Every company past a certain age has one. The server, or the licensed platform, that predates everyone's tenure, runs something important, and appears in conversation only as "yeah, we should really do something about that". The reason nothing happens is that "it still works" feels free. It isn't, and it's worth pricing honestly.
What it actually costs
Start with the visible line: the annual licence or hosting renewal, which quietly rises a few percent a year precisely because the vendor knows leaving feels hard. Add the maintenance tax: the patching, the certificate renewals, the disk that fills up every few months, each one landing on your most senior engineer because nobody else dares touch it.
Then the two costs that never appear on an invoice. Risk: ageing systems concentrate outage and security exposure exactly where you have the least knowledge and the fewest escape routes; when it fails, it fails on a public holiday, and the one person who understands it is at the beach. Opportunity: every hour your best engineer spends keeping the old thing alive is an hour not spent on the product. At most small companies this is the largest cost of the four, and the least discussed.
Why teams stay stuck
Because migrations have a reputation for drama, and the reputation is earned; most migration horror stories come from big-bang cutovers done under deadline pressure. So the renewal gets signed for one more year, seven years running.
The boring way out
Good migrations are dull. Build the replacement alongside the old system, as code. Run both in parallel until the outputs match and everyone's bored of checking. Cut over in a low-traffic window with a rollback path you've rehearsed. Then decommission, and cancel the invoice, which I maintain is one of the genuine pleasures of this job.
I've done this with CI platforms, enterprise file transfer, analytics workloads and whole toolchains. Not one of those cutovers was noticed by a customer. That's the standard the parallel-run approach buys you.
A useful forcing question
At the next renewal, don't ask "can we afford to migrate?" Ask "would we choose this system today, at this price, with this risk?" If the answer is no, the renewal isn't a default, it's a decision, and it deserves a real alternative costed against it.
Got one of these? The Cloud Migration service is the boring way out, priced fixed. Tell me what's ageing.