The Revenue-Impact Incident Classifier

Some incidents matter more than others to the business. The classifier that maps incident impact to revenue, and the conversations it enables.

The model

The classifier has two dimensions: per-service revenue per minute (calculated from traffic, conversion, and average order value), and per-failure-mode percentage affected. Cross them and the headline output is dollars per minute of incident.

• Per-service revenue per minute. Calculated rate per service from traffic, conversion, average order value. The foundation number.
• Per-failure-mode percent affected. Impact share per mode. Total outage 100 percent; partial degradation 30-60 percent.
• Cross the two. Dollars-per-minute composite per incident. The headline number for stakeholder conversations.
• Documented assumptions per service. Captured input data per service. Supports later “where did this number come from” reviews.

How to apply

Apply the classifier at postmortem time. Per-incident dollar figure goes in the postmortem; quarterly aggregation drives reliability-investment decisions because leadership conversations need numbers, not narratives.

• Per-incident dollar figure. Post-resolution number per incident. “This 45-minute outage cost about $120k.”
• Figure in the postmortem. Visible cost line per postmortem. Stakeholders see the cost; investment becomes rational.
• Quarterly aggregation. Total-incident-cost versus reliability-investment view per quarter. ROI becomes visible.
• Named author per incident. Responsible analyst per incident. Honest accounting.

Avoid

Three failure modes destroy the classifier. False precision (pretending estimates are exact); metric optimisation (treating the proxy as the target); public shaming (using the number to blame teams instead of inform leadership).

• False precision. Range-not-exact framing per figure. Estimates are estimates; do not pretend otherwise.
• Optimising for the metric. Metric-versus-reliability rule per team. The metric is a proxy, not a target.
• Public shaming. No-team-shaming rule per team. Revenue impact is a leadership conversation.
• Documented use policy. Published “how we use this number” doc per org. Catches misuse.