The 30-Day Phone Swap: A Week-by-Week VoIP Plan for Small UK Teams

There is a particular kind of dread reserved for the phrase we are switching the phones over on Friday. For a small UK business, the phone system is the one piece of technology that customers, suppliers and that one important client all touch directly. Getting it wrong is loud, public and expensive.

The good news is that a VoIP migration is a well-trodden path. The bad news is that most of the stress comes from squeezing it into a weekend. Give it a month, work through it in stages, and keep a parachute, and the whole thing becomes pleasantly boring.

Why a month is about right

Too quick and you skip testing. Too slow and momentum dies, people lose interest, and the old system limps on for another year. Four weeks is enough to audit what you have, port numbers properly, sort out the network and handsets, train the team, and still have a week in hand for the thing nobody predicted.

It also gives you room to deal with the backdrop. The old analogue network is being retired, with the final switch-off scheduled for the end of January 2027, so plenty of businesses are being nudged along whether they like it or not. If you are reading this in late 2026, you are not early and you are not late. You are right on time.

Week one: count what you have actually got

This is the unglamorous week, and it saves the most pain later.

Start with numbers. Every number you own, what each one is used for, and which ones customers actually recognise. Mainline, direct dials, the fax line nobody has touched since 2019, the number on the side of the van. Then map the behaviour: who answers what, when calls get diverted, how the out-of-hours message works, whether anyone relies on a hunt group or a receptionist transferring calls all day.

Next, the awkward stuff hiding in cupboards. Alarm panels, lift phones, card payment terminals, door entry systems, franking machines and any redcare-style monitoring lines. These are the classic migration casualties, and the fix is usually to keep a working analogue line for them rather than trying to force them onto VoIP.

Finally, look at your internet connection. VoIP is only as good as the link underneath it. Upload speed and consistency matter far more than headline download figures, so run a few tests during busy periods and note the jitter as well as the speed. Check your switch has enough power over Ethernet ports, or plan for injectors. Have a look at Wi-Fi coverage if anyone will be making calls on a mobile app.

By Friday you should have a spreadsheet, a list of contract end dates, and a clear idea of whether you want desk phones, softphones, or a mixture.

Week two: numbers and call routing

The single most important rule of number porting: do not cancel your old service. Porting a number can quietly terminate the line it sits on, and if that line is also carrying your broadband, you have just created a much bigger problem. Let the new provider manage the port and let them tell you what to do about the old contract.

Gather the paperwork early. You will typically need a letter of authorisation and a recent bill showing the numbers in their current format. Porting dates are booked, not instant, so decide the order. Many businesses port a secondary number first as a live test, leaving the main customer-facing number until the actual cutover.

While you wait, build your call flows on paper. Opening hours, lunch cover, overflow to mobiles, voicemail to email, a simple menu if you get enough calls to justify one. Keep it simpler than you think you need. Every extra option is another thing to explain to a caller and another thing to test.

Ask one specific question: how does caller ID present when we ring out? A surprising number of migrations quietly change this, and a client who does not recognise your number may not pick up.

Week three: handsets, headsets and the network

Now the visible bit. Desk phones still make sense for reception, sales floors and anyone who spends the day on the phone. Softphones suit people who move around and are already living in a laptop. Warehouse and yard teams often want a DECT handset that can survive being dropped. Headsets are not a luxury; they are the difference between a call that sounds professional and one that sounds like a wind tunnel.

Sort the network before the phones arrive, not after. Voice traffic needs priority, which in practice means a separate voice network or at least sensible quality-of-service settings on the switch. Plug 20 desk phones into an unmanaged switch behind the photocopier and you will spend week four apologising.

Check the boring infrastructure too: cabling, port count, and a small uninterruptible power supply for the switch and router so a brief power cut does not take the phones down with it. If the broadband fails, you want calls to divert to mobiles automatically, so test that behaviour now rather than discovering it at the worst moment.

Week four: train, cut over, keep a parachute

Pick a pilot group of three or four people, ideally including the least patient person in the building. Their complaints are your quality assurance. Give everyone a one-page cheat sheet and a ten-minute demonstration. Show them how to transfer a call, put someone on hold, and check voicemail. That is genuinely most of it.

Cut over outside business hours, on a Tuesday or Wednesday rather than a Friday, so you have working days to fix anything. Then run the test matrix: inbound to main number, inbound to a direct dial, outbound, transfer, hold, voicemail, conference if you use it, and a test call to 999 where you confirm the registered address is correct and remember not to actually connect.

And build the rollback plan. Keep the old system live and paid for until you are 30 days past cutover. Forward the main number rather than porting it on day one if you can. Keep at least one analogue line for the alarm and the lift. Know, in writing, who to ring and what the process is if it all goes sideways.

The checklist

• Every number listed, with its purpose and owner
• Alarms, lifts, PDQs and door entry identified and protected
• Upload speed, jitter and an Ethernet port count checked
• Port dates booked and old service left well alone
• Call routing mapped in writing, including out-of-hours
• Caller ID presentation confirmed
• Phones, headsets and power provisioned
• Voice traffic prioritised on the network
• Pilot group trained and cheat sheets handed out
• Cutover on a quiet weekday, with a 30-day rollback window

The traps worth knowing about

Most VoIP horror stories come from the same handful of causes: porting before testing, no network priority, no training, and nobody clearly owning the project. Add in an alarm line that nobody remembered and you have a perfect storm. None of it is difficult to avoid. It just needs a month, a spreadsheet and someone willing to chase.

Do it that way and the thing nobody says afterwards is how much they miss the old system.