Two operations can run the same ERP, the same forecasting model, and the same alerting rules, and still land three days apart on how fast a flagged exception turns into an executed decision. The gap isn't the technology. It's everything that happens after the alert fires.
Definition
Decision latency is the time between a signal arriving and an approved action reaching the system of record. It's the composite measure across all five layers, Signal, Route, Approve, Execute, Audit, not a single-stage metric.
A strong decision intelligence layer can still produce high decision latency, because intelligence measures the quality of the recommendation, not the speed of what happens after someone reads it. The Decision Latency Diagnostic scores each of the five layers separately and reports where the time is actually leaking, usually not where operations assume.
Why does a "governed" decision intelligence layer still produce high decision latency?
Decision intelligence and decision latency measure different things. Intelligence measures whether the recommendation was good. Latency measures how long it took, after the recommendation arrived, for someone to act on it. That gap is governed by Route, Approve, Execute, and Audit, not by the model that produced the signal. An operation can have an excellent recommendation engine and still take three days to act on what it recommends.