David Divitt, Senior Director of Sales Engineering at Visa with over two decades in the fraud and risk industry, discusses how e-commerce businesses can support and promote inclusivity while effectively combating digital fraud.
In 2023, global e-commerce sales reached an astounding $4.9 trillion—a figure that reflects the digital transformation of commerce worldwide. However, with this growth comes an increase in e-commerce fraud. According to a new study from Juniper Research, the value of e-commerce fraud is predicted to rise from $44.3 billion in 2024 to $107 billion in 2029—a growth of 141%.
As fraudsters evolve their techniques, businesses must adopt sophisticated, multi-layered fraud prevention methods. However, it’s equally important to ensure these methods remain inclusive.
Despite the massive growth in online commerce and development of smart technologies, there are still a huge number of people who are being excluded from online services.
Would you believe that the third largest nation on Earth does not have any borders? According to research by Sumsub, nearly 650 million residents—referred to as the Greenflag nation by Sumsub—are underserved or marginalized by outdated fraud prevention techniques and risk policies. These individuals may come from high-risk regions or demographics, lack technological experience, or possess minimal financial footprints in conventional settings, needing or preferring to carry out their commercial activities in more traditional ways. At the same time, they shouldn’t be excluded from the modern digital world.
This article explores how modern fraud prevention enables businesses to welcome legitimate customers while effectively blocking fraudulent activities.
When we think about e-commerce and digital services, inclusivity isn’t often the first thing that springs to mind. Those accustomed to digital interactions may take for granted that these methods are straightforward for everyone. However, individuals with minimal digital engagement may feel vulnerable and exposed online. Rapid technological evolution further complicates their navigation of digital commerce. According to research from Link, almost a quarter of UK adults feel digitally excluded. Around 24% classify themselves as such, with the highest rates among 18–24-year-olds (42%) and those aged 75 and over (30%).
Individuals in high-fraud-risk countries often face additional barriers to digital inclusion as businesses may avoid onboarding them due to perceived risks.
Imagine someone with a minimal digital footprint. They use cash for most transactions, lack a credit card, but have a bank account (set up in-branch) with a debit card. Now, they want to explore e-commerce and online finance.
To start, they buy a new mobile phone to get online and manage their finances. They sign up for an online marketplace to find a good deal. The process is simple, and they quickly add a phone to their virtual basket.
Then comes the payment.
The marketplace sees the “higher risk item” from a first-time purchaser—first red flag. Traditional fraud prevention measures use relatively limited data-sets to make decisions. This is an unfortunate effect of the natural silos which are created when individual businesses only see the data that relates to them. These more traditional approaches often use home-grown, or basic fraud systems which are limited in their sophistication and don’t see patterns beyond their four walls.
The marketplace applies simple rules and flags that several mobile phones have recently been purchased from this region, with fraud rates exceeding the threshold—red flag two.
A local data source the marketplace connects to runs the phone number through a search and finds very little history—red flag three.
The order is declined, not even sent for authorization, and this citizen of the Greenflag nation’s attempt to digitize is thwarted. They are excluded from modernization. Often, this exclusion pushes individuals toward less secure services, increasing their vulnerability to fraud.
Suggested read: Addressing the Digital Divide in 2024
As digital commerce expanded, fraudsters adapted alongside. Pre-pandemic fraud prevention, especially in brick-and-mortar businesses, lagged behind digital-native counterparts. It relied on simple rules, minimal external data, limited AI and Machine Learning (ML/AI), and a “one-and-done” tooling approach.
Global lockdowns forced businesses and individuals to rapidly adapt to a new digital reality, and traditional fraud prevention methods revealed their weaknesses. Fraudsters excel at exploiting vulnerabilities, and the sudden influx of businesses new to online operations—along with millions of inexperienced users, the citizens of the Greenflag nation—created ample opportunities for fraud.
As online interactions continue to dominate, it is essential that fraud and risk management tools evolve. Inclusivity must be integral to all strategies, as it benefits everyone. Sumsub’s research reveals that digital services could lose $1.75 trillion in revenue by excluding digitally marginalized individuals—users eager to engage in the digital economy.
The key to inclusivity, especially as it relates to fraud and risk, is about connecting the world because as you bridge the silos that exist naturally across our technology ecosystems, you unlock the potential to dramatically improve inclusivity, and the experience for everyone.
By treating fraud prevention as a team sport and utilizing signals beyond just the data an individual merchant or bank sees, it ensures there are no blind spots in which fraudsters can hide.
Build your fraud prevention around a toolset that is made up of multiple techniques, each specializing in the thing they do best to ensure coverage across the many attack vectors a fraudster could exploit. And by always putting the customer, and their experience, front and center, we can design out many of the things that alienate the citizens of the Greenflag nation. Fortunately, we have the technology today to make this a reality.
Revisiting our citizens of Greenflag trying to purchase a new phone online for the first time…
Then comes the payment.
This time the marketplace is using a combination of tools to help manage fraud and risk. Instead of simple rules, they are now using ML/AI models trained on billions of data points, not just a few. They are scanning various pieces of their registration info across global databases and registries to find even the smallest digital footprint. They are also using a step-up risk assessment to run a simple identity check prior to the purchase. Using a quick selfie and a query out to a national registry, the marketplace now can clearly understand the true risk of Greenflag citizens, and sees that the risk is actually low.
Our digital explorer is approved, the phone is shipped, and the beginning of a fruitful online existence is underway, no longer marooned in the borderless nation of Greenflag.
Applying this success story to millions of excluded individuals gives them a voice and participation in the modern world while opening significant growth opportunities for our economies, all while keeping fraudsters at bay.
So what does a truly modern fraud prevention solution look like? We should think about it as 3 pillars:
Fraud changes constantly. Weaknesses are probed, new techniques arise, and new criminal actors pop in and out along the way—not to mention how the evolution of our own businesses drives a constant change in the options for good and bad. Any fraud system must keep up. So, making small changes regularly (limits, areas of focus, risky items, etc), or large changes now and then (journey flows, policies, new payment types, etc) shouldn’t be a big headache, and shouldn’t require specialist intervention. And with this flexibility comes the ability to fine tune your strategies to minimize impact to genuine customers and maximize it with fraudsters. As in our example above, with minimal friction, we were able to serve a new customer, and keep our business safe.
It’s essential to utilize ML/AI algorithms in your fraud prevention strategies. Predictions suggest that AI-driven fraud detection could save businesses approximately $10.4 billion globally by 2027.
Suggested read: Machine Learning and Artificial Intelligence in Fraud Detection and Anti-Money Laundering Compliance
However, AI is not a silver bullet to stop all your fraud. While AI plays a critical role in modern fraud detection, integrating it into a multi-layered framework enhances its effectiveness. The best defence is a layered approach, and one that has a broad set of approaches and technologies. From ML/AI to rules, from pre-order analysis to post-order monitoring, from device biometrics to biometric authentication, they all have a role to play, and will all place barriers in front of criminals trying to abuse. Your specific business risk will determine how broad and resilient to go, but don’t underestimate the benefits of having as many layers of defence as you can manage—just make sure they all play nice together and aren’t causing extra effort.
A flexible, broad fraud solution is limited if it only sees the data you see. Unfortunately, criminals are very aware of the siloed nature of technology, across regions, across industries, and across businesses. They use the blind spots created here to carry out the same attacks, using the same technology, and the same devices over and over again until they run out of blind spots. Then they morph slightly and carry on. We can dramatically reduce these blind spots by using consortiums wherever we can. Vendors that specialize in a few types of intelligence but collected over millions or billions of entities become incredibly powerful at securing those blind spots and stopping the repeat offenders in their tracks, even if it’s the first time they have tried their crime with your business. Organizations utilizing comprehensive threat intelligence platforms report a 30% increase in the detection of repeat offenders due to enhanced visibility across multiple data points.
Rethinking approaches to the Greenflag nation’s citizens can yield significant benefits. Through thoughtful combinations of modern technologies, such as ID verification, pre-authorization fraud checks, ML/AI, and shared datasets we can reduce the global digital divide, include millions of previously excluded individuals, foster safe business growth, and deliver excellent customer experiences.
