Screening threshold playground
Configuration
Matches scoring above this are flagged for review
Simulated results
7 flagged8 cleared100.0%
weighted score
100.0%
weighted score
96.4%
weighted score
95.4%
weighted score
94.6%
weighted score
94.0%
weighted score
87.4%
weighted score
79.4%
weighted score
75.8%
weighted score
72.2%
weighted score
71.6%
weighted score
68.6%
weighted score
52.8%
weighted score
49.8%
weighted score
45.6%
weighted score
Understanding weighted screening scores
Sanctions screening typically uses a weighted scoring model that combines multiple factors to produce a single match score. The three primary factors are:
Name Similarity — measured using algorithms like Jaro-Winkler, accounting for transliterations, spelling variations, and name permutations. This is usually the highest-weighted factor.
Date of Birth — when available, DOB matching significantly reduces false positives. An exact DOB match increases confidence that a name match is a true positive.
Country/Nationality — matching the customer's country against the sanctioned person's nationality or country of operation provides additional confirmation signal.
This playground is a teaching tool; it doesn't reflect FirmScreen's own settings. Higher name weights catch more potential matches but increase false positive volume. Adding DOB and country weighting helps filter out false positives without missing true matches.
FAQ
Questions firms ask
Common questions about screening thresholds and scoring.
Still have a question?
hello@firmscreen.io
A screening score threshold is the minimum similarity score at which a potential match is flagged for review. Setting the threshold too low produces excessive false positives that overwhelm compliance teams. Setting it too high risks missing true matches. The optimal threshold depends on your risk appetite, customer base demographics, and regulatory expectations.
False positive rates are influenced by the screening threshold, the size and composition of your customer base, the fuzzy matching algorithm used, the granularity of identifying data available, and how name variations are handled. Populations with common names (e.g., Mohammed, John Smith, Li Wei) naturally generate more false positives. Field-level weighting that incorporates DOB and nationality can significantly reduce false positives.
There is no one-size-fits-all threshold. The FATF recommends a risk-based approach where the threshold reflects your business risk profile. Common practice is to set a primary threshold around 85-90% for initial review and a critical threshold above 95% for immediate escalation. FirmScreen allows you to configure separate thresholds for different risk categories.
From possible match to recorded decision
In FirmScreen, every possible match becomes a case with a risk level and an owner, closed with a recorded reason.
Book a demo