Plenty of small businesses still have a server in a cupboard. It runs the accounts software, holds the shared drive or handles logins. It’s usually been there longer than anyone remembers, and it gets attention only when it stops.
We looked back at a year of tickets, 19 August 2025 to 18 August 2026, with a server, virtual machine or storage device in the title. There were 64, from 19 clients, with 22.8 hours of work.
What the server tickets were
Most were about getting to the server rather than the server itself: people who couldn’t sign in to a remote server, got dropped from a session, or found a mapped drive missing. One business raised seven of those in the year.
Then there was the work that changes where things live.
- An education provider had servers decommissioned in November 2025, after a run of emails about its servers and an upgrade.
- A design business moved onto a new server. The tickets that followed were the ones you’d expect: error messages in its design software, setting up backup for the new server, a licence key for the backup software, and mapping drives on new PCs.
- A concrete business had a network storage device put in, a three-hour job in July 2026.
- Another business had its phone system moved to a new server.
Signs a server is near the end
The tickets don’t record how old each server was, so we can’t give you ages. The warning signs are consistent, though.
- Its operating system is out of support, or will be soon.
- The warranty has lapsed, so a failed part means hunting for spares.
- Remote logins are getting slower, or dropping people more often.
- Backups are taking longer or failing.
- Nobody is sure what’s still running on it.
Replace it, or move to the cloud?
For many small businesses, most of what the server does can move. Shared files fit in SharePoint. Email is often already in Microsoft 365. Logins can live in Microsoft’s cloud directory.
What usually keeps a server in the cupboard is one piece of software: accounts or practice software, design software that expects a local server, or a database a vendor only supports on site. That’s a question for the vendor before it’s a question for us.
Sometimes the right answer is a smaller, newer server, or a network storage device for large files that don’t suit cloud sync.
Planned versus failure-driven
A planned replacement happens on a day you chose. Data is copied and checked. Backups for the new server are running before the old one is switched off. Staff know what’s changing and when.
A failure-driven replacement happens on the day it dies. You restore from whatever backup you have, onto whatever hardware you can get, while nobody can work. In our experience it’s the more expensive way to do the same job, before you count the lost day.
Where to start
Find out three things about your server: how old it is, when its operating system support ends, and what’s still running on it. If you’d like us to check, get in touch.