, firewall rules written and peer-reviewed, rollback steps documented per task. The checklist gets a dry run on paper with maintenance in the room
• Shutdown week: execute, verify, document. Work the list, test each system as it returns, and write down what the environment now looks like while it is fresh. The updated network diagram from this week is next year’s audit answer
• Restart day minus 1: the proving run. Bring systems up in dependency order, ERP, MES, label printing, EDI, the same order recovery would use (the sequencing logic from ransomware downtime applies verbatim), and prove order-to-ship on paper before the first shift clocks in. The goal is a restart morning that is boring

The multi-plant note

Operators with several facilities should resist synchronizing everything. Stagger the deep work across sites or across years, keep one site’s staff reachable while another cuts over, and reuse the staged configurations so plant 2’s shutdown costs half of plant 1’s. Sponsor-backed operators will recognize the pattern: a documented shutdown playbook is exactly the kind of repeatable artifact that makes a platform’s next add-on integration cheaper.

The window is coming whether or not it gets used. Ten weeks is enough to use it well. Atticus Rowan plans and staffs shutdown-week IT work for manufacturers, including the restore tests and cutover coverage, and October is when those plans get real. If your plant goes dark at the end of December, start a conversation this month.