Downtime Cost Calculator

Revenue per hour times the outage length, plus staff who cannot work, plus the engineering hours spent recovering, plus any service credits owed. The last three are invisible on any report and often exceed the first.

Sometimes deferred rather than lost. A customer who could not buy at 11am may buy at 2pm, which makes the revenue line an overstatement — unless they bought from a competitor instead, in which case it is not.

Published figures range from hundreds to tens of thousands per minute and they average over businesses nothing like each other. Your own revenue per trading hour is a far better starting point than any benchmark.

Yes. They are being paid regardless and the output is not recoverable, which makes it a real cost even though no invoice records it. It is frequently the largest single line for an internal system.

Through the annualised figure. One incident rarely justifies infrastructure; four a year frequently does, and one more nine of availability removes roughly 90% of the downtime and therefore 90% of the cost.

Real and genuinely hard to quantify, so it is deliberately not modelled here. Treat the calculated figure as a floor rather than a total, particularly for a customer-facing outage that lasted long enough to be noticed publicly.

Annual revenue divided by trading hours, and be honest about which hours those are. An always-on service uses 8,760; a business that only transacts in office hours uses far fewer and has a much higher hourly figure.

Considerably. An outage at peak trading costs a multiple of one at the quietest hour, so an annual average understates a peak-time incident and overstates an overnight one.

Yes, and count the opportunity cost too. Engineers responding to an incident are not doing what they were going to do, and on a long incident that displaced work is a real part of the bill.

Usually a percentage of the monthly fee, applied on request rather than automatically, and capped well below what the outage cost the customer. They are a gesture rather than compensation.

Not in this figure, and frequently larger than your own. For a business-to-business service it is also what drives churn after an incident, which is a cost that arrives months later.

Compare the annualised downtime cost against the annual cost of the redundancy. Sometimes accepting the outage is genuinely cheaper, and saying so is a legitimate engineering position rather than a failure.