Generic cost-per-minute figures are rarely useful for a specific company. A defensible downtime estimate starts with your transaction volume, staffing, contracts, and customer impact.
Direct transaction impact
For a revenue-producing workflow, start with the transactions that could not complete during the incident. Separate delayed transactions from permanently lost ones.
A simple estimate is:
lost contribution = failed transactions × average contribution per transaction
Use contribution rather than gross revenue when the decision requires an economic estimate. Record seasonal or time-of-day assumptions, and do not assume that every interrupted transaction is lost.
Response and recovery work
Add the time spent by engineering, support, operations, communications, and management. Use the company's normal loaded labor rate if available. Include follow-up work only when it was caused by the incident.
Useful time buckets are detection and triage, mitigation, recovery verification, customer support, incident review, and corrective work.
Contractual costs
Review customer SLAs and vendor contracts separately. A customer credit your company owes is not necessarily offset by a credit from the failed vendor. Eligibility, exclusions, measurement windows, and claim procedures differ by contract.
The SLA calculator converts a percentage into an allowed-time budget; it does not determine whether a contract was breached.
Customer and operational effects
Some effects should be recorded even when a reliable dollar value is not available:
- support volume and backlog
- delayed shipments or processing
- missed internal deadlines
- customers requiring individual follow-up
- sales or renewal conversations affected by the incident
- regulatory or contractual reporting work
Avoid assigning a universal churn percentage or reputation value without company evidence. Describe the observed effect and keep uncertain estimates labeled.
Third-party outages
When an external provider is the root cause, measure the effect on your service in the same way as an internal failure. Record when your application first showed impact, when the vendor acknowledged the issue, which fallback was used, and when customer service returned to normal.
A repeatable worksheet
A range with clear assumptions is more useful than a dramatic industry average. Use the estimate to prioritize recovery work, not to manufacture precision.