How to Reduce Sanctions Screening False Positives

To reduce sanctions screening false positives, you tune three things: the quality of your own data, the way your tool matches names, and the identifiers you use to rule people out. Most false positives are not a software failure. They come from thin data and loose settings that flag anyone with a similar name.

A false positive is a screening alert that turns out to be innocent. Your customer shares a name with someone on a sanctions list, but they are not the same person. Too many of these slow your team down and hide the real hits. This guide shows you how to cut the noise without missing a true match. It sits alongside our guide on how to resolve a false positive, which covers what to do once an alert lands.

Why false positives happen

Sanctions lists hold common names. The OFAC SDN list alone runs into the thousands of entries, and many share names with ordinary people and firms. When your tool compares a name against a list, it looks for similarity, not certainty. A loose comparison flags too much. A tight one risks missing a real party.

Three root causes drive most noise. Poor input data gives the tool little to work with. Match settings that are too broad catch near-strangers. And a lack of secondary identifiers means the tool cannot tell two same-named people apart. Fix these and the volume drops fast.

Start with your own data quality

Screening is only as good as what you feed it. If you send a tool a first name and a country, expect a flood of alerts. Clean, complete records are the cheapest way to cut false positives. This is work you control, and it costs nothing in software.

Collect more than a name

Capture full legal names, dates of birth for individuals, and full registered names for companies. Add address, country, and any registration or tax number you have. Every extra field gives the tool a way to clear a partial name match. A shared name plus a different date of birth is an easy dismissal.

Standardize before you screen

Fix the formatting problems that create fake mismatches and fake matches. Consistent input helps the tool compare like with like.

• Split first, middle, and last names into separate fields.
• Remove titles, honorifics, and role words such as Mr, Dr, or Director.
• Use one country and date format across all records.
• Transliterate non-Latin names in a consistent way.
• Strip legal suffixes like Ltd or GmbH into their own field so they do not distort the core name.

Bad data is the top cause of noise. Our denied party screening best practices cover how to keep records clean over time. For a broader view of the process, see how to process denied party screenings.

Tune your match settings

Most tools let you set how close a name must be before they raise an alert. This is often called the match threshold or fuzzy matching score. Set it too low and you catch everyone. Set it too high and you may miss a real party who used a slight name variant. The goal is balance, backed by testing.

Understand the threshold

Fuzzy matching scores how similar two names are, often as a percentage. A score of 100 means an exact match. A score of 75 means the names are close but not identical. A low threshold treats weak similarity as a hit. Raising the threshold cuts weak matches, but you must test before you move it.

Setting	Effect if too loose	Effect if too tight
Match threshold	Floods you with weak name hits	May skip real parties who vary spelling
Phonetic matching	Matches names that only sound alike	Misses transliteration variants
Lists enabled	Screens against lists you do not need	Leaves a required list switched off

Match on the right lists

Screening every record against every list in the world creates noise you do not need. Enable the lists your obligations demand and your risk requires. A US exporter must screen against OFAC and the relevant US export lists. You can view the primary source at the Office of Foreign Assets Control. Many firms also use the Consolidated Screening List, which bundles several US lists into one feed.

Set fuzzy matching per list where your tool allows it. A high-risk list can carry a looser threshold so you catch more. A low-risk list can run tighter to cut noise. Adjust by customer geography, product type, and risk. New customers in high-risk regions warrant more caution than long-standing domestic accounts.

Use secondary identifiers to clear alerts

A name match alone rarely proves anything. Secondary identifiers are the extra data points that separate your customer from the listed party. Sanctions list entries often carry a date of birth, a nationality, a passport or ID number, and known addresses. Compare these against your own record and most false positives fall away.

• Date of birth. A different year or date clears an individual fast.
• Nationality and country. A mismatch in country of residence or citizenship weakens the hit.
• Passport or ID number. A direct number check gives a clean answer.
• Address. A different city or country supports dismissal.
• Company registration number. For entities, a registry number beats a name every time.

Feed these identifiers into your tool where it supports them. Many platforms can filter out a hit automatically when the date of birth or country does not match. That keeps the alert off your analyst’s screen entirely. Watch the 50 percent rule for entities, since ownership can pull a firm into scope even when its name is clean.

Whitelist known-good matches

Some customers will alert every time you screen them. They share a name with a listed party, and nothing about that will change. Once you have cleared and documented a match, add it to a whitelist, sometimes called a good-guy or exception list. The tool then suppresses that specific alert on future runs.

Whitelisting saves your team from resolving the same false positive again and again. Do it with care. Record who approved the entry, the evidence used, and the date. Review the list on a schedule, because a whitelisted party can change or the underlying listing can shift. Keep the audit trail. See record keeping for denied party screening for what to store.

Test, measure, and adjust

You cannot tune what you do not measure. Track your false positive rate and review it by segment. If one customer group or one list drives most of the noise, focus there. Small, tested changes beat one big guess. Never loosen a threshold without checking what real hits it would drop.

A simple tuning loop

1. Measure your current false positive rate and time spent per alert.
2. Pick one change, such as raising a threshold on a low-risk list.
3. Test it against past alerts to confirm you still catch true hits.
4. Roll out the change and document the reason.
5. Re-measure after a set period, then repeat.

Backtest against historical data before you push a change live. This shows whether a tighter setting would have missed a real party. Keep every rule change timestamped so an auditor can see your logic. This links back to your wider program in the denied party screening hub and the screening checklist.

What not to do

Cutting false positives is worth doing. Cutting them the wrong way creates legal risk. The point is a cleaner queue, not a blind one.

• Do not raise thresholds so high that you miss real matches. That is a compliance failure, not an efficiency gain.
• Do not disable a required list to cut volume.
• Do not whitelist a party without evidence and a documented approval.
• Do not skip re-screening. Lists change and today’s clear record can alert tomorrow.
• Do not treat a dismissed alert as undocumented. Write down why you cleared it.

The line between a false positive and a missed hit is thin. Stay on the safe side and keep proof of your decisions. Study the red flags of sanctions evasion so you do not dismiss a hit that deserves a closer look.

Frequently Asked Questions

What is a good false positive rate for sanctions screening?

There is no fixed number that fits every firm. Rates depend on your customer base, the lists you use, and your risk appetite. The better measure is trend. Track your rate, then work to lower it without dropping true hits. A rate that falls while your true-hit catch holds steady is the sign of good tuning.

Will raising the match threshold cause me to miss real hits?

It can, which is why you test first. Backtest any threshold change against past alerts before you go live. If the tighter setting would have dropped a real party, do not make the change. Tune the threshold per list so high-risk lists stay sensitive while low-risk lists run tighter.

Is whitelisting allowed under sanctions rules?

Yes, when done with discipline. A whitelist suppresses repeat alerts for a party you have already cleared. Regulators expect evidence, a named approver, and a review schedule. Never whitelist to hide a party you have not investigated. Keep the audit trail so you can show why each entry is safe.

Does better customer data really cut false positives?

Yes, and it is often the biggest single lever. A name plus a date of birth, a country, and an ID number lets the tool clear near-matches on its own. Thin data forces the tool to alert on names alone. Collect and clean identifiers at onboarding and the noise drops before you touch a single setting.

Key Takeaways

To reduce sanctions screening false positives, work on data, settings, and identifiers together. Collect full names, dates of birth, and ID numbers, then standardize them before you screen. Tune your match thresholds with testing, enable only the lists you need, and set fuzzy matching by risk. Use secondary identifiers to clear name-only hits, and whitelist proven matches with a documented reason. Measure your rate, backtest every change, and keep an audit trail. Do all of this without ever loosening controls to the point where you miss a real party. If you want help building or tuning a screening program, contact us.