Small Business Internet Uptime, SLAs & Redundancy: What to Demand

A small business can lose transactions, calls, and staff time during an Internet outage. A service-level agreement (SLA) describes a provider's stated performance commitment and remedy; a backup connection helps keep work running. Compare both before signing for a business circuit.

Start with your real tolerance for downtime. A point-of-sale system may need a working backup path during business hours, while a less time-sensitive office may accept a lower-cost plan. A high advertised speed is not a substitute for a clear repair process.

What uptime percentages mean

Over a full 365-day year, a 99% uptime target allows about 87.6 hours of downtime; 99.9% allows about 8.8 hours; 99.99% allows about 52.6 minutes. These are arithmetic illustrations, not predictions for a particular provider. Actual SLA calculations may exclude planned maintenance, customer equipment failures, or other events. Read the contract's measurement period and exclusions.

Choose a target based on the cost of interruption to your operation, and design a tested failover path if a single outage would be expensive. No published percentage by itself guarantees uninterrupted card payments or calls.

Learn the four terms sales will use

1. Uptime % — how consistently the circuit stays up.
2. SLA (service level agreement) — the written remedy if the provider misses that target.
3. Failover — automatic cutover to a backup path when the primary dies.
4. Redundancy — designed backup capacity (second ISP, cellular, multi-WAN) ready before the outage.

Credits on next month’s bill are not the same as recovering lost revenue. Treat SLA credits as consolation; design failover so the outage never reaches the register.

SLA red flags that should slow you down

• “Best effort” with no published numerical target
• Vague or open-ended maintenance windows that never count against uptime
• Force majeure language so broad that weather, fiber cuts, and peering issues all vanish
• Token credits that max out at a few dollars after a multi-hour Saturday outage
• Response time ≠ resolution time — “acknowledge in 4 hours” only means they noticed, not that a tech is fixing it

Ask for both acknowledgment and repair targets in writing. A published number helps only when its measurement and remedy are clear.

Failover tiers most SMBs actually buy

1. Cellular failover — a multi-WAN / LTE-5G appliance (Cradlepoint-, Peplink-class, or ISP-supplied) detects primary failure and flips to cell in seconds. Lowest cost, most common.
2. Multi-network failover — backup can roam across carriers (and sometimes nearby Wi-Fi) so one cell outage does not strand you.
3. Dual-ISP redundancy — two independent providers active (or hot-standby). Usually the highest ongoing cost; useful when the business cannot tolerate a single access-network outage.

Manual “we will tether a phone when it breaks” is not failover. By the time a human notices, the transaction already failed.

What to demand before you sign

• A published numerical uptime target for the business product you are buying
• Clear credit math, response vs resolution times, and force majeure limits
• An automatic failover option you can test during install
• Confirmation that POS, VoIP, and camera NVRs will ride the backup path without reconfiguration
• A quote that separates primary circuit, backup SIM/data, and managed router fees

A one-hour diligence script for owners

Before renewing a “business” circuit that is really a consumer plan with a different label:

1. Ask for the exact uptime percentage in the product name you are buying — not a corporate network average.
2. Request a sample SLA PDF and highlight force majeure, maintenance, and credit caps.
3. Price cellular failover hardware + data as a line item next to one more year of single-path risk.
4. Test failover during install: unplug the primary WAN and confirm POS and phones recover without a reboot ritual.
5. Document MTTR contacts (NOC chat, escalation phone) in a binder staff can find on Saturday.

That hour costs less than one bad lunch rush on a dead cable modem.

Bottom line

Shop for clear outage terms, support response, and a tested backup connection as well as speed. Select an uptime target based on your downtime tolerance, and read the SLA measurement rules and remedies before signing. Service credits cannot recover missed sales; a working failover path may help keep operations running.