How the cost is calculated
An outage costs money in three ways: sales that do not happen, employees who cannot work normally, and the extra effort needed to recover. The calculator adds them up:
Lost revenue per hour = annual revenue × % that depends on the system
÷ operating hours per year
Lost productivity per hour = affected employees × cost per employee per hour
× % of productivity lost
Cost per hour = lost revenue per hour + lost productivity per hour
Cost per incident = cost per hour × outage duration in hours + recovery costsThe cost per minute is the cost per hour ÷ 60. Recovery costs are counted once per incident, so they are not part of the cost per hour. The chart shows how the incident’s total splits into lost revenue, lost productivity and recovery.
Worked example
A SaaS company earns $12,000,000 a year, and 70% of that revenue depends on its platform being available (the rest comes from long-term contracts and services that an outage does not stop). The platform runs 24/7 (8,760 hours a year). During a 90-minute outage, 40 employees in support, sales and engineering, at a loaded cost of $55 per hour, lose 60% of their productivity. Recovery costs $3,000 in overtime and SLA credits.
- Lost revenue per hour: $12,000,000 × 0.70 ÷ 8,760 = $958.90
- Lost productivity per hour: 40 × $55 × 0.60 = $1,320
- Cost per hour: $2,279 ($37.98 per minute)
- Lost revenue in 1.5 hours: $1,438
- Lost productivity in 1.5 hours: $1,980
- Recovery: $3,000
- Cost of the incident: $6,418
Amounts are rounded exactly as the calculator shows them; each step uses the unrounded values.
In this example, lost productivity is larger than lost revenue, and recovery is the single biggest item. Both are easy to forget when people talk about “lost sales”.
Cost by outage duration
With the same assumptions, this is how the cost of a single incident grows with its length:
| Outage duration | Lost revenue + productivity | Recovery | Cost of the incident |
|---|---|---|---|
| 15 minutes | $569.73 | $3,000 | $3,570 |
| 1 hour | $2,279 | $3,000 | $5,279 |
| 4 hours | $9,116 | $3,000 | $12,116 |
| 8 hours | $18,231 | $3,000 | $21,231 |
| 24 hours | $54,694 | $3,000 | $57,694 |
Short incidents are dominated by fixed recovery costs; long ones by lost revenue and productivity. Cutting the time to detect and fix a problem pays off quickly.
Tips to reduce the cost of downtime
- Remove single points of failure. Run critical services across several availability zones and keep spare capacity for the parts that cannot fail.
- Detect problems before your customers do. Monitoring and alerting on what users actually experience (errors, latency, checkouts) shortens every incident.
- Prepare runbooks and an on-call rotation. A clear, rehearsed response is faster than improvising at 3 a.m.
- Deploy safely. Many outages come from changes. Gradual rollouts, feature flags and one-click rollbacks limit the damage of a bad release.
- Test your backups and restores. A backup that has never been restored is a guess, not a plan.
- Communicate during the outage. A status page and quick updates reduce support load and protect customer trust.
- Learn from every incident. Blameless post-mortems with concrete follow-up actions prevent the same failure from happening twice.
What this estimate does not include
Long-term effects are not included: customers who leave for good, damage to your reputation, contractual penalties beyond the recovery costs you enter, regulatory fines and the cost of delayed projects. On the other hand, some lost sales may only be delayed. Use the result as an order of magnitude to prioritize investments in reliability, not as an exact figure.
Frequently asked questions
How do you calculate the cost of downtime?
Add the revenue lost per hour (annual revenue × share that depends on the system ÷ operating hours) to the productivity lost per hour (affected employees × hourly cost × share of productivity lost). Multiply by the outage duration in hours and add the one-off recovery costs of the incident.
What counts as recovery costs?
Any extra spending caused by the incident itself, such as overtime for the on-call team, external support or consultants, credits paid to customers under an SLA, data recovery and emergency hardware or cloud capacity.
Is all revenue during an outage really lost?
Not always. Some customers simply come back later and buy anyway, while others go to a competitor. If you know that part of your sales is only delayed, lower the percentage of revenue that depends on the system to reflect only the sales you would truly lose.
Which operating hours should I use?
Use 8,760 hours (24 × 365) for services that earn money around the clock, such as online stores or SaaS platforms. If the system only generates revenue during business hours, use about 2,080 hours (40 hours × 52 weeks), so that each hour carries its real share of revenue.
Is the data I enter stored anywhere?
No. All calculations run in your browser. Values only leave your device if you use the share button, which puts them in the link you copy.