Incident Downtime Cost Calculator

The four buckets

Lost revenue is the obvious one and usually not the largest. Staff cost covers everyone pulled into the incident plus everyone unable to work during it. Recovery cost includes overtime, data reconstruction, vendor escalation and the backlog cleared afterwards. Then there is the residual: customers who left, credits owed under SLAs, and any regulatory penalty. Estimates that count only lost revenue typically capture less than half the total.

Revenue loss is not linear with time

A five-minute outage during a quiet hour may lose almost nothing, because most transactions are simply retried. A five-minute outage during a peak sales event loses the whole window and some of the trust behind it. Averaging annual revenue across the year to get a per-minute figure understates peak incidents badly and overstates quiet ones, which is why a single blended number is a poor input to any decision.

Staff cost compounds quietly

Twenty engineers idle for two hours is forty person-hours at fully loaded cost. The incident response team's time is on top. The genuinely expensive part is the aftermath: interrupted work does not resume where it stopped, context has to be rebuilt, and a major incident routinely costs the following day as well. Counting only the outage window misses that.

Reputational cost is real and hard to size

Customers who churn after an outage take their lifetime value with them, and a public incident affects deals in progress that nobody attributes to it. Rather than inventing a number, model a churn range — even a fraction of a percent of the base, multiplied by average lifetime value, usually exceeds the direct costs and makes the point without false precision.

Use it to size prevention, not to assign blame

The productive use is comparing the expected annual cost of downtime against the cost of reducing it. If outages cost 200,000 a year and multi-region redundancy costs 80,000 a year while removing most of that exposure, the decision is straightforward. If the same redundancy costs 400,000, accepting the risk is the rational choice — and saying so explicitly is more valuable than pretending the risk is not there.

MTTR is the lever you actually control

Preventing every incident is impossible; shortening them is not. Cost scales with duration, so investment in detection, alerting, runbooks and rollback capability often returns more than investment in preventing failure. An organisation that detects in two minutes and rolls back in five has a different cost profile from one that finds out from customers an hour later, with identical underlying reliability.