Cost of Downtime Calculator

Lost revenue, plus anything else an hour of outage costs you.

Formatting only, no conversion.

One incident

What a single outage of this length costs.

minutes
$5,000
1h at $5,000 per hour

What each availability tier costs

Every target permits some downtime. This is what that allowance is worth per year, and what buying the next tier would save you.

TargetDowntime allowedCostSaved vs tier below
99%two nines3d 15h$438,300
99.5%1d 19h$219,150$219,150
99.9%three nines8h 45m$43,830$175,320
99.95%three and a half nines4h 22m$21,915$21,915
99.99%four nines52m 36s$4,383$17,532
99.999%five nines5m 16s$438$3,945

The right way to read this column: the saving is your budget for the engineering that buys the extra nine. If reaching 99.99% costs more per year than the number beside it, the target is not worth having.

What this leaves out

  • Reputation and churn, which outlast the incident.
  • SLA credits you owe customers, often a multiple of the lost revenue.
  • Engineering hours spent on the incident and the follow-up.
  • Outages rarely land at quiet hours. Weighting by traffic usually raises the figure.

Every one of these pushes the number up, so treat the result as a floor rather than a forecast.

Availability targets are budget decisions. This turns each one into a number you can put in front of a finance team.

Downtime cost FAQs

How do I calculate the cost of downtime?

Multiply the hourly cost of an outage by its duration in hours. The hourly figure is usually lost revenue, which you can derive by dividing annual revenue by 8,766 hours. That gives an even spread, so weight it upward if your traffic concentrates in business hours.

How much downtime does each availability target allow?

Per year: 99% allows about 3.65 days, 99.9% about 8.8 hours, 99.95% about 4.4 hours, 99.99% about 53 minutes, and 99.999% about 5 minutes. Each additional nine cuts the allowance to a tenth, and typically costs considerably more than a tenth to achieve.

Is the next nine worth buying?

Compare the saving column against what the engineering would cost. Moving from 99.9% to 99.99% removes about 90% of the allowed downtime, so it saves about 90% of what that tier costs you. If multi-region failover, on-call rotation and the testing to prove it all works cost more per year than that saving, the target is not worth having.

What does this calculation leave out?

Reputation damage and churn, SLA credits owed to customers, engineering hours spent on the incident and the follow-up, and the fact that outages rarely land at quiet hours. All push the number up, so treat the result as a floor rather than a forecast.

Which period should I use?

Whichever matches the conversation. Annual figures make the case for investment; monthly figures fit an error budget review. A month here is 30.44 days, one twelfth of an astronomical year, so it agrees with the annual figure rather than assuming 30 days.

Is anything sent to a server?

No. The arithmetic runs in your browser and the page works offline, which matters when the inputs are your revenue figures.