When the NBN goes down: keeping your business running

Most Australian offices treat the NBN like electricity—until it isn’t there. A dropout isn’t just an annoyance for someone refreshing their inbox. It stalls EFTPOS, breaks cloud apps, kills VoIP calls, and sends staff into a holding pattern that still costs you money. The NBN is better than it was, but it is not infallible—and betting the whole business on one line is a plan you only notice when it fails.

This guide is a practical look at what outages cost, what you can do before the next one, and how options scale from a small Newcastle office to a business that genuinely cannot tolerate an hour offline. Layer 3 helps clients design redundancy that matches real impact—not every firm needs Starlink on day one, but every firm should know what happens when the link drops.

Quick answers

Why do NBN outages hurt Australian businesses so much?

Most workflows assume always-on internet: cloud email, accounting, CRM, EFTPOS, remote access, and voice over IP all stop when the link dies. Unlike a single app fault, a WAN outage is site-wide—everyone is affected at once.

What does an hour offline typically cost?

It varies by industry, but think in lost sales, idle wages, missed calls, delayed dispatch, and reputational damage—not just the NBN ticket you raised. Even a modest team can burn hundreds to thousands of dollars per hour when trading or production depends on connectivity.

What is the simplest first step for redundancy?

Document what must keep working during an outage, then add automatic failover—usually a 4G or 5G backup modem that takes over when the primary NBN link fails. Test it quarterly so you are not discovering misconfiguration during a real outage.

Can staff work from home if the office NBN is down?

Often yes—if you planned for it. That means laptops, VPN or secure remote access, cloud apps, phone diversion or softphone, and staff who know the drill. Sending people home without those pieces in place just moves the problem to households that may also rely on the same NBN infrastructure.

When is a second fixed connection worth it?

When the financial impact of any outage exceeds the cost of a redundant link—common for multi-site firms, heavy EFTPOS/trading floors, or teams that cannot pause for half a day. Options include a second NBN from another provider, fixed wireless, or Starlink as diverse path—not a duplicate of the same failure mode.

The real cost of “the internet’s down”

Outages are usually measured in minutes on a provider status page and in hours in your P&L. Worth listing explicitly for your business:

• Lost sales—EFTPOS, e-commerce, quotes that cannot be sent, appointments not booked.
• Staff productivity—payroll still runs while output stops.
• Customer experience—unanswered calls, “our systems are down” messages, social complaints.
• Operational knock-ons—missed cut-offs, delayed invoicing, supply chain pings that fail silently.
• Recovery time—catching up after the link returns often takes longer than the outage itself.

Once you put a rough dollar figure on an hour offline, redundancy stops feeling like an optional extra and starts looking like insurance with a known premium.

Why one NBN line is a single point of failure

The NBN is national infrastructure—and local reality. Fibre, FTTN, HFC, and fixed wireless all have different failure modes: street works, weather, provider backhaul, power at the node, faulty CPE, or a fault in the building. Your RSP can be excellent and still helpless when the underlying path is broken.

Redundancy is not about blaming NBN Co—it is about accepting that any single path can fail and deciding what your business does in the meantime.

Options that scale with your business

Micro and small offices (roughly 1–15 staff)

Start with a written outage plan and a tested 4G/5G failover router. Keep critical traffic alive: EFTPOS, email, phones, and remote access for key staff. For many firms this is enough—if failover is automatic and someone checks it actually works.

• 4G/5G backup with business-grade data allowance (consumer SIMs often throttle or run out).
• Cloud-first apps and MFA already in place—no “everything lives on the server in the cupboard”.
• Simple runbook: who declares an outage, who switches to mobile hotspots if failover fails, who updates customers.
• Optional: one mobile hotspot in the drawer as a last-resort spare.

Growing SMBs with higher daily revenue at risk

Add diversity, not just backup. A 5G modem that shares the same street cabinet as your NBN may fail alongside it. Consider a second path—different technology or provider—plus clearer work-from-home playbooks.

• Dual-WAN router with automatic failover and alerting when on backup.
• Second connection: Starlink, fixed wireless, or NBN via a different RSP/path where physically diverse.
• VoIP call forwarding to mobiles or a secondary site when primary voice fails.
• Priority list of which systems must survive vs which can wait.

Businesses that cannot tolerate hours offline

Design for continuity, not hope. Redundant links, monitored failover, regular failover tests, and documented RTO-style targets for connectivity—not just for servers. Pair network design with backup and recovery planning so data is protected even when staff relocate.

• Monitored dual links with SLA-backed RSPs and documented escalation.
• Secondary site or co-working fallback for critical teams.
• Quarterly “pull the primary cable” tests with logged results.

Plan B: send staff home—only if you planned Plan B

When the office link is dead but homes still have working internet, sending people home can keep revenue and service moving. It fails embarrassingly often when nobody rehearsed it. Treat work-from-home during an outage as a remote support scenario that must work without a desk visit first.

• Remote access: VPN, Entra ID–joined laptops, or secure zero-trust access—not ad hoc TeamViewer on personal PCs.
• Permissions and data: staff can reach what they need without exposing the whole network.
• Phones: divert extensions to mobile or softphone apps; test call quality and caller ID.
• Collaboration: Teams/Zoom already licensed and familiar—not first-use during a crisis.
• Staff training: how to take customer calls from home, mute properly, use noise-cancelling headsets so barking dogs and school pickup do not land on client calls.
• Security: no printing sensitive docs on home printers; brief people on phishing spikes during outages.

Run a 30-minute tabletop once a year: “NBN is down until 2pm—what do we do?” Fix the gaps while coffee is still hot.

4G and 5G failover—what good looks like

A backup modem in a box is not a strategy. Good failover is automatic, monitored, and sized for your critical apps—not the whole office streaming video in HD.

How 5G fits into a typical office network

Most SMB failover setups place cellular backup on the WAN side of your router or firewall—not as a separate Wi‑Fi network staff join manually. When the NBN link drops, the router should move default routing to the 5G modem automatically; EFTPOS, VoIP, and key PCs keep using the LAN while traffic exits via the backup path. Larger uploads can wait until primary returns.

The diagram below is a simplified reference layout: NBN as primary WAN, 5G modem on a secondary WAN port, and your office LAN unchanged on the inside. Exact hardware, port labels, and VLANs vary by site—use this to align conversation with your provider before an outage, not as a substitute for a site-specific design.

Click to enlarge

• Business-grade router with dual SIM or external modem and health checks.
• Failover triggers on WAN loss, not manual USB tethering panic.
• Alerts when you are running on backup so you know to avoid large uploads.
• Data plan sized for EFTPOS + email + voice, with headroom for a full day if needed.
• Quarterly test: disable NBN WAN and confirm EFTPOS or a key app still flows.

Second fixed connections: Starlink, second NBN, and diversity

Failover cellular buys time; a second fixed link buys resilience when outages run long or cellular is congested. The goal is path diversity—different technology, entry point, or provider—not two labels on the same fragile segment.

• Starlink: useful diverse path where install and obstructions suit; plan for power and weather.
• Second NBN service: only valuable if physically diverse—same trench, same outcome.
• Fixed wireless or enterprise Ethernet: worth comparing where available.
• Load-balancing vs active/passive: most SMBs want passive failover first; balancing adds complexity.

Match spend to impact. A sole trader may live on 5G backup; a busy clinic with EFTPOS and cloud records may need dual fixed links plus cellular as tertiary.

How Layer 3 helps

We map what your business actually loses per hour offline, then recommend proven redundancy—not vendor buzzwords. That might be a configured 5G failover router for one client and dual-WAN plus Starlink for another. We document failover, test it with you, align remote access and phones with the plan, and support the stack when something breaks at the worst time.

Based in Newcastle, we design for Australian NBN reality: mixed technology types, RSP handoffs, and teams that need practical runbooks—not a 40-page policy nobody reads.

FAQ

Is mobile hotspot enough instead of a failover router?

For one person, maybe—for a site, rarely. Hotspots are manual, easy to misconfigure, and weak on EFTPOS and VoIP stability. A proper failover router is designed to detect WAN loss and switch automatically.

Will Starlink replace NBN for our office?

For some rural or reliability-sensitive sites it is primary or secondary—but it is not a drop-in for every business. Latency, weather, and install matter. We usually position Starlink as diverse backup or primary only where NBN is persistently poor.

Does backup internet slow our normal connection?

A well-configured passive failover should not affect day-to-day speed. Active load-balancing is different—useful but more complex to support.

Who should declare a connectivity outage internally?

Name a role, not a vague “IT person”—office manager, ops lead, or MSP contact. They trigger the runbook: failover check, customer comms, work-from-home decision, provider ticket.

How often should we test failover?

At least quarterly for businesses that rely on EFTPOS or cloud ops; twice a year minimum for everyone else. Log the result like a fire drill—pass/fail and fixes assigned.

Can Layer 3 manage our routers and failover modems?

Yes—as part of managed IT we monitor links, patch firmware, and respond when backup paths fail or alerts show you have been on cellular longer than expected.

Discuss outage impact for your site: Contact Layer 3.