How to Set Up Multi-Site VoIP for Branch Offices

For UK businesses operating across multiple locations, unifying voice communications under a single, centrally managed VoIP platform is one of the most impactful infrastructure decisions you can make. Whether you run two branch offices or twenty, a well-architected multi-site VoIP deployment eliminates the fragmented, costly patchwork of separate phone systems that plagues so many organisations — and replaces it with seamless, four-digit extension dialling, centralised call routing, and a single management portal that covers every site from Aberdeen to Plymouth.

Multi-site VoIP is not simply a matter of installing the same system in each office. It requires careful planning around network architecture, inter-site connectivity, number allocation, bandwidth provisioning, and disaster recovery. Get it right, and your staff can collaborate as if they were all under one roof. Get it wrong, and you end up with call quality problems, siloed systems, and frustrated employees.

This guide walks through every aspect of setting up multi-site VoIP for branch offices in the UK — from choosing between centralised and distributed architectures to planning bandwidth per site, configuring extension dialling across locations, and building resilience into the system from day one. Whether you’re expanding into new offices, consolidating legacy PBX systems, or preparing for the PSTN switch-off, you’ll find practical, actionable guidance here.

Centralised vs Distributed VoIP Architecture

The first and most consequential decision in any multi-site VoIP deployment is the choice between a centralised and a distributed architecture. Each approach has distinct advantages and trade-offs, and the right choice depends on your organisation’s size, network infrastructure, and operational requirements.

Centralised Architecture

In a centralised model, a single VoIP platform — whether cloud-hosted or on-premise at your headquarters — handles all call processing for every branch office. Remote sites connect back to the central system over your WAN (wide area network), with IP phones at each branch registering to the central call server. All call routing logic, voicemail, auto-attendant menus, call recording, and management functions live in one place.

This approach offers simplicity and consistency. There is one system to manage, one set of dial plans to maintain, one configuration to back up. Adding a new branch is a matter of connecting phones to the network and pointing them at the central platform. For organisations with up to around 20 sites and reliable inter-site connectivity, centralised architecture is typically the best choice.

The trade-off is WAN dependency. Every call at a branch office — including local calls between colleagues sitting in the same room — must traverse the WAN link back to the central system. If that link goes down, the branch loses telephony entirely unless you have built in local survivability (more on this later).

Distributed Architecture

In a distributed model, each branch office runs its own local VoIP platform or session border controller (SBC), with inter-site calls routed between the locations. Each site can operate independently if the WAN link fails, processing local calls without needing the central system. A master management platform provides unified oversight, configuration synchronisation, and centralised reporting.

Distributed architecture suits larger organisations with many branches, particularly those in locations with less reliable connectivity (rural offices, industrial sites) or where local call volumes are high. It reduces WAN bandwidth consumption because intra-office calls stay local, and it provides inherent resilience at each site.

The trade-off is complexity. You are managing multiple VoIP instances, keeping configurations synchronised, and maintaining hardware or virtual appliances at each site. Licensing costs may also be higher, depending on the platform.

Inter-Site Connectivity: MPLS vs SD-WAN

Your choice of WAN technology directly determines the quality, reliability, and cost of voice traffic between sites. For multi-site VoIP, you need a network that can deliver low latency, minimal jitter, and guaranteed bandwidth for voice packets — even during periods of heavy data traffic. The two dominant options in the UK market are MPLS and SD-WAN.

MPLS (Multi-Protocol Label Switching)

MPLS has been the gold standard for enterprise WAN connectivity for over two decades. It creates a private, managed network between your sites, operated by a single carrier. Traffic is label-switched rather than routed at each hop, which provides predictable performance and inherent Quality of Service (QoS). Voice packets can be assigned to a high-priority class of service that guarantees bandwidth and minimises latency, regardless of how much data traffic is flowing simultaneously.

For VoIP, MPLS offers excellent call quality because you are not sharing infrastructure with public internet traffic. Latency between UK sites typically stays below 20 milliseconds, jitter is negligible, and packet loss is effectively zero under normal conditions.

However, MPLS is expensive. A typical 10 Mbps MPLS circuit between two UK sites costs £400–£800 per month, and you need a circuit to every branch. For a business with five offices, WAN costs alone can reach £2,000–£4,000 per month. Lead times for new circuits can stretch to 60–90 days, making rapid expansion difficult.

SD-WAN (Software-Defined Wide Area Network)

SD-WAN is the modern alternative that has rapidly gained market share in the UK since 2022. Rather than relying on a single expensive MPLS circuit, SD-WAN aggregates multiple lower-cost connections — business broadband, FTTP, 4G/5G, and optionally a smaller MPLS circuit — into a single logical WAN. Intelligent software monitors the performance of each connection in real time and dynamically routes traffic over the best available path.

For VoIP, SD-WAN applies application-aware routing that recognises voice packets and steers them over the lowest-latency, lowest-jitter connection. If the primary broadband link develops packet loss, SD-WAN can switch voice traffic to a backup 4G connection within milliseconds, without dropping the call. Many SD-WAN platforms also include built-in packet duplication and forward error correction that further improve voice quality over imperfect connections.

The cost savings are substantial. Replacing MPLS with SD-WAN over commodity broadband typically reduces WAN costs by 40–70%, while delivering comparable or superior performance for voice traffic.

WAN Cost Comparison for a Five-Site UK Business

Single-Tenant vs Multi-Tenant Systems

When selecting a multi-site VoIP platform, you will encounter two fundamentally different hosting models: single-tenant (dedicated) and multi-tenant (shared). Understanding the distinction is critical for making the right choice for your organisation.

Multi-Tenant (Shared Platform)

In a multi-tenant deployment, your organisation shares the VoIP infrastructure with other customers of the same provider. Each tenant has their own logically isolated configuration — separate extensions, call routes, voicemail boxes, and admin portal — but the underlying servers, software, and SIP trunks are shared. This is the model used by the vast majority of hosted VoIP and UCaaS providers in the UK.

Multi-tenant platforms are cost-effective and fully managed. The provider handles all hardware, software updates, security patches, and capacity planning. You pay a predictable per-user, per-month fee that includes everything. For UK businesses with up to around 500 users across multiple sites, this is typically the most practical and affordable option.

Single-Tenant (Dedicated Platform)

A single-tenant deployment gives your organisation its own dedicated VoIP infrastructure — either physical servers in a data centre or dedicated virtual machines in a private cloud. Nobody else shares your platform. This provides complete control over configuration, security policies, update schedules, and customisation.

Single-tenant suits organisations with strict regulatory requirements (financial services under FCA/MiFID II, healthcare with NHS Data Security and Protection Toolkit compliance), very high call volumes (contact centres handling 10,000+ calls per day), or highly customised call workflows that cannot be accommodated on a shared platform.

The trade-off is cost. Single-tenant deployments typically cost two to four times more than multi-tenant equivalents, and you take on responsibility for capacity planning and some maintenance tasks, even in a managed hosting arrangement.

Factor	Multi-Tenant	Single-Tenant
Typical cost per user/month	£8 – £25	£20 – £60
Customisation	Platform features only	Full customisation available
Update schedule	Provider-managed, automatic	You choose when to update
Security isolation	Logical separation	Physical/VM-level separation
Best for	SMEs and mid-market (5–500 users)	Enterprise and regulated sectors (200+ users)

Extension Dialling Between Sites

One of the most immediate, tangible benefits of multi-site VoIP is unified extension dialling. Instead of staff at your Manchester branch needing to dial a full telephone number to reach a colleague in London, they simply dial a short extension — typically three or four digits — exactly as if the person were in the next room.

Designing Your Extension Plan

A well-designed extension numbering plan is essential for a smooth multi-site deployment. The most common approach is to use a site-prefix system where the first digit or digits identify the location, and the remaining digits identify the individual user.

For example, a business with four UK offices might structure extensions as follows:

Site	Extension Range	Example
London HQ	1000 – 1999	1001 (Reception), 1042 (Sales Manager)
Manchester	2000 – 2999	2001 (Reception), 2015 (Accounts)
Birmingham	3000 – 3999	3001 (Reception), 3008 (Engineering)
Edinburgh	4000 – 4999	4001 (Reception), 4022 (HR)

This structure is intuitive — staff quickly learn that “2xxx” extensions are Manchester, “3xxx” are Birmingham, and so on. It also scales well, accommodating up to 999 extensions per site. For smaller branches, you might use three-digit extensions (100–199 for London, 200–299 for Manchester) to keep things even simpler.

Hunt Groups and Ring Groups Across Sites

Multi-site VoIP allows you to create cross-site hunt groups that span multiple locations. A sales enquiry line, for example, can ring simultaneously at sales desks in London, Manchester, and Birmingham — whichever agent answers first takes the call. This distributes workload naturally across sites and ensures callers are never told “that department is at our other office, let me transfer you.”

You can also configure time-based overflow. If the London sales team is busy (or the office has closed for the day), calls automatically overflow to Manchester or Edinburgh, providing extended coverage without additional staffing costs.

Centralised Management and Administration

Managing phone systems independently at each branch office is a time-consuming, error-prone nightmare. Different firmware versions, inconsistent configurations, separate admin portals, and no unified view of call activity across the business. Multi-site VoIP eliminates this entirely.

Single Pane of Glass Management

A properly deployed multi-site VoIP system provides a single web-based admin portal from which your IT team (or your managed service provider) can manage every aspect of telephony across all sites. From this portal, you can:

• Add, modify, or remove users at any site — provision a new starter in Edinburgh from your desk in London in under two minutes
• Configure call routing — set up auto-attendants, call queues, hunt groups, and time-based routing rules that span multiple locations
• Monitor call quality in real time — view Mean Opinion Score (MOS), jitter, latency, and packet loss metrics per site, per user, or per call
• Pull consolidated reports — analyse call volumes, missed call rates, average answer times, and agent performance across all branches in a single dashboard
• Apply configuration templates — create standardised phone configurations and push them to all sites simultaneously, ensuring consistency
• Manage firmware updates — schedule handset firmware updates across all locations from one interface, avoiding version fragmentation

Role-Based Access Control

Enterprise multi-site VoIP platforms support granular role-based access. A local office manager might have permission to add users and modify hunt groups at their own site but cannot alter global call routing or access other sites’ configurations. IT administrators can be given full access across all locations, while C-level executives might receive read-only access to analytics dashboards. This approach balances local autonomy with central governance.

Bandwidth Planning Per Site

Insufficient bandwidth is the single most common cause of poor VoIP call quality in multi-site deployments. Each concurrent VoIP call requires a predictable amount of bandwidth, and you must provision enough at every site to handle peak call volumes alongside regular data traffic.

Bandwidth Requirements Per Call

The bandwidth consumed per call depends on the codec used. Here are the standard figures including all protocol overhead:

Codec	Audio Quality	Bandwidth per Call	Use Case
G.711	Excellent (HD Voice)	87 Kbps	LAN and high-bandwidth WAN links
G.729	Good	31 Kbps	WAN links with limited bandwidth
Opus	Excellent (adaptive)	20–90 Kbps	Variable-quality links, SD-WAN
G.722	Very good (wideband)	80 Kbps	Internal calls between sites

Calculating Bandwidth Requirements

To calculate the bandwidth needed at each site, multiply the maximum number of concurrent calls by the per-call bandwidth of your chosen codec, then add a 20% overhead margin for signalling, network overhead, and headroom.

For a branch office with 30 staff where a maximum of 10 people are typically on calls simultaneously using G.711:

• 10 calls × 87 Kbps = 870 Kbps for voice
• Add 20% overhead = approximately 1.05 Mbps dedicated to voice
• Add existing data traffic requirements (email, web, cloud apps, file transfers)
• Total bandwidth should be provisioned at minimum 20 Mbps, ideally 50+ Mbps, with QoS reserving the voice allocation

Bandwidth Adequacy by Site Size

UK Multi-Site Deployment Considerations

Deploying VoIP across multiple UK sites brings specific challenges and opportunities that differ from single-site installations. Here are the key factors to plan for.

Geographic Connectivity Variation

The UK has significant regional variation in broadband quality. Urban offices in London, Manchester, and Birmingham typically have access to full-fibre (FTTP) connections delivering 900+ Mbps, making bandwidth a non-issue. However, branch offices in rural areas of Wales, Scotland, or the West Country may be limited to ADSL (up to 10 Mbps) or FTTC (up to 80 Mbps), which requires more careful bandwidth planning and potentially the use of lower-bandwidth codecs like G.729.

For sites with poor fixed-line broadband, consider 4G/5G backup or primary connections. Business-grade 4G routers with external antennas can deliver reliable 30–80 Mbps in areas with good signal, and 5G coverage is expanding rapidly across UK urban and suburban areas.

The PSTN Switch-Off and Multi-Site Migration

With BT’s PSTN switch-off scheduled for January 2027, any branch offices still using traditional ISDN or analogue lines must migrate to VoIP. For multi-site businesses, this is actually an opportunity to consolidate. Rather than replacing each site’s legacy system with a standalone VoIP solution, use the switch-off as a catalyst to deploy a unified multi-site platform.

Openreach has designated certain exchanges as “stop-sell” areas where new PSTN lines can no longer be ordered. Check whether any of your branch locations fall in these areas — if so, migration is not optional, it is mandatory.

Regulatory Considerations

UK multi-site VoIP deployments must comply with several regulatory requirements:

• Emergency services (999/112) — each site must be configured with its correct physical address for emergency call routing. When a user dials 999, the system must relay the correct location to the emergency operator. This is handled via ELIN (Emergency Location Identification Number) configuration or, in cloud platforms, by registering each site’s address with the provider
• Ofcom numbering regulations — geographic numbers (01/02) must relate to the geographic area they represent. You cannot present a London 020 number from a Manchester office without specific regulatory provisions
• GDPR — call recordings stored centrally must comply with data protection requirements, including lawful basis for recording, secure storage, access controls, and retention policies
• FCA/MiFID II — financial services firms must record and archive all relevant communications, with recordings stored securely for a minimum of five years

Number Planning for Multi-Site Deployments

Telephone number planning is a critical but often overlooked aspect of multi-site VoIP. A well-structured numbering plan ensures callers reach the right location, your business presents a professional image, and you comply with UK numbering regulations.

Geographic vs Non-Geographic Numbers

Geographic numbers (01/02 prefixes) are tied to specific UK area codes. Having a local geographic number in each area where you operate builds trust with local customers — a Birmingham customer is more likely to answer a call from a 0121 number than a generic 03 number. Multi-site VoIP makes it easy to present the correct local CLI (Caller Line Identity) for outbound calls from each branch.

Non-geographic numbers (03 prefixes) are location-independent and charged at local call rates. They are ideal for national contact numbers that route to whichever branch is best placed to handle the call. A single 03 number can serve as your main business line, with intelligent routing directing calls based on the caller’s location, time of day, or agent availability.

Recommended Number Structure

Number Type	Purpose	Example
Main national number (03xx)	Single point of contact for all enquiries	0330 123 4567
Local geographic per site	Local presence and outbound CLI	020 1234 5678 (London), 0161 234 5678 (Manchester)
Freephone (0800)	Sales and customer service lines	0800 123 4567
DDI ranges per site	Direct dial-in to individual staff	020 1234 5600–5699 (London block)

Number Porting Across Sites

When consolidating multiple legacy phone systems, you will likely have numbers from multiple providers that need porting to your new unified platform. In the UK, number porting between VoIP providers typically takes 5–10 working days, but porting from legacy ISDN/analogue to VoIP can take up to 30 working days, depending on the losing provider.

Plan your porting in phases. Port the headquarters first, verify everything works perfectly, then port branch offices one at a time. This limits the blast radius if anything goes wrong and gives your team confidence in the process. Always maintain temporary call forwarding from old numbers during the porting window to ensure no calls are missed.

Disaster Recovery Across Sites

Multi-site VoIP inherently provides better disaster recovery than single-site deployments — but only if you design for resilience from the outset. A well-architected multi-site VoIP system should survive the complete failure of any single site without any calls being lost or callers being aware of a problem.

Geographic Redundancy

The fundamental advantage of operating across multiple sites is that you already have geographic diversity. If your London office suffers a fire, flood, or prolonged power outage, calls can be automatically rerouted to Manchester, Birmingham, or Edinburgh within seconds. The VoIP platform detects that London endpoints are unregistered and triggers failover rules that redirect incoming calls to surviving sites.

For this to work, you need pre-configured failover routes for every inbound number and hunt group. Do not wait for a disaster to figure out where calls should go — document and test the failover plan quarterly.

WAN Redundancy

Each branch should have at least two independent internet connections from different providers, ideally using different last-mile technologies (e.g., FTTP from one provider and 4G/5G from another). SD-WAN makes managing dual connections straightforward, automatically failing over voice traffic if the primary link degrades.

For critical sites like headquarters or contact centres, consider a leased line plus FTTP plus 4G triple-redundancy configuration. The leased line provides guaranteed bandwidth and SLA-backed uptime, FTTP provides high-capacity backup, and 4G serves as a last-resort connection that works even if the physical cable infrastructure is damaged.

Cloud Provider Redundancy

If you are using a hosted VoIP platform, ask your provider about their data centre redundancy. The best UK VoIP providers operate from at least two geographically separated data centres (typically London and a northern UK location like Manchester or Leeds), with automatic failover between them. Your service should continue seamlessly even if an entire data centre goes offline.

Disaster Recovery Readiness Score

Step-by-Step Multi-Site VoIP Deployment Plan

Deploying VoIP across multiple sites is a structured process that, when followed carefully, minimises disruption and delivers a working unified system in four to eight weeks depending on the number of locations.

Phase 1: Discovery and Design (Weeks 1–2)

1. Audit every site — document existing phone systems, line counts, call volumes, number inventories, and any analogue devices (fax machines, alarms, franking machines) connected to phone lines
2. Assess network readiness per site — run bandwidth, latency, jitter, and packet loss tests at each location during peak hours. Identify sites needing broadband upgrades
3. Design the architecture — choose centralised vs distributed based on your site count, connectivity quality, and resilience requirements
4. Plan the numbering scheme — design extension ranges per site, DDI allocations, and main number routing logic
5. Define failover rules — document where calls route if each site becomes unreachable

Phase 2: Infrastructure Preparation (Weeks 2–4)

1. Upgrade broadband where needed — order FTTP, leased lines, or 4G backup circuits for under-provisioned sites. Lead times can be 2–6 weeks
2. Deploy SD-WAN or configure QoS — install SD-WAN appliances or configure existing routers with QoS policies to prioritise voice traffic
3. Provision the VoIP platform — set up the centralised system, create site configurations, user accounts, and extension assignments
4. Order and pre-configure hardware — IP handsets, conference phones, headsets, and any local survivability gateways

Phase 3: Phased Rollout (Weeks 4–6)

1. Pilot at headquarters — deploy VoIP at your main site first, running in parallel with the legacy system. Iron out any issues before touching branch offices
2. Roll out to branches one at a time — start with your second-largest or most technically capable branch, then work through remaining sites. Allow 2–3 days per small branch, up to a week for larger offices
3. Port numbers progressively — begin the number porting process for each site as you deploy, with temporary forwarding in place during the porting window
4. Configure cross-site features — once two or more sites are live, set up inter-site hunt groups, shared auto-attendants, and cross-site call routing rules

Phase 4: Testing, Training, and Optimisation (Weeks 6–8)

1. End-to-end testing — test inter-site extension dialling, cross-site hunt groups, failover scenarios, and emergency (999) calling from every site
2. Staff training — conduct on-site or video training sessions for each branch, tailored to their specific call handling needs
3. Decommission legacy systems — once all numbers are ported and testing is complete, cancel old phone lines and return legacy equipment
4. Baseline and optimise — establish performance baselines for call quality at each site, refine QoS settings, and tune auto-attendant menus based on early usage data

Cost Breakdown for a UK Multi-Site VoIP Deployment

Understanding the full cost picture helps you build an accurate business case. Here is a realistic breakdown for a UK business deploying VoIP across five branch offices with a total of 100 users.

Three-Year Total Cost of Ownership

The savings from consolidating five separate legacy phone systems into a single hosted VoIP platform are substantial — typically £30,000–£50,000 over three years for a 100-user deployment. The savings come from eliminating duplicate line rentals, maintenance contracts, and hardware refresh cycles at each site, combined with lower per-minute call costs and reduced IT management overhead.

Cost Component	One-Off	Monthly (all sites)
VoIP platform (100 users @ £15/user)	—	£1,500
SD-WAN (5 sites)	£2,500 (appliances)	£750
IP handsets (100 @ £120 avg)	£12,000	—
Network switches/upgrades	£3,000	—
Number porting (50 numbers)	£500	—
Installation and configuration	£4,000	—
Training	£1,500	—
Total	£23,500	£2,250

Common Pitfalls and How to Avoid Them

Multi-site VoIP deployments can be transformative, but several common mistakes trip up even experienced IT teams. Here are the pitfalls we see most frequently and how to avoid them.

1. Deploying all sites simultaneously — always roll out in phases, starting with your headquarters. Trying to switch five offices at once is a recipe for chaos if anything goes wrong
2. Ignoring bandwidth at smaller branches — it is tempting to focus capacity planning on your largest sites, but a small rural branch with 10 Mbps ADSL will be the first to suffer call quality issues. Every site needs assessment
3. Not configuring QoS — without Quality of Service policies, a large file download at any site can consume all available bandwidth and make voice calls unusable. QoS must be configured on every router and switch in the path
4. Forgetting emergency services configuration — each site must relay its correct physical address for 999 calls. If all calls route through a central platform, the system must identify which site the call originated from and pass the correct address
5. Skipping failover testing — it is not enough to configure failover rules. You must physically test them by disconnecting WAN links and verifying calls route correctly. Test quarterly, not just at go-live
6. Underestimating training needs — a multi-site VoIP system has features staff have never used before (inter-site transfers, cross-site pickup, presence, conferencing). Budget time and resources for proper training at every location
7. Poor number porting coordination — porting numbers from multiple legacy providers simultaneously can create gaps in service. Port in phases, maintain temporary forwarding, and always have a rollback plan

Why Cloudswitched for Multi-Site VoIP

Deploying VoIP across multiple branch offices is not a commodity IT project — it requires expertise in network architecture, WAN design, telephony engineering, and project management. At Cloudswitched, we have designed and deployed multi-site VoIP for UK businesses ranging from three-office professional services firms to nationwide retail chains with 50+ locations.

• End-to-end project management — from initial site surveys and network assessments to phased rollout and post-deployment optimisation
• Network design expertise — we design the WAN connectivity (MPLS, SD-WAN, or hybrid) alongside the VoIP platform, ensuring they work together seamlessly
• Vendor-neutral advice — we recommend the right platform for your specific needs, whether that is a cloud-hosted multi-tenant solution or a dedicated on-premise system
• UK-wide deployment capability — our engineering team deploys on-site across the UK, from central London to the Scottish Highlands
• Ongoing managed service — 24/7 monitoring, proactive issue resolution, and quarterly performance reviews across all your sites
• PSTN switch-off readiness — we audit every site for legacy dependencies and ensure full migration before the January 2027 deadline