Tech for good: Using data and technology to pick out the bad apples

05 March 2018


In our first blog we talked how organisations can protect themselves against financial crime, for the second blog in our series we’ll look into more detail about how to use data and technology to understand your customers.

How would you spot a criminal when you’re out and about? Criminals – particularly those involved in the money laundering industry – tend to look and on the surface, behave just like everyone else. Which often makes them very difficult to spot.

Banks and other financial institutions know that some customers, whether business or personal, are riskier than others. So how do you separate out the potential criminals without upsetting your honest customers?

“Know Your Customer” (KYC) procedures have required banks to gather data and build the risk profile of their customers, at the time they first join the bank and then into the future. But this can be a blunt tool fraught with unintended consequences, including lengthy checking processes that impact customer service, spiralling compliance costs, and innocent customers being caught in the net.

Now more than ever, it’s essential that firms can accurately and efficiently identify their high-risk customers and segment them out for closer monitoring and due diligence. There is plenty of data available to feed into customer risk rating (CRR) models – but it’s intelligent analysis of the right combination of good quality data that’s really key.

The most effective CRR models take a 360⁰ view of each customer, through a series of lenses. In other words, they don’t just look at location or the sector the customer operates in, but consider every angle – who they are, where they are, how they transact, and if the customer is a company, how it’s structured and other factors that make for high-risk ownership – to build a comprehensive risk profile. Everything is considered in context, not in isolation. A nail bar owned by a mother and daughter in Romford might not ring many alarm bells – but one owned through an opaque corporate structure with a branch in rural Mexico certainly might.

Successful CRR modelling is like attempting to complete a complex jigsaw without a picture to guide you, or even knowing if you have all the pieces. Some data may be missing, or of poor quality, and a decision has to be made about how to fill in the gaps. In other cases the level of detail in the data will be crucial – for example when assessing the risks associated with different business that deal with the same or very similar products or services. We’ll come back to data quality in more detail later in this blog series.

CRR models have a huge impact on a firm’s operations – driving the level of due diligence – and creating them is inevitably subjective. As a result, emotions and implicit bias play a part. These must be carefully governed by due process. High risk customers mean extra costs, but missing even one brings significant risk. The definition of high and low risk levels needs to be clear, as do the thresholds that the firm feel comfortable with; it’s a difficult but important process that needs strong governance around it. But in a world where, according to our previous Global Financial Crime Survey, money laundering is worth up to $2 trillion a year, it’s one of the best examples of how data analytics can pick out  financial criminals and help build trust in society.

Today, we work with some of largest financial services firms in the world, to help them solve problems and catch criminals. If you would like to discuss how we can help you please contact [email protected] or [email protected]

Scott Samme|Director, Financial Crime Analytics|+44 (0)7753 605698

Jeremy Davey|Director, Financial Crime Analytics|+44 (0)7867 502 988