Name screening sounds simple: check a name against a list. In practice it is one of the hardest problems in compliance, because names are messy, lists are imperfect, and the same human can be written a dozen ways. The result is false positives — matches that look like a hit but are actually a different person. Manage them well and screening produces clear signal; manage them badly and the real risk drowns in noise.
Why a name match is hard
| Challenge | Example |
|---|---|
| Transliteration | محمد can be Mohammed, Muhammad, Mohamad, Mehmet… |
| Common names | Thousands of real "Maria Santos" or "Wei Chen" exist |
| Name order | Family-name-first vs given-name-first conventions |
| Nicknames & initials | Bill/William, J. Smith vs John Smith |
| Typos & OCR | Data-entry and scanning errors in source records |
A screening system that demands an exact match misses all of these and gives dangerous false confidence. So screening uses fuzzy matching.
What fuzzy matching does
Fuzzy matching scores how similar two names are rather than requiring them to be identical, so near-misses still surface. It is essential — a sanctioned person spelled slightly differently must still be caught — but it has an unavoidable trade-off.
Disambiguation: turning hits into decisions
A raw match is not an answer; it is a question. Disambiguation answers it.
A screen flags each case. Decide whether it's a genuine match to act on, or a false positive to clear.
Your subject's full name, date of birth AND nationality all match a sanctioned individual.
What good practice looks like
- Use identifiersCompare date of birth, nationality, location and known IDs — not just the name string.
- Don't clear on spelling aloneA different spelling is expected; only mismatched identifiers clear a hit.
- Escalate strong matchesWhere identifiers align, treat as a real match and follow the process.
- Record the reasoningDocument why each hit was cleared or escalated — that is the audit trail.
- Tune, then reviewCalibrate matching to your risk, and review the rate of false positives over time.
The cost of getting it wrong
Where Probitas fits
A Probitas screen aims for signal over volume: it surfaces credible matches with the identifying context you need to disambiguate, and anchors each to its source so you can judge it quickly. It reduces the noise; the clear/escalate decision, and its documentation, remain yours.
False
What is a false positive in name screening?
A screening "match" that turns out not to be your subject — usually a different person who shares a name, or a near-match thrown up by fuzzy matching. Clearing false positives accurately is most of the work in screening.
What is fuzzy matching?
A technique that scores how similar two names are rather than requiring an exact match, so near-misses (different spellings, transliterations, name order, typos) still surface. It is essential for catching real matches but inevitably produces more false ones.
Why do common names cause so many false positives?
Because thousands of unrelated real people share them. Without additional identifiers — date of birth, nationality, location — a common-name match is impossible to resolve and floods the analyst with noise.
How do you reduce false positives without missing real matches?
Disambiguate using identifiers rather than tightening matching to the point of missing variants. Compare date of birth, nationality and location; never clear a hit on spelling differences alone; and document the reasoning for every decision.
Is it better to have more matches or fewer?
Neither extreme. Too many matches buries real risk in noise and wastes resource; too few risks missing a genuine sanctioned match. The aim is precision — surfacing the right matches with enough context to decide quickly.
Sources
This guide is written from primary sources. Each is linked below; claims in the text link to the specific reference they rely on.