If your PEP and sanctions checks keep returning long lists of people who are obviously not your client, the problem is often not the screening tool. It is how the name went in. Reversed name order, a missing date of birth or one spelling of a name that has six common versions can each turn a clean check into an afternoon of reviewing false matches.
This guide covers the most common causes of false matches and a set of entry rules any small team can follow. It also explains the one thing you should never do to reduce the noise: weaken the search until nothing comes back.
Why the same person can have many spellings
Screening tools use fuzzy matching on purpose. A sanctioned or politically exposed person might appear on a list under a different spelling from the one in your client’s documents, so the tool returns close variations rather than exact matches only. That protects you, but it also means the quality of what you enter decides how much noise you get back.
Names change shape for a few predictable reasons:
- Transliteration. Names written in Arabic, Cyrillic, Chinese or other non-Latin scripts have to be converted into Latin letters, and there is often more than one accepted way to do it. The same Arabic name can produce several different Latin spellings.
- Name order. Many Chinese, Korean, Vietnamese and Hungarian names put the family name first. If your form assumes ‘first name, last name’, the parts can end up in the wrong fields.
- Multiple surnames. Spanish and Portuguese naming customs often use two surnames. Dropping one changes what the tool compares.
- Patronymic elements. Arabic names may include ‘bin’ or ‘bint’ (son or daughter of) and a chain of ancestors’ names, which can be recorded in full or shortened.
- Diacritics and special characters. Characters such as ü, é or ø are often stripped or converted on official documents and in databases.
- Changes over time. Marriage, divorce and legal name changes mean a person can have earlier names that still appear on lists.
Use the passport’s machine-readable zone as your reference
For individuals with a passport, the two lines of characters at the bottom of the photo page (the machine-readable zone, or MRZ) are a useful reference. Under the International Civil Aviation Organization (ICAO) standard for travel documents, names in the MRZ appear in Latin characters only, with national characters transliterated and punctuation removed. The MRZ also separates the primary identifier (usually the surname) from the secondary identifier (given names).
That makes the MRZ a consistent source for which part of the name is the family name and how it is spelled in Latin characters. Use it alongside the printed name, not instead of it, because the printed name may show a fuller or more familiar version.
Entry rules that cut false matches
| Situation | What to enter | Why it helps |
|---|---|---|
| Every individual | Full name exactly as on the primary ID, plus date of birth and nationality or country of residence | Date of birth and country rule out most unrelated results |
| Family name written first | Put the family name in the surname field, whatever order it appears in on the document | Stops the tool comparing given names against surnames |
| Two surnames | Enter both surnames as they appear on the ID | Keeps the full name the lists are likely to hold |
| Non-Latin script name | The Latin spelling from the passport MRZ or printed transliteration, plus any other spelling the client uses | Matches how the name is most likely recorded |
| Known aliases or former names | Run a separate check for each recorded alias or previous name | A list may hold the person under an older name |
| Name also written in its original script | Enter the original script name as well, if your tool accepts it | Lets the tool compare the name before transliteration |
| Initials only on the form | Ask for the full given names before screening | Initials greatly increase the number of possible matches |
| Companies | Full registered name, registration number and country of incorporation | Distinguishes entities with similar names |
| Company legal suffixes | Test with and without suffixes such as “Pty Ltd” or “Limited” if your tool does not handle them automatically | Suffixes vary between registers and lists |
Gender and address can help too where your tool accepts them, but date of birth and country do most of the work. A check run with a full date of birth will usually return a fraction of the results of the same name with no date.
What not to do
There is a tempting shortcut when a common name produces dozens of results: trim the name, misspell it slightly, or switch off fuzzy matching until the results disappear. Do not do this. A screening check that cannot find a real match is not a control, and a clean result produced that way will not stand up if your process is reviewed.
Match rate settings have a place, but set them in line with your risk-based policy and record the setting you use, rather than tightening them check by check until a result disappears. The right way to reduce noise is to add information, not remove it. If results are still unmanageable after you have entered every identifier you hold, the answer is a structured review of each result, not a weaker search.
A quick worked example
Illustrative example. A bookkeeping practice onboards a new client whose passport shows the name NGUYEN Thi Lan, with the family name first. A staff member enters ‘Nguyen’ as the given name and “Lan” as the surname with no date of birth. The check returns a long list of results, because Nguyen is one of the most common family names in Vietnam and the fields are reversed.
Re-entered with ‘Nguyen’ as the surname, ‘Thi Lan’ as the given names, the full date of birth and Australia as country of residence, the same check returns a short list that can be reviewed in minutes.
Keep a note of what you entered
Whatever you enter becomes part of your record. If you screen a name in more than one way, such as two spellings or a former name, keep each result. Someone reviewing your file later should be able to see exactly which details were checked and why. Our guide to recording screening decisions sets out what the record should contain.
Frequently asked questions
Why does a common name return so many results?
Because the tool is matching on the name alone. Adding a date of birth and country is the quickest way to narrow the results to people who could genuinely be your client.
Should I switch off fuzzy matching to get fewer results?
No. Fuzzy matching exists so that spelling differences do not hide a real match. Reduce noise by adding identifiers instead.
Which spelling should I use for a name in a non-Latin script?
Start with the Latin spelling on the client’s passport, which follows the international travel document standard, and also screen any other spelling the client commonly uses.
Do I need to screen a client’s maiden name or former names?
Where you know a client has used a different name, screening it as well is sensible, because a list may record the person under that name.
How do I enter a company name?
Use the full registered legal name with the registration number and country of incorporation, and screen any trading or former names separately.
How NameScan can help
NameScan lets you add details such as date of birth, gender, country and the name in its original script to each check, screen companies as well as individuals, and run single or batch checks on a pay-as-you-go basis. When potential matches do come back, you can record your decision on each one in the dashboard so your file shows the details you entered and how you reviewed them.
Start screening with NameScan, or read our guide to reviewing a potential match.
0 Comments