What an hour of downtime actually costs

Every decision about hosting, monitoring and maintenance is implicitly a bet about what an outage costs. Almost nobody has calculated that number, which means the bet gets made on feeling — and feelings run in both directions. Some businesses over-insure a site that would barely notice an outage; others run a revenue-critical store on the cheapest hosting available.

The three layers of loss

Direct loss is the one everybody thinks of and the smallest of the three.

The permanent layer is the one that matters most and resists measurement hardest, which is exactly why it gets left out. Somebody who tried to buy from you, found nothing there, and bought elsewhere does not appear in any report. They appear as a slightly lower number next month, indistinguishable from noise.

Timing changes everything

An hourly average is a useful starting figure and a poor planning tool, because outages are not evenly distributed in consequence.

• An hour at three in the morning in your main market may cost close to nothing.
• An hour during your busiest trading window can cost many multiples of the average.
• An hour while a paid campaign is running costs the sales and the ad spend, since you pay for clicks that land nowhere.
• An hour on the day of a launch, a sale or an event costs the outage plus everything spent driving people to it.

What the number is for

Once you have an honest hourly figure, a set of previously vague decisions become arithmetic.

1. Is better hosting worth it? Compare the annual difference to the hours of downtime it plausibly prevents.
2. Is monitoring worth it? Monitoring does not prevent outages; it shortens them. Value it as hours saved per incident.
3. Is a maintenance plan expensive? If a single hour offline costs more than the year's plan, it is not.
4. Is a backup strategy sufficient? Restore time multiplied by hourly cost is the real exposure, which is why an untested backup is so dangerous — its restore time is unknown.

The second point is worth dwelling on. Monitoring is frequently sold as protection and it is nothing of the sort. Its entire value is in the gap between when something breaks and when a human finds out, which for a site without monitoring is often measured in customer complaints.

When the answer is that it does not matter much

For a good number of brochure sites, an hour offline genuinely costs very little. A handful of visitors, most of whom will return, and no transactions to lose. That is a legitimate and useful conclusion.

It means you can stop paying for reliability you do not need, and redirect that budget to something that does move the business. The calculation is worth doing precisely because it can come back low — knowing it is low is what lets you spend the money elsewhere with confidence rather than anxiety.