War-game a health score rebuild by testing whether signal capacity, review gates, and rebuild deadlines still shrink soft weights when several account surfaces share one schema and the next renewals board is close.
Health score rebuilds fail when "approved" means a checkbox without a decision artifact, when teams dual-count a draft as both reviewed and live, and when money-adjacent signal paths still pin to one thin glossary because the map is incomplete. The review needs one written inventory of signal classes, one owner who can deny a risky weight, and a documented consequence when a failed review ages past the deadline.
Assemble the rebuild packet
List signal classes by surface, weekly weight volume, review gate, last rebuild-skip reason, owner of record, and evidence required to keep the rule. Attach open findings, overdue drafts, and any emergency weight waiver in the last quarter. Name which classes are out of scope and why. If a finance-adjacent signal class is omitted, show the residual risk in the same packet.
Decide allow, revise, or hold in advance. Include the fields that compute overdue-review status so two reviewers reach the same label from the same exports.
Claim matrix
| Claim | Evidence required | Fail condition |
|---|---|---|
| Class is staffed | Named reviewer roster and measured time-to-rebuild check | Roster is empty or still points to a single untested operator |
| Definition is complete | Schema export showing approved state for the claimed fields | Ticket closed while export still shows open |
| Priority is unique | Sample accounts proving no double-booked signal | Priority invents uniqueness the glossary never enforced |
| Override is controlled | Logged use with time-bound restore of review | Shared override or unlogged use |
Walk four cases in Pingpong: a surface still pinned to one overloaded glossary, a weight that left a partner renewals chart live without CS review, a class that mixed batch and interactive signals, and a request to skip a finance signal because volume is rare. Start from the documented rebuild language. Ask who can extend a rebuild window and which export proves review. Any step that depends on an unnamed person becomes a rebuild condition.
Replay one prior weight incident under the proposed rules. If the incident would still age without a reverse owner, revise the rules before calling the rebuild ready.
Pair with the customer success ops lead seat, a customer health score review, and a customer success playbook review. More operating decisions live in the war-game decisions hub.
Close only when approved and blocked labels can be reproduced from stored exports without a verbal override.