How Machine Learning is Revolutionising Fraud Prevention in E-commerce: An Expert’s Perspective

In the realm of e-commerce, fraud prevention is paramount to safeguarding businesses from financial losses and maintaining consumer trust. Jonathan Hougardy, Head of Business Development E-commerce at Worldline, sheds light on the significance of combating e-commerce fraud and the role of advanced technologies in addressing this challenge.

Let's delve into his insights on the operational benefits of AI and machine learning-based fraud prevention solutions and the transformative impact they bring to the e-commerce landscape.

Trends and talking points

Q: What is e-commerce fraud and why is it a significant issue?

A: E-commerce fraudsters use financial information deceptively to make illicit transactions on online shopping platforms. Over time, e-commerce fraud has evolved dramatically, with fraudsters employing increasingly sophisticated techniques. Phishing attacks, botnets, and identity scams have grown in complexity, making detection more challenging.

In 2023, online fraud caused industry losses exceeding $48 billion for the first time (source: Juniper Research), eroding consumer trust and inflicting financial strains on businesses. This comes in light of advancements in payment security through efforts like PSD2 and strong customer authentication underscoring the need for businesses to adopt more advanced and adaptable fraud prevention solutions.

AI and machine learning-based solutions offer a dynamic approach to fraud prevention

Q: How can AI and machine learning help in fraud prevention?

A: In the e-commerce landscape, static rule-based fraud prevention engines—like those monitoring IP addresses, BIN numbers, or blacklists—are no longer effective against the evolving and sophisticated nature of fraud. Fraudsters use complex tactics which can easily bypass static rule-based systems.

However, a combination of static rules and AI/machine learning-based solutions offers a dynamic and adaptable approach to fraud prevention that does not throw out the efforts businesses have already made in preventing fraud but enhances them. These technologies analyse large volumes of data in real-time, detecting subtle and emerging fraud patterns, adapting to evolving threats, and enabling early and accurate detection of fraudulent activities.

Q: What is Microsoft Dynamics 365 Fraud Protection (DFP) and how can it assist merchants?

A: Microsoft Dynamics 365 Fraud Protection (DFP) is a cutting-edge hybrid fraud management solution offered by Worldline Online Payments. By combining rules (including Device Fingerprinting) and machine learning atop an extensive dataset, DFP enhances accuracy and efficiency beyond the capabilities of rules or machine learning alone.

Its unrivalled fraud detection capabilities make it indispensable for global merchants lacking an in-house fraud tool or sizable fraud team. DFP helps reduce fraud costs, increase acceptance rates, and enhance the overall customer shopping experience.

Q: Can you explain the scalability of fraud prevention solutions like DFP?

A: With over 20 patents for fraud machine learning and over 6 billion transactions processed yearly, DFP’s machine learning is one of the most advanced and extensively trained on the market. The extended data network includes data from merchants across all verticals and regions. Whether in retail, travel, or other sectors, these solutions can identify suspicious behaviours and anomalies, providing effective protection against a wide range of fraud.

AI and machine learning-based solutions increases conversion rates by allowing more legitimate transactions

Q: What are ‘false declines’ and why are they a problem?

A: False declines happen when legitimate transactions are mistakenly marked as fraudulent, causing revenue loss and customer dissatisfaction. In 2023, an estimated $700 billion in global e-commerce revenue was lost due to false declines (source: Statista), exacerbating the fraud issue. Our ability to minimise false declines is crucial for enhancing merchant brand confidence and enabling them to achieve new levels of growth.

Q: How can AI and machine learning help reduce false declines?

A: Machine learning-based solutions play a crucial role in reducing false declines by proactively analysing vast amounts of transactional data and not only identifying suspicious patterns but also safe patterns. Unlike traditional methods relying on static rules, AI-driven machine learning solutions adapt to changing user behaviour and fraud patterns, distinguishing between legitimate and fraudulent transactions with increased accuracy.

Through our partnership with Microsoft and its DFP solution, we've successfully decreased false declines by approximately 1.4% for our global merchants. This has had a substantial positive impact on performance and subsequent revenue growth.

Q: How does reducing false declines improve e-commerce performance?

A: Reducing false declines boosts e-commerce performance by increasing conversion rates and nurturing customer loyalty. Studies reveal that 28% of customers abandon purchases after a payment decline, with 14% opting for another merchant to complete their purchase (source: Riskified). AI-based fraud prevention not only raises conversion rates by streamlining legitimate transactions, thus driving up sales and revenue but also fosters customer trust by ensuring a seamless shopping journey.

Moreover, our Worldline exemption engine relies on the DFP prevention score, assisting merchants to improve their frictionless transaction rate. In this way, Worldline optimises the performance-risk balance, empowering merchants to enhance profitability and outsmart fraudsters.

New fraud prevention technologies contribute to reducing chargebacks

Q: How do fraud prevention technologies reduce chargebacks and improve trust between merchants and their banks?

A: Fraud prevention technologies contribute to reducing chargebacks by detecting and blocking fraudulent transactions before they occur. By identifying these suspicious activities in real-time, fraud prevention solutions enable merchants to avoid not only the losses from fraudulent transactions but also the fees associated with chargebacks.

Worldline’s global team of fraud experts and DFP’s advanced machine learning empower merchants to make smart, accurate, and quick decisions to reduce their chargeback rate and increase their acceptance rate.

Adapting these fraud prevention technologies highlights merchants' commitment to transaction security and minimises risks for banking partners, fostering trust and collaboration between them. The resulting enhanced trust can lead to additional benefits, including improved payment terms and increased financial support for future business endeavours.

Q: Can you provide recent statistics on the cost of dispute management and the amount of money lost as a result?

A: Chargebacks pose a substantial financial challenge for merchants. Recent findings reveal that for every Euro lost to online fraud, merchants face chargeback costs exceeding €3. Furthermore, for every euro of online fraud, merchants lose around an additional €4 in processing fees, fraud-related costs, and revenue loss. By investing in a fraud prevention tool, a merchant can save greatly on dispute management costs.

Increase operating efficiency

Q: How do fraud prevention technologies based on machine learning contribute to operational efficiency?

A: Machine learning-powered fraud prevention technologies leverage advanced algorithms to automatically detect and block fraudulent transactions, slashing the need for manual review. Implementing machine learning-based fraud prevention solutions has resulted in more than an 80% decrease in manual review rates, achieving a mere 0.03% manual review across all platforms for Microsoft.

This reduces operational costs associated with manual reviews and enhances efficiency by allowing teams to focus on more strategic tasks, improving overall productivity and performance. Depending on the size of the merchant's fraud team, they can manage their fraud account with Worldline's global team of fraud experts or independently.

In today's fast-paced e-commerce landscape, the demand for advanced fraud prevention solutions is more critical than ever. Traditional rule-based methods fall short against sophisticated fraud tactics. However, machine learning-based solutions like Microsoft Dynamics 365 Fraud Protection (DFP) offer a proactive approach to combating fraud.

Our solution not only reduces chargebacks and false declines by utilising advanced algorithms and a vast dataset but also boosts operational efficiency by minimising manual reviews. Scalable and adaptable across industries, it fosters trust between merchants and their banking partners while providing robust protection against emerging fraud threats. In conclusion, Activating DFP on Worldline Online Payments ensures businesses can confidently navigate the complexities of e-commerce fraud with a single onboarding.