The Date Somebody Else Set: Why Migration Waits on Your Slowest Supplier
Windows 10 reaches end of support today, and a great many organisations have not finished. The operating system was never the hard part. The date an organisation can actually meet is set by whichever of its software suppliers is slowest.
Windows 10 reaches end of support today. The date has been published for years, and a great many organisations will not have finished.
That is rarely for want of trying. The operating system was almost never the hard part.
1. The Deadline Was Never Hidden
Support end dates are published a long way ahead — this one was known before most of the machines it affects were even bought.
Organisations were not complacent about it. Testing a large estate takes months, hardware baselines ruled out a proportion of existing devices, and the work competed with everything else the year had already committed to.
Even so, the shape is familiar — four years of notice, and four months of work.
2. The Operating System Was Never the Problem
Moving a desktop estate to a new version of Windows is well understood. The tooling exists, the vendor wants it to happen, and most of it can be automated.
The difficulty is in what runs on top of it. A planning tool, a control system interface, a piece of engineering software written against a version two releases back, maintained by a supplier with four other customers.
Each has to be tested, and a proportion will fail. Which ones is not knowable in advance — it is discovered, one application at a time, by somebody trying it.
3. Your Timetable Belongs to Your Suppliers’ Suppliers
When an application fails testing, the fix belongs to whoever wrote it. The organisation can raise it, escalate it, offer to pay for it — none of which moves the date the release lands.
The deadline was set by one supplier. The date that can be met is set by whichever of the others is slowest.
Nothing in the plan makes that visible. It shows up as a milestone that keeps moving, with no obvious reason attached to it.
4. Some Organisations Cannot Simply Stop
In most businesses there is an answer available: stop using the application until it is ready. Inconvenient, and survivable.
That option does not exist everywhere. An organisation running water treatment, a network or an emergency service cannot suspend a tool because a supplier is behind — the work does not pause and wait.
So the estate has to move and the application has to keep running, at the same time, on foundations that no longer fit together.
5. The Workaround Is a Legacy System You Chose
The practical answer is containment. The application carries on running on the old operating system inside an isolated virtual machine, fenced off from the network, tightly controlled and used for nothing else.
It works, and it is the right call. It is also an unsupported system the organisation has agreed to keep running — deliberately, with an accepted risk and no date on which it ends.
Most legacy is inherited. This kind is created on purpose, in daylight, by people doing the sensible thing.
6. Paying to Extend Is a Real Option
Buying extended support is a defensible answer. It costs money and buys time, and for an estate with real dependencies that is a reasonable trade — particularly where the alternative is rushing the testing.
The difficulty is that most organisations which bought it did not choose it. They ran out of runway and took the only thing still available.
One of those is a decision with a review date attached. The other is an invoice.
7. The Exceptions Are the Whole Problem
Containment only stays safe while somebody is watching it, and watching costs almost nothing if four things are written down and reviewed:
- which applications have been tested on the new version, and which have only been assumed to work
- which suppliers have given a date in writing, rather than an intention
- what each isolated exception costs to run, and who reviews it
- when each one is expected to end, and who hears about it if it does not
The last is the one that gets skipped, and it is why temporary containment becomes permanent.
8. The Calendar You Did Not Write
A share of every year’s change capacity is spoken for before the year begins. Certification expiry, support windows, hardware floors, browser requirements — none of it chosen, and all of it compulsory.
Most plans do not show that share. It arrives instead as an interruption, and displaces something the business actually wanted.
The organisations that handle this well have stopped being surprised by it. The enforced work sits in the plan as a standing line, and what remains is what there actually is to allocate.
Final Thought
Support deadlines are not really technology events. They are a recurring claim on capacity, made by suppliers, on dates the organisation had no hand in setting and cannot negotiate.
Treated as a surprise, that claim is paid for twice — once in the scramble, and again in the exceptions nobody remembered to close.