Combating fraud by choosing the right strategy and implementing machine learning

Friday 2 June 2023 11:57 CET | Editor: Claudia Pincovski | Interview

Discover the value of machine learning in combatting fraud in 2023. Learn how to choose the right fraud prevention strategy, with insights from Sift's community-driven approach.

As fraud continues to be on the rise globally, what are some of the main fraud prevention trends merchants should focus on in 2023? Is machine learning gaining momentum and, if so, can you explain its importance?

A concerning trend is the commercialisation of fraud, which is the ability for anyone to commit fraud, even those who previously lacked the skills or resources to do so. Easily accessible fraud forums and marketplaces abound on the dark and deep web, selling the tutorials and tools that enable anyone with malicious intent to commit fraud or to engage third parties in ‘fraud as a service’. And many of these forums, particularly on apps like Telegram, are openly marketing their services to regular consumers.

In response, we’re seeing new rules like Visa’s Compelling Evidence 3.0 (CE 3.0) enacted. Visa CE 3.0 updates a process for online merchants to dispute chargeback fraud with an extended list of evidence, including an IP address or device ID, a shipping address, or user account. As a baseline, online merchants need an online payment fraud prevention solution that can continuously monitor and collect these vital analytics.

With fraud becoming increasingly prevalent and easier to execute, businesses that want to stay secure and protect their revenue must invest in technology that relies on machine learning (ML).

Organised fraud actors are already leveraging automation tools to accelerate attacks, so there is no way for legacy services to keep up, especially if they are dependent on manual review. ML, when paired with the right data, is the foundation for building a fraud prevention program that places user experience as a top priority.

The last thing to consider is that economic uncertainty is a driving force behind raising fraud rates. Unfortunately, this economic uncertainty also pinches corporate budgets, resulting in staffing shortages. Across the board, organisations are being tasked to do more with reduced resources. Consequently, machine learning is non-negotiable for any business that wants to fight fraud with accuracy and efficiency while simultaneously enabling growth through revenue protection and customer retention.

Global fraud prevention solution providers are innovating with new approaches to better mitigate digital risk and help businesses grow safely. What are, in your opinion, the main differences between various types of fraud prevention platforms on the market?

Currently, there are three primary types of fraud prevention strategies businesses can choose from a build-you-own model, an insurance model, and a comprehensive Digital Trust & Safety platform.

When choosing between these options, businesses should keep in mind these key considerations or attributes:

  1. Does the solution align with business incentives?

  2. What is the decision accuracy?

  3. How much visibility and control will your risk team have?

  4. What capital and people efficiency does it provide?

  5. What is the solution’s time to value?

The first option – an in-house fraud decisioning strategy – requires massive investment and commitment to execute successfully. Businesses that choose this fraud prevention method must commit to outsized capital and people to support, and can expect to wait a longer period before they see their investment pay off.

Fraud prevention providers aligned to the insurance model can be initially appealing because they offer a chargeback guarantee and a virtual outsourcing of an entire function. However, this guarantee also means that they are more likely to reject a disproportionate number of overall transactions, including many legitimate ones, which can have a real negative impact on user experience and customer retention.

Digital Trust & Safety platforms, on the other hand, offer the optimal balance across all of the above-mentioned attributes by providing nuanced and accurate fraud risk scoring, efficiency with capital and people, and real-time responsiveness – leveraging the wisdom of machine learning, powered by a global data network and management structures focused on transparency and control.

When choosing a fraud prevention solution provider, how important is it to focus on transparency and control in fraud operations?

Transparency and control are two of the top factors to consider when evaluating fraud prevention solutions (the others are automation, investigation, and scalability). These two factors are important because online merchants need to be able to make and apply informed decisions to their business – and do so quickly.

Transparency is the ability to obtain visibility into fraud risk with the context to analyse patterns – both red flags and positive signals alike. For example, an online merchant may be able to determine if particular customer segments are at greater risk for fraud.

Control is the ability to apply the right decisions to transactions and modify risk thresholds, based on unique business needs. For example, an online merchant may determine that they don’t want to challenge low-value sales because the risk of an abandoned shopping cart is not worth the potential risk of fraud. Solutions like Sift enable merchants to apply dynamic friction, which evaluates the risk score of each transaction on a case-by-case basis.

Finally, it takes a network to fight a network, both human and machine, which means solution providers, merchants, and end-users must all work together to stay one step ahead of fraudsters. How is Sift’s new customer community portal help reducing the gap in the field?

Sift’s new customer community, Sifters, is one of many ways that we’re levelling the playing field for fraud fighters. The cybercriminal underground on the dark and deep web is well-networked and fast-acting, quick to share the newest fraud techniques and security flaws within their own communities. Sift is forging a community for the ‘good guys’.

The community is called Sifters and it allows our customers to connect with each other and with our Trust and Safety Architects to collaborate and share best practices and information on emerging fraud threats. Sifters is the human layer to our digital network – a partnership among digital risk practitioners who embrace the mission to fight fraud and grow safely.

About Armen Najarian

Armen Najarian is Sift’s Chief Marketing Officer. Armen Najarian is Sift’s Chief Marketing Officer. He’s built extensive strategic experience serving as CMO at fast-growing fraud, identity, and cybersecurity firms, including Outseer, Agari, and ThreatMetrix.

About Sift

Sift is the leader in Digital Trust & Safety, empowering digital disruptors to Fortune 500 companies to unlock new revenue without risk. Sift dynamically prevents fraud and abuse through industry-leading technology and expertise, an unrivaled global data network of one trillion (1T) events per year, and a commitment to long-term customer partnerships. Global brands such as DoorDash, Twitter, and Wayfair rely on Sift to gain a competitive advantage in their markets. Visit us at, and follow us on LinkedIn.

Free Headlines in your E-mail

Every day we send out a free e-mail with the most important headlines of the last 24 hours.

Subscribe now

Keywords: fraud management, machine learning, fraud prevention, merchants, chargebacks, online payments, data, insurance, transactions , cybercrime
Categories: Fraud & Financial Crime
Companies: Sift
Countries: World
This article is part of category

Fraud & Financial Crime


Discover all the Company news on Sift and other articles related to Sift in The Paypers News, Reports, and insights on the payments and fintech industry:

Industry Events