Every IT vendor brochure has a number with some nines in it. Most of the time, no one in the room knows what the number actually buys them. So let's do the maths.
How much downtime each nine actually allows
99% uptime is 3 days 15 hours of downtime a year.
99.9% is 8 hours 45 minutes.
99.99% is 52 minutes.
99.999% is 5 minutes.
A "three-nines" SLA sounds bulletproof until you realise it permits the equivalent of a full working day offline, spread across the year. For most small businesses, that's actually fine.
What the SLA doesn't cover
Read the fine print. Most SLAs exclude scheduled maintenance windows (often 4 hours per week), regional internet outages, acts of god, attacks, and any time the customer caused the issue. Your "99.9% uptime" can quietly become 99% if you count everything.
A good provider will tell you the real number from the last 12 months on request.
How much each extra nine costs
Going from 99% to 99.9% typically adds 30–40% to your IT cost. Going to 99.99% can double it. For most Geelong businesses, that money is far better spent on backup and recovery — which lets you bounce back from anything, not just the things the SLA covers.
For what it's worth: our managed-service customers run at around 99.92% real-world uptime, measured across the last calendar year.
What "uptime" really measures
Uptime is the percentage of time a system is available and working. It sounds simple, but the number hides a lot. A provider quoting 99.9% is promising no more than about 8 hours and 45 minutes of unplanned downtime across an entire year. Drop to 99% and you are allowing more than three and a half days. Climb to 99.99% and you are down to under an hour. Each extra nine costs real money in redundancy, monitoring and engineering, so the right target is the one that matches what downtime actually costs your business.
Why the contract wording matters more than the number
Most service level agreements quietly exclude scheduled maintenance windows, regional internet outages, power failures and anything the customer caused. That means a headline figure of 99.9% can become far lower in practice once those carve-outs are applied. When you read an SLA, look for how downtime is defined, how it is measured, what is excluded, and what you actually receive if the target is missed. A credit of one month of fees rarely covers the cost of a day of lost productivity, so the prevention matters far more than the compensation.
What we do to protect uptime
Real-world uptime comes from boring, consistent work: 24/7 monitoring that alerts us before you notice a problem, patching that closes vulnerabilities without breaking things, tested backups that let us recover quickly, and redundancy where it counts. Our managed clients run at around 99.9% real-world availability measured across the year, and we report the real figure rather than the marketing one. If a system is critical to your revenue, we design the recovery plan around how long you can genuinely afford to be without it.
Matching the target to your business
The honest answer to "how much uptime do I need" is "it depends on what downtime costs you". For some businesses an hour offline is a minor irritation; for others it is lost revenue and missed deadlines by the minute. Rather than sell the highest number, we work out what an outage actually costs you and design the monitoring, redundancy and recovery to match. That way you pay for the resilience you genuinely need, and not for nines that look impressive on paper but make no difference to your day.