Customer Impact Quantification
Most postmortems undercount customer impact. The quantification that produces honest numbers.
Count affected
Customer-impact quantification starts with unique users, not raw requests. One user retrying 100 times is one impact, not 100. Users is the severity calibrator that matches customer pain; segment cuts (enterprise versus free) reveal the parts of the customer base that carry different stakes.
- Unique users affected. Distinct affected-user count per incident. Pulled from real data, not estimates.
- Users, not requests. The unit that matters. One user with 100 retries is one impact.
- User count over request count. User-driven severity calibration. Better proxy for customer pain than raw request counts.
- Per-segment cut. Impacted-customer split by tier. Enterprise versus free carry different stakes.
Per-user duration
Per-user duration captures how long each user actually felt the failure. Aggregate incident-window time hides the long-tail user pain that drives churn. Median plus p95 covers the typical and the worst cases together.
- Time-in-failure per user. Captures the actual experience, not just the incident window.
- Average and median. Both stats per incident. Median surfaces typical pain; average catches outliers.
- p95 user duration. Long-tail metric per incident. Catches users who felt the outage for hours.
- User-by-region cut. Duration split per region. Supports regional comms decisions.
Revenue impact
Revenue translation grounds impact in business terms. Lost transactions times average value pulls from billing data; correlated churn surfaces the long-tail business cost. SLA-credit estimates per customer support honest customer communication.
- Lost transactions times average value. Missing-revenue arithmetic per incident. From billing data, not estimates.
- Customer churn correlation. Churn-uplift per incident where measurable. Surfaces the long-tail business cost.
- SLA-credit estimate. Contractual credit owed per customer. Supports honest customer comms.
- Quarterly impact-trend chart. Revenue-impact trajectory per quarter. Catches recurring patterns.