Screening for personal names always throws up challenges, and every language has its own features we must understand if we are to be effective in matching. Russian names are currently all over Sanctions lists, and they have their own nuances.
Russians use the Cyrillic alphabet, which contains some characters with look very different to Latin characters, some character that look similar, and some that are identical. But even when they look the same they are completely different and nearly always make a different sound.
It is impossible for the human eye to tell the difference between a Cyrillic character and a Latin one when they are the same shape, but to a computer they are quite different. sometimes when a Cyrillic name is input into a system we can end up with a mixture of Cyrillic and Latin letters, which makes no sense at all. This can even happen on Sanctions Lists, and it can certainly happen if you have any Latin/Cyrillic switchable keyboards used for data entry into your systems.
Here is an example from the EU Asset Freeze List:
You might be thinking what is wrong with that name, it is clearly in the Latin Alphabet, but what should be a Latin I in ROMASKIN is actually a Ukrainian Dotted Cyrillic I. The name should look like this:
You can probably not tell the difference unless you are viewing it with a font that has a slight difference in the way the letters are displayed.
There are many other examples on the EU list where there is this kind of mix-up, but they are all Cyrillic names with an odd Latin character thrown in:
In this example the first letter is a Latin A, but the 6th, 14th, 23rd, and 25th letters are all Cyrillic A.
The way to find issues like this is through Data Profiling, which will identify the different types of character that are in use.
If you are only using Latin characters, then you can “fix” these issues by substituting the Cyrillic character for the corressponding Latin character prior to matching (I have come across banks who do this). If you are only using Cyrillic characters then you can do the same in the opposite direction. However if you are using both Latin and Cyrillic then you will probably need to perform some data remediation and list management exercise to get these types of error fixed.
These Russian False Friend characters can cause disruption to matching, a screening system will see them as completely different letters, even if they sound the same. They can happen by accident quite easily, but could be used to purposefully hide a name from screening.
At SQA Consulting we can not only give you hints and tips on screening effectively and data management, we are leaders in measuring the effectiveness and efficiency of screening.
Please contact us to find out more about this article, or download our brochure here.
