Soaring sales and less bad debt for a retail finance giant

Our client, a market leader in retain finance, needed a better way stop dead-end lending and push profitable applications through faster.

The answer lay in clamping down on organised crime. Specifically, finding the hidden links to OCG applications guaranteed to turn bad, while green-lighting legitimate customer credit at speed.

To achieve this, our client turned to Synectics’ predictive AI, rock-solid data and rapid risk screening tools: the vital trio for flagging early warning signs and subtle links to organised crime activity that erodes retail finance growth. 

The results were remarkable. Referral rates were cut by over 50%, fraud detection speed increased by 64% and sales significantly increased as a result of, more accurate application decisions.

Learn more about their strategic approach and its impact below.

The client

Our client ensures high street brands can provide customers with credit products that make general merchandise goods affordable and accessible.

They currently process over 350,000 applications every year, linked to purchases ranging from gadgets, appliances, and high-end tech, to toys, household furnishings and health and beauty products.

They wanted a better way to predict and defend against fraud, in particular organised fraud, in order to say yes to more good customers and screen out the bad.

The context

Organised crime (3rd party fraud) is a major threat to retail finance operations. Sophisticated scams involving identify theft, account takeovers, and the creation of synthetic identities are increasingly commonplace – costing the industry millions every year.

There are also broader business implications. If you don’t know a ‘lend is bad’, you chase it. Racking up time and recovery costs on top of the eventual write off amounts.

Similarly, without a mechanism in place to accurately predict credit applications linked to fraud - enabling them to be rejected and removed from genuine customer risk assessments - a ‘financial loss snowball’ starts rolling.

Probability of Default (PD) predictions lose accuracy and Expected Credit Loss (ECL) calculations become inflated, ultimately resulting in excess funds being set aside to cover anticipated losses, affecting profitability, business agility, and opportunities for growth.

The challenge

Against this backdrop, our customer had a clear goal. Help us find the often-well-hidden indicators of organised crime to predict and avoid bad lending at speed. As while fraud detection is key, so is fast decisioning. Consumers will quickly lose interest in application processes that are unduly lengthy or complicated, resulting in lost sales.

We knew we could help our customer address all these challenges, and ease pressure on a busy fraud prevention team in the process.

The solution

We proposed Precision - our award-winning AI-powered predictive analytics tool. When used to screen customers, its powerful algorithms leverage a range of data sources to detect tell-tale behaviours suggestive of organised crime connections. It spots, and flags, even the subtlest of clues. And continuously learns as it goes to keep pace with new and emerging threats.

Building on solid ground

In this case, we had a good foundation on which to build. A year earlier, our customer had started using tailored rules applied to National SIRA in order to risk assess applications. Implementing this mechanism had already led to more accurate fraud scoring, and a 50% reduction in referral rates.

But they still lacked the predictive power so useful for identifying applications highly likely to be linked to organised crime groups (OCGs). Doing so is vital to avoiding unnecessary write-offs and endless time chasing a dead-end recovery.

Working with their team, we tailored a 3rd party fraud screening Precision model to give them the boost they wanted.

Drawing on diverse data sources

Risk intelligence from our proprietary database National SIRA, data from the National Fraud Initiative (NFI), and information from other relevant third-party sources - which in this case included device data (type/IP addresses) and wider credit bureau information - were all ingested into the model developed.

Data points within the model were weighted, and continue to be regularly recalibrated, in line with our customer’s ongoing needs, and to optimise learning. Every Precision model we build for a customer gets smarter with every application assessed.

Fast result with clear reasoning

In less than 5 seconds, Precision delivers not only a risk score, but the decisioning behind it. For example, scores may be higher where a device or email is not typically linked to applicant name, or where there are subtle differences in personal information supplied to different data sources assessed.

The scores enable applications to be grouped into risk cohorts (with thresholds dictated by our customer), to fast-track good customers, filter out bad ones, and refer those that truly need closer attention. Such proactive due diligence can cut organised retail fraud off at the source: screening for links to profit-eroding crime and stopping the “snowball effect” of bad lends in seconds

The outcomes

From being dubious about the added value AI could offer, our client now uses Precision as their primary fraud prevention and risk analysis tool. When you hear the words “we’ll never go back to life before Precision”, you know it’s working.

1. Prevention rate up, workload down

Our customer’s love for Precision is backed up by clear facts. Since adopting the 3rd party screening model, 10% more fraud is detected (and subsequent bad lending prevented). At the same time, referral rates have more than halved to just 2.5%.

“With Precision, we’re detecting previous levels of fraud 64% faster, have increased our prevention rates, and dramatically reduced the volume of cases that require working.”

2. Staying ahead of developing threats 

Precision’s ability to detect the majority of emerging threats and trends was recently borne out, when a new organised attack involving the use of fraudulent gift cards emerged.

"Precision alerted us to the issue, and was already preventing 90% of activity related to this attack. Aware of the risk, and with intelligence on how it was manifesting, we could then quickly work with Synectics to fine tune rules to plug the remaining gap".

3. Securing more sales, supporting more customers

Precision’s speed means shoppers applying for finance from our customer get a decision quickly, meaning they are less likely to abandon their basket. A win for both sides – reflected in the fact that sales have significantly increased since Precision’s 3rd party fraud screening was introduced.

"The explanatory data provided enabled us to  support shoppers who are themselves victims of fraud, for example where identity theft has occurred as part of organised crime efforts".

Answering the important questions

What is the probability of a borrower defaulting? Are you financing baskets destined to become a loss to sell on?

Answering these questions accurately is central to reducing so many cost drivers in retail finance. It’s also a question you can answer with up to 100% certainty when you know the borrower is linked to, or has been used by, OCGs.

The accuracy with which Precision can identify suspicious activity means our customer is better equipped than ever to minimise fraud related defaults and the costly recovery process, and to serve more good customers, faster.