Usage Metering & Billing Pipeline · View 05 of 22 · People and journeys
The finding
- The sentiment line is worst at the start, not the end: a figure found wrong after issue is a correction with a customer attached, and one found before issue is just a draft. That is the entire argument for the anomaly gate in view 13.
- Confidence recovers at Scope and Replay because both are bounded by construction — an unbounded restatement is refused, and a replay cannot mutate anything.
What this forces
- Replay must use the same code path as the original metering, or the restated figure is a second opinion rather than a correction.
- Both figures survive, with reason and authorising identity, so the history of a dispute is itself auditable.
- The remedy choice — credit memo or superseding invoice — is recorded per correction, because the policy varies by case and auditors ask which was used.
Risks
- Scope creep under pressure: the temptation at 2 a.m. is to restate every account rather than the one that is wrong. The scope guard is a control, not a convenience.
- Evidence that has aged into the cold tier turns a 30-minute diagnosis into a retrieval wait. The 400-day hot window is sized for the dispute window plus an annual cycle, and nothing longer.