Four in the Afternoon in Perth
Three staff at the Perth branch, six callers waiting, and the queue getting longer. In Adelaide, two people have wrapped up for the day and the phones are quiet. Both locations belong to the same business. Neither can help the other.
The capacity to answer those six calls exists, is already paid for, and is sitting idle. What is missing is not staff. It is a route.
The whole multi-site argument in two sentences
Meanwhile a customer rings the Newcastle branch at 5:15pm, gets no answer, and never discovers that the Maitland branch is open until six and would have taken the booking. That call is not recorded as lost anywhere, because it is not recorded at all — it rang out on a system that does not report to anything.
Both of those situations are routing problems with five-minute technical solutions, and both persist for years in businesses that are otherwise very well run. It is worth understanding why.
Why It Looks Like This
- Location one chose a phone system with proper thought. Nobody wrote down the reasoning.
- Location two opened under fit-out pressure. Someone local arranged a service that week, because the alternative was opening without a phone.
- Locations three to six each repeated location two, with whatever was quickest locally. Different handsets, different voicemail conventions, four different ideas of what "missed" means.
- An acquisition arrived with its own platform, its own numbers and a contract with eighteen months left on it. Nobody diarised the end date.
- A manager left and took the only knowledge of one site's admin login with them.
- Head office noticed — usually when a customer complained about not being able to reach a branch, or when somebody tried to build a report and found there was nothing to build it from.
This is worth saying out loud before any project starts
Every decision in that list was the fastest correct decision available at the time. The estate is not evidence of carelessness; it is evidence of nine deadlines. Projects that begin by implying otherwise lose the site managers in the first meeting, and the site managers are the people who have to make the new arrangement work.
The Four Costs, Smallest First
1. Duplicated charges
Line rentals, minimum spends, per-site support and maintenance, multiplied by nine. Visible on invoices, genuinely worth recovering, and almost always the smallest of the four.
2. Staffing for nine separate peaks
Each location rosters for its own busiest hour, so each is over-staffed at eleven and short at nine. Nine peaks staffed independently is materially more expensive than one network peak shared.
3. The inconsistent experience
A customer who deals with two of your locations meets two businesses. Different greetings, different hold behaviour, different odds of reaching anyone after five. Your brand ends at the phone line.
4. Unanswered calls
The big one, and it appears on no invoice. Individually trivial per site, enormous across nine, and invisible because nothing counts them in one place.
Cost four deserves an actual number, and you can get one this week without buying anything. For each location, from whatever reporting exists, count last month's calls that rang out unanswered, went to voicemail with no message left, or were abandoned while waiting. Add them across the estate. Estimate the share that were genuine enquiries rather than repeat attempts. Multiply by your average transaction value.
What businesses usually find
The figure exceeds the entire annual telephony spend, frequently by a multiple. That changes the nature of the project: it stops being a cost-reduction exercise that competes with everything else for capital, and becomes a revenue recovery exercise that happens to reduce cost as a side effect. It also relocates the conversation from IT to the people who own the revenue, which is where it belongs and where it gets decided.
A Worked Comparison
A nine-location business, 62 staff, illustrative and rounded. The shape is what matters, not the figures.
| Nine separate services | One platform | |
|---|---|---|
| Monthly service cost | $4,180 across four providers, nine bills | $2,790, one bill |
| Per-site support and maintenance | $690/month total, three contracts | Included |
| Admin effort | Nine logins, four interfaces, no single view | One view, delegated per site |
| Unanswered calls last month | Unknown — no consolidated figure exists | Reported, per site and total, one definition |
| Overflow between sites | Not possible | By time of day, queue depth or availability |
| Cross-site internal calls | Charged as external calls | Four digits, no charge |
| Adding a location | A new procurement, 4–8 weeks | A configuration, same week |
| Annual visible saving | ~$25,000 — and this is the least interesting line in the table | |
| Recovered calls, at 8 per site per week | ~3,700 calls a year that previously rang out, at whatever your conversion and transaction value are | |
Whatever your numbers, the ratio between the last two rows tends to hold. The saving pays for the project; the recovered calls are the actual return. Our note on what business phone systems cost covers how to normalise the first half properly.
What One Platform Changes
| Change | What it means day to day |
|---|---|
| Each location is an object, not a box | Its own numbers, hours, greetings, queues, holidays and staff. Edit one without touching the rest. Nothing on site but handsets, apps and a router. |
| One dial plan | Four digits between any two locations, any state, no charge. Internal calls stop being external calls. |
| Shared queues | A queue can be served by people in three locations. The caller does not know or care where the person answering is sitting. |
| Delegated administration | The site manager changes their own hours and staff. They cannot change another site, or override what the business has fixed centrally. |
| People, not desks | Someone covering three locations is one identity who logs in wherever they are. No second handset, no second licence, no forwarding chain. |
| One reporting definition | Per site, per region, consolidated. A comparison between two locations becomes a real comparison instead of an argument about how each system counts. |
None of this is a premium tier. It falls out of the architecture: when the exchange is a shared platform rather than equipment in a comms room, a location is a boundary in configuration. The background is in what a cloud phone system is.
Keeping the Local Number and the Local Feel
The first objection to consolidation is always the same, and it is a good one: customers ring the branch, not head office, and a local number is part of how a location belongs to its town. Consolidating the platform does not mean consolidating the identity.
Local numbers stay local
Each location keeps and answers its own published number, and presents it on outbound calls. Nothing about the platform requires a single national number.
Local greeting, local voice
The site's own recording, its own hours, its own on-hold messages about its own promotions. Recorded by someone who works there, if that is what suits.
A network number as well
A 1300 or 1800 for advertising, routed by caller location or availability, sitting alongside the local numbers rather than replacing them.
The bit that actually changes
When the local number is not answered, something happens instead of nothing. That is the whole difference, and it is invisible to every caller who does get answered locally.
One thing to check while you are in there
Confirm whose name each published local number is registered in. In estates built up over years — especially with acquisitions or changes of local operator — numbers are frequently registered to a previous owner or an entity that no longer trades. That decides whether the number can be moved at all, and it is far better discovered now than during a migration. See whose number it really is.
The Three Hours Nobody Uses
Australia spans three time zones, and for a business with locations in more than one, that is either an inconvenience or two extra hours of coverage a day at no additional staffing cost. Almost every multi-site business treats it as the first.
| Time | What happens now | What could happen |
|---|---|---|
| 7:00am eastern | Eastern locations are not open. Calls ring out or reach voicemail. | Nothing yet — nobody is at work anywhere. This is an AI answer or a message, not a routing problem. |
| 8:00am eastern | Eastern locations opening, busiest hour of the day, understaffed. | Perth is at 6:00am — no help. But Adelaide at 7:30am and Brisbane already open can absorb the eastern peak between them. |
| 4:30pm eastern | Eastern locations winding down; late callers get voicemail. | Perth is at 2:30pm and mid-shift. Eastern late calls route west and are answered by a person. |
| 6:00pm eastern | Everything east is closed. Perth is at 4:00pm and open. | Two full extra hours of live answering for the eastern states, using staff already rostered and already paid. |
| 7:00pm eastern | Perth closing too. Calls go nowhere. | The genuine after-hours boundary, and where an AI answer or an on-call path belongs. |
The gain is not theoretical: for an eastern-state customer, the difference between a business that answers at 5:40pm and one that does not is often the difference between a booking and a competitor. And it costs one routing rule.
The Report You Cannot Currently Build
Five lines. Every multi-location business wants them and almost none can produce them, for the simple reason that four platforms count four different ways.
| Line | Why it changes decisions |
|---|---|
| Calls offered, per location, per hour | Shows where demand actually is, which is frequently not where the staff are. The single most useful roster input available. |
| Answered within target, per location | One target across the network, so a comparison means something. Look at the distribution, not the average — an average hides two failing sites behind seven good ones. |
| Unanswered and abandoned, per location | The revenue leak, finally visible and finally owned by someone. |
| After-hours volume and outcome | Usually the biggest surprise in the first report ever produced, and usually the cheapest thing on the list to fix. |
| Repeat callers within a day | Somebody calling twice in one day is a first-contact failure. A better indicator of customer experience than any survey you will get returned. |
Introduce it in the right order
Give each location its own numbers first, privately, and let them act on them for a month before anybody sees a comparison. Reporting introduced as a tool gets used; reporting introduced as a ranking gets gamed — and the fastest way to improve an answer-rate ranking is to answer instantly and transfer the problem elsewhere. Pair the answer measure with the repeat-caller measure from day one, and say that you are doing so.
Who Gets Credit for the Call
The technical work to let Adelaide answer Perth's overflow takes about five minutes. The conversation about who books the resulting sale takes longer, and it must happen first.
- Decide what a receiving location may do. Take a message only? Take a booking? Quote? Complete a sale? The answer can differ by call type, and it should be written down rather than assumed.
- Decide where the revenue lands. With the location that owns the customer, the location that answered, or split. Any of the three works; ambiguity does not.
- Decide how it is visible. Both locations should be able to see overflow in and overflow out. Hidden flows breed suspicion faster than unfavourable ones.
- Then enable it. Start with sequential fallback — unanswered here, try there — which is the least contentious pattern and captures most of the benefit.
- Review after a month. Look at which locations gave and received. If one is consistently subsidising another, that is a staffing conversation the reporting has just handed you for free.
This matters most where locations are separately owned or separately measured on profit. Where they are simply branches of one business, steps one and four are usually enough.
The 000 Detail That Multi-Site Businesses Get Wrong
On any cloud platform, a number is not tied to a physical place — that is what makes cross-site working possible. It also means the address associated with each number is a record somebody maintains, and in a multi-location business that record goes stale faster than anywhere else, because locations open, close, relocate and share staff.
| Do this | Because |
|---|---|
| Hold a service address per location, not one for the company. | An emergency call from a regional branch should not present a head office address in another state. |
| Put an address check on the fit-out and relocation checklist. | Relocations are where this breaks. Nobody thinks about a phone record when a lease changes. |
| Tell staff who use mobile or laptop apps to state their location out loud. | They will not know this unless somebody tells them, and it is the one fact that matters most. |
| Put it in the induction at every site. | Two sentences, once, to every new starter across nine locations. |
The full detail, including how a call is held and returned, is in how 000 works from a cloud phone and what to tell staff about the silence before it connects.
The Order to Do It In
There is no version of this that should involve a single weekend and nine locations.
- Build the inventory. Every location: published numbers, provider, monthly cost, contract end date, whose name each number is registered in, what equipment is on site, who has the admin login. This spreadsheet is most of the project.
- Design one template. Greeting structure, hours pattern, queue behaviour, after-hours path, reporting definitions. Once, not nine times — designing per location just rebuilds the current mess somewhere newer.
- Pilot the awkward location. Not the biggest, not the easiest. The one with an unusual requirement: an on-call roster, a paging system, a legacy analogue device, a shared reception with another business. The template either survives that or it changes now.
- Convert as contracts expire. Not by size or by importance. This avoids paying two providers for the same location, and turns a risky programme into a routine rolling one.
- Enable overflow after the attribution conversation. Sites working individually first; inter-site routing once the credit question is settled.
- Publish reporting last. Locations see their own numbers before anyone sees a comparison.
Numbers need the most lead time: porting runs to an industry framework with standard windows, and a location whose numbers are registered to a third party may need a signed authority from an entity that has stopped trading. Start that on day one. See what actually goes wrong when businesses switch.
Four Reasonable Objections
| Objection | The honest answer |
|---|---|
| "Our locations are genuinely different and a standard will not fit." | Partly right, which is why the template covers structure rather than content: the shape of the greeting, the maximum wait before something happens, and how metrics are defined. Hours, wording, staffing and local routing stay local. If a location genuinely cannot fit the structure, it is the pilot site. |
| "One system means one point of failure." | The reverse of the usual concern, and worth answering properly. Nine independent systems means nine independent single points of failure, each unmonitored, each with a different recovery path. One platform on an operated network with per-site diversion means a location losing connectivity has its calls rerouted rather than dropped. The redundancy question in full. |
| "Our site managers will not want head office in their phone system." | Legitimate, and the fix is delegation rather than persuasion. They keep their own hours, greetings, staff and routing. What they gain is a report of their own missed calls, which most managers have never had and quickly find they want. |
| "We have contracts running for another two years." | Then the programme runs for two years, converting locations as they fall due. That is the normal shape of this work and it is cheaper than breaking contracts. The inventory is still worth building this month, because it is what makes the sequence obvious. |
What We Do
We operate our own network and platform, and in it a location is a configuration boundary rather than a piece of equipment. Practically, that means each of your sites keeps its own published numbers, its own hours, its own greetings and its own queues, administered by whoever you nominate, with anything you decide to fix centrally left unchangeable locally. Extensions dial each other on four digits across every state at no cost. Overflow between locations is a rule you can edit yourself — by time, by queue depth, by who is available. Reporting arrives per location, per region and consolidated, from one source with one definition, so comparing two branches is a comparison rather than a dispute.
Numbers we allocate are registered with rights of use in the name you nominate, which is what keeps an advertised local number from leaving with a departing operator. And we migrate location by location as your contracts fall due, because a coordinated nine-site cutover has no upside for anyone except a project plan.
In short
Nobody designed your nine-location phone estate; nine deadlines did. The visible cost is duplicated charges and the real cost is unanswered calls, which you can quantify this week from the reporting you already have — and the figure usually exceeds the entire telephony spend. One platform keeps every local number and every local greeting while adding shared queues, overflow, four-digit dialling between sites and one reporting definition. Australia's three time zones become two extra hours of live answering for the eastern states at no extra staffing cost. Settle who gets credit for an overflowed call before you enable overflow, keep a correct service address per location for emergency calls, and convert locations one at a time as contracts expire — starting with the awkward site, not the easy one.
Related reading: the five numbers your system should report, what it costs, what goes wrong when you switch, whose number it is, and how redundancy works.
Frequently Asked Questions
Can multiple business locations share one phone system?
Yes, and on a cloud platform it is the ordinary arrangement rather than a special build, because once the exchange runs as software in a data centre a location becomes a boundary in configuration rather than a piece of equipment. Each site exists in the platform with its own published numbers, its own opening hours, its own greetings, its own queues, its own holidays and its own staff, so you can change one without touching any other. There is one dial plan across the whole business, so any extension can ring any other extension in any state on four digits at no cost, and internal calls stop being charged as external calls. Queues can be served by people in several locations at once, and the caller neither knows nor cares where the person answering is sitting. Administration can be delegated, so a site manager changes their own hours and staff but cannot alter another site or override what has been fixed centrally. Someone who covers three locations is one identity who logs in wherever they are, rather than three handsets and a chain of forwards. And reporting is available per site, per region and consolidated from a single definition, which is what makes a comparison between two locations meaningful instead of an argument about how each system counts a missed call.
What does running separate phone systems at each location actually cost?
Four costs, and the invoice only shows the smallest one. Duplicated charges come first: line rentals, minimum spends and per-site support or maintenance contracts multiplied by the number of locations, which is real money and worth recovering but rarely the main event. Staffing for separate peaks is second: each location rosters for its own busiest hour, so each is over-staffed in the middle of the day and short at opening, and nine peaks staffed independently costs materially more than one network peak shared. Inconsistent experience is third: a customer dealing with two of your locations meets two different businesses, with different greetings, different hold behaviour and different odds of reaching anybody after five. Unanswered calls are fourth and almost always the largest, and they appear on no invoice at all. You can size that one this week without buying anything: for each location, count last month's calls that rang out, reached voicemail with no message, or were abandoned in a queue, add them across the estate, estimate the share that were genuine enquiries, and multiply by your average transaction value. Businesses doing this for the first time usually find the figure exceeds their entire annual telephony spend, which turns the project from cost reduction into revenue recovery.
Will consolidating our phone system mean losing our local numbers?
No, and this is the most common objection to consolidation as well as the easiest to answer. Consolidating the platform is not the same as consolidating the identity. Each location keeps and answers its own published local number, and presents that number on outbound calls, so a customer ringing the branch still reaches the branch and sees a local number when the branch rings them. Each site keeps its own greeting in its own voice, its own opening hours and its own on-hold messages about its own local promotions, recorded by somebody who actually works there if that is what suits. You can add a network 1300 or 1800 number for advertising alongside the local numbers rather than instead of them, routed by caller location or availability. The only thing that genuinely changes is what happens when a local number is not answered: instead of ringing out, the call can go to another location, a shared queue, or an after-hours answer. That is invisible to every caller who is answered locally, and it is the entire benefit. While you are reviewing numbers, check whose name each one is registered in, because in estates built up over years numbers are often registered to a previous owner or an entity that has stopped trading, and that determines whether they can be moved at all.
How can locations in different time zones cover each other's calls?
With a routing rule, and Australia's three-hour spread makes this unusually valuable for any business with locations in more than one zone. The pattern that pays best is late-afternoon coverage flowing west. At half past four in the eastern states, eastern locations are winding down and late callers are getting voicemail, while a Perth office is at half past two and mid-shift, so eastern calls can route west and be answered by a person. By six in the evening eastern time, everything east is closed and Perth is still at four in the afternoon, which delivers roughly two extra hours of live answering for eastern-state customers using staff who are already rostered and already paid. The same logic works in reverse in the morning: eastern locations opening into their busiest hour can share load with each other and with Brisbane, though Perth is too early to help. For an eastern-state customer, the difference between a business that answers at twenty to six and one that does not is frequently the difference between a booking and a competitor, and the change costs one routing rule. The boundary where an AI answer or an on-call path belongs is when the last location in the west closes, not when the first one in the east does.
Who gets credit for a sale when another branch answers the call?
That is a decision to make before enabling overflow rather than after, because the technical change takes about five minutes and the commercial argument does not. Work through four questions in order. First, what is a receiving location permitted to do with another location's call — take a message only, take a booking, provide a quote, or complete a sale? The answer can legitimately differ by call type, and it should be written down rather than assumed. Second, where does the revenue land: with the location that owns the customer, the location that answered, or split between them? Any of the three arrangements works in practice; ambiguity does not. Third, how is it made visible? Both locations should be able to see overflow received and overflow sent, because hidden flows breed suspicion much faster than unfavourable ones. Fourth, start with the least contentious pattern — sequential fallback, where a call unanswered at one location is offered to another — which captures most of the benefit with the least argument. Then review after a month and look at which locations gave and which received; if one is consistently subsidising another, the reporting has just handed you a staffing conversation you could not previously have had. This matters most where locations are separately owned or separately measured on profit.
How do you move several locations to a new phone system safely?
One location at a time, sequenced by contract expiry, with no coordinated network-wide cutover at any point. Start by building an inventory covering every location: published numbers, current provider, monthly cost, contract end date, whose name each number is registered in, what equipment is on the premises, and who holds the admin login. That single spreadsheet is most of the project, and it surfaces the numbers registered to unexpected entities early, which is when you want to find them. Next design one template rather than nine configurations — greeting structure, hours pattern, queue behaviour, after-hours path and reporting definitions — because designing once and applying it many times is the whole economic argument, while designing per location simply rebuilds the existing mess inside a newer platform. Then pilot the awkward location rather than the biggest or easiest one: a site with an on-call roster, a paging system, a legacy analogue device or a shared reception. Convert the rest as their contracts fall due so you never pay two providers for the same site. Enable inter-site overflow only after the attribution question is agreed. Publish reporting last, showing each location its own numbers before anyone sees a comparison. Start the number work on day one, because porting runs to standard industry windows and a number registered to a company that no longer trades needs a signed authority.
Isn't one phone platform across all our sites a single point of failure?
It is the reverse, though the concern is a fair one to raise. Nine independently procured systems are nine separate single points of failure, each one unmonitored, each with a different recovery path, and each dependent on whoever at that site happens to know the admin login. When one fails, nothing routes around it because there is nothing to route to. A single platform operated on a provider network changes the failure mode rather than adding one: because the exchange is not at any of your sites, a location that loses its internet connection loses its desk phones while its inbound calls continue to arrive at the platform and can be diverted automatically to another location, to a mobile, to a shared queue or to an answer that takes messages. A power failure behaves the same way, and mobile and laptop applications keep working entirely because they are not on your network. The honest framing is that the comparison is not between reliable and unreliable but between an outage that takes a location's calls off the air and an outage that reroutes them. What determines which you get is whether the diversion path has been configured and tested, so ask for it specifically and then deliberately disconnect one site on a quiet afternoon to confirm the calls land where you expected.
