Screening Threshold
The level of fuzzy matching or scoring used by a screening system to determine when a potential match (alert) is generated. A lower threshold generates more matches, but also more False Positives.
Threshold is a trade-off knob, not a correctness setting: lowering it catches more genuine variants of a listed name at the cost of more false positives for analysts to clear; raising it reduces alert volume at the cost of missing genuine variants. There is no universally correct value — the right threshold depends on the risk profile of the customer base being screened and the corroborating data available (dates of birth, nationality, structured identifiers) to help distinguish true matches from coincidental name overlaps.
A high match score is a ranked hypothesis, not a verdict. The system should present enough context — which list, which programme, what corroborating identifiers agree or disagree — for a person to resolve the alert, rather than auto-deciding on score alone in either direction. Auto-blocking on score alone blocks real customers; auto-clearing on score alone eventually clears a designated one.