Before You Decide Anything
Two questions worth answering honestly before you go to market, because they change what you should be shopping for.
What is actually wrong?
Price, call quality, support response, missing capability, or a service being retired? Businesses that move for “a better deal” without naming the problem tend to reproduce the same problem somewhere cheaper.
What is forcing the timing?
A contract date, a product retirement, an office move, or nothing at all. If nothing is forcing it, you have the luxury of a slow, tested changeover — which is the kind that goes well.
If a product retirement is driving this, the date is not negotiable and the queue gets longer as it approaches. The legacy small business products being retired and their dates covers that situation specifically, and the operative advice there is to move early rather than into the rush.
One more piece of preparation that pays for itself repeatedly: get a call report from your current provider before you start. Volume by hour, by day, by number. You need it to size what you are buying, and you will want it afterwards to prove the new arrangement is performing. It is much harder to obtain once you have given notice.
Reading Your Exit Terms
The single most common unpleasant surprise in a changeover is financial, not technical, and it is entirely avoidable by reading four things.
| What to look for | Why it matters |
|---|---|
| The end date, per service | Businesses commonly have several services on different dates — lines, internet, mobiles — often signed at different times. There is rarely one contract end date, and assuming there is causes most of the trouble |
| Early termination charges | Usually the remaining months of the service charge, sometimes plus unamortised hardware. Get a written figure, per service, before you commit to anything |
| Notice period | 30 days is typical. A notice period that runs past your intended cutover means paying for two services for a period — which is often the correct decision anyway, but should be a choice rather than a discovery |
| Bundling and hardware | If phones, internet and mobiles are bundled, removing one may reprice the rest. And leased handsets usually have to be returned or bought out |
The distinction that catches people
Porting a number away is not the same as cancelling the service, and it does not terminate your contract. The port moves the number. The contract, and any early termination charge, continues to exist until it is dealt with separately. Businesses that assume the port ends everything discover the position on the next invoice. Conversely — and this is the more expensive mistake — never cancel a service in order to move it. See the porting section below.
Worth knowing that consumer protections in this area were strengthened during 2026 as the ACMA moved to replace the industry TCP Code with a directly enforceable standard, with clearer obligations around contract information and switching. What the enforceable standard changes covers that shift. It does not make an early termination charge disappear, but it does mean your provider should be able to state the position in writing quickly, and it is reasonable to insist on that.
What Actually Moves, and How
Your phone number is not owned by your provider. It is allocated under the national numbering arrangements and can be moved between providers — that is what number portability means, and it is why a business is not trapped by the number it has advertised for fifteen years.
The mechanics run under an industry Local Number Portability framework which sets the operational procedures between providers, including standard hours of operation and activation timeframes. Practically, you sign an authority with the gaining provider, they lodge it with the losing provider, and on an agreed date the number moves. You do not contact the losing provider to arrange it, and you should not.
| Simple port | Complex port | |
|---|---|---|
| Typically | A single standalone number | A number range, multiple services, or legacy technology such as ISDN |
| Timeframe | Days | Weeks — plan on several, and more if anything is unusual |
| Cutover | Often same-day, sometimes near-immediate | Scheduled window, agreed in advance between providers |
| Main risk | Details on the authority not matching the losing provider's records | Range splitting, incomplete inventories, and dependent services nobody listed |
Most businesses with a main number and a set of direct numbers are in the second column, and should plan on that basis. Three things are worth knowing before you start.
Ranges usually move whole
If you hold a block of numbers, taking part of it and leaving the rest is often not possible. Establish early whether the whole block is coming, because it changes the plan.
The authority must match exactly
Business name, address and account number as the losing provider holds them — not as you would write them today. A trading name where the records say the ACN entity is the most common rejection.
Never cancel to move
Cancelling the service releases the number, which then goes to quarantine. At that point it is no longer yours to port. This is the one genuinely irreversible mistake available in this process.
Say this to whoever handles your accounts
If the losing provider rings to ask whether they should “close the account” while a port is in flight, the answer is no, do nothing. Someone in the business, trying to be helpful and tidy up, agreeing to a cancellation mid-port is a genuinely common way to lose a number permanently. Tell the person who answers those calls, in advance, in writing.
For the number-level detail — authority forms, what a rejection looks like and how to fix it — our porting guide covers that ground. This article is about everything around it.
Inbound Numbers Are a Separate Job
A 1300 or 1800 number is not a geographic number and does not move by the same mechanism. It is an inbound service with its own arrangements, and the practical consequences are worth planning around.
| Point | What it means for your changeover |
|---|---|
| Different process, different timeline | Run it as a parallel workstream from day one rather than assuming it rides along with the geographic numbers |
| Answer points are configuration, not wiring | Where a 1300 number terminates is a setting. Once the new service is live, redirecting to it is fast — which makes inbound numbers a useful safety net during cutover |
| Routing rules travel with it | State-based routing, time-of-day rules and overflow behaviour all need to be reproduced. Document the current rules before anything changes; nobody remembers them afterwards |
| Smart numbers have their own paperwork | A memorable or word-based number may sit under separate arrangements. Check ownership and any rights of use early rather than at cutover |
Use them as your safety net
Because inbound number answer points are just configuration, an advertised 1300 number gives you a fast, reversible way to steer traffic during a changeover — point it at the old service until the new one is proven, then move it in one change. If your business advertises a 1300 number as its primary contact, your cutover risk is materially lower than a business advertising a geographic number. Background in our 1300 and 1800 number guide.
The Devices Nobody Remembers
This is the section that earns its place. A phone line at an Australian business is rarely carrying only phone calls, and the other things on it fail silently — they keep looking normal at your end while nothing arrives at the other.
| Device | How it fails | Severity |
|---|---|---|
| Back-to-base alarm dialler | The panel arms, beeps and lights up normally. Nothing reaches the monitoring centre. Discovered only when there is a break-in | Critical |
| Medical alert pendant | Identical failure mode. The pendant works, the base unit responds, no signal arrives | Critical |
| Fire panel dialler | Stops reporting. May also create a compliance and insurance problem independent of the safety one | Critical, and usually the longest lead time |
| Lift emergency phone | Someone presses the button in a stuck lift and nothing happens | Critical, separate contractor, book early |
| EFTPOS terminal on a line | Declines or fails to settle. Highly visible and immediately expensive | High — but at least it announces itself |
| Gate and door intercoms | Visitors cannot be let in. Annoying rather than dangerous, and always discovered at the worst moment | Moderate |
| Fax | Still real in legal, medical and freight. Behaves poorly over some services and needs a deliberate plan, not an assumption | Moderate, occasionally critical |
| Franking machines, ATMs, vending, irrigation controllers | Whatever they do, they stop doing it quietly | Varies |
The procedure is the same for all of them and it is not complicated.
- Walk the building. Every site. Look for anything plugged into a phone socket, including in ceiling spaces, comms cupboards, lift motor rooms and behind reception. Do not do this from a list — do it with your feet.
- Identify the contractor for each. Alarm, medical alert, fire and lift are four different companies with four different lead times. Fire and lift are usually the longest and therefore the critical path.
- Contact them before you set a cutover date, not after. Their availability determines your date, not the other way round.
- Get a written signal test per device after cutover. Not a verbal assurance, not an assumption. A test, and a record of it.
If you do one thing from this article
Do the walk. Every genuinely serious changeover failure this playbook is written to prevent involves a monitored device that stopped reporting and was not noticed for weeks. Everything else on this page produces inconvenience, an invoice or a bad afternoon. This one produces the incident.
The Six-Week Timeline
Six weeks is comfortable for a typical multi-number business. It compresses to three if it must, and if a retirement date is forcing you, start earlier than you think you need to — everyone else in the same position is booking the same contractors.
| Week | What happens | Owner |
|---|---|---|
| −6 | Exit terms in writing per service. Call report obtained. Site walk done and the device inventory exists | You |
| −5 | Contractors contacted for alarm, medical alert, fire, lift and EFTPOS. Their availability sets the cutover date | You |
| −4 | Port authority lodged with correct legal entity details. Number range scope confirmed. Inbound number workstream started | New provider |
| −3 | New service built and tested alongside the old one. Call flows configured. Existing routing rules documented and reproduced | New provider |
| −2 | Handsets on desks, apps installed, staff shown how it works. Test calls in and out on temporary numbers | Both |
| −1 | Cutover date confirmed in writing. Contractors booked for the day after. Rollback position agreed and written down | Both |
| 0 | Cutover. Tuesday or Wednesday morning | Both |
| +1 | Device signal tests. Old service kept alive but idle. Nothing cancelled | Both |
| +2 to +4 | Verify traffic against the old call report. Then, and only then, close the old services | You |
Two dates that are not arbitrary
Tuesday or Wednesday morning, so a problem has three business days and available contractors before the weekend. Friday afternoon cutovers are how a silent alarm failure gets a 65-hour head start. And run the old service in parallel for a few weeks where you can. The overlap costs one month of line rental and is the cheapest insurance in the entire project.
Cutover Day, Hour by Hour
A well-prepared cutover is undramatic. Here is what the day looks like when the preparation has been done.
| Time | What happens |
|---|---|
| Before start | Everyone knows it is today. One named person is coordinating. The rollback position is written down and agreed |
| Port window | Numbers move. Expect a short period where behaviour is inconsistent — some calls old, some new. This is normal and not a fault |
| +30 min | Call every main number from an external mobile. Not from the office — from a mobile on a different network, which is what a customer actually is |
| +1 hr | Test outbound, including caller ID presentation. A changeover that presents the wrong number outbound is a real and common fault worth catching immediately |
| +1 hr | Test the call flows: main menu, each queue, transfer between people, voicemail, after-hours path |
| +2 hr | Test a 000 call if your provider supports a test procedure, and confirm the registered service address is correct. Do not skip this |
| Afternoon | Walk the floor. Ask people whether anything is odd. Frontline staff notice things monitoring does not |
| End of day | Written note of anything unresolved, with an owner against each item |
The emergency-call check deserves emphasis. A cloud phone service carries a registered service address used to route an emergency call, and that address needs to be right on day one rather than at some later tidy-up. Triple Zero from a cloud phone covers the obligations and the specific things to verify.
The First Week After
Cutover day catches the obvious. The first week catches the rest, and it is where a good changeover distinguishes itself from a merely completed one.
| Check | Why |
|---|---|
| Signal test every monitored device | In writing, per device. The single highest-value item on the list |
| Compare call volume to the old report | If Tuesday used to bring 90 calls and now brings 60, something is not arriving. This is the check that finds a broken routing rule nobody noticed |
| Ring the number on your website, van and invoices | Businesses port the numbers they remember. The old fax number on the back of the invoice book is a real category of problem |
| Check after-hours and weekend behaviour | Configured last, tested least, and it is when a large share of enquiries arrive |
| Ask the team, once, in a meeting | “What is annoying about the new phones?” Small irritations get worked around silently and then become permanent |
Then close the old services deliberately — checking that nothing depends on them first, and keeping written confirmation of what was cancelled and when.
How the Bad Ones Actually Fail
Six failure modes account for nearly every changeover that goes badly. All six are preventable and none is technically difficult.
| Failure | Prevention |
|---|---|
| Someone cancelled the old service to “help” and the number was lost permanently | Tell the accounts contact in writing, before you start, that no cancellation is to be agreed to while a port is in flight |
| A monitored alarm stopped reporting and nobody knew for six weeks | The walk, the contractors, and a written signal test per device after cutover |
| The port was rejected repeatedly on entity name mismatches, and the date slipped past a retirement deadline | Use the legal entity exactly as the losing provider records it. Lodge early enough to absorb one rejection |
| Only part of the number range could move, discovered late | Confirm range scope in week −4, before anything is scheduled or announced |
| An early termination charge nobody knew about | Written figures per service, in week −6. Then decide with the number in front of you |
| Cutover on a Friday, problem found Monday, contractors booked out | Tuesday or Wednesday morning. This is free and it removes an entire class of outcome |
6 wks
Comfortable lead time
Tue/Wed
The only good cutover days
8+
Device types on a typical site
0
Services you cancel yourself
None of this is difficult. It is just unglamorous, and it happens in the weeks when everyone would rather be discussing handsets. A business that does the walk, reads its exit terms, lodges the authority with the right entity name and picks a Tuesday will have a changeover nobody talks about afterwards — which is exactly the outcome you want.
If you are still choosing rather than moving, the comparison of business phone systems in Australia and what a phone system actually costs cover that stage, and plug-and-play provisioning covers why the handset part of week −2 is usually the easiest part of the entire project.