Financial tech services brought rapid innovation to banking. Businesses everywhere can manage capital faster and with more flexibility, making room for speedier growth and greater expansion.
But these advancements also give fraudsters more opportunities to do what they do best: exploit anything in the customer journey that reduces friction for trusted users. In a recent webinar, Sift Sr. Product Manager, Jim Payne, and Adam Gibson, Head of Product at Flinks, discussed what fintech merchants can do to protect customers from evolving types of fraud, even as they scale to meet new demand.
Protection, Growth, and the Pitfalls of Friction
Protection and growth shouldn’t have to be tradeoffs, but many e-commerce companies rely on fraud prevention tactics and tools that sacrifice one for the other. Recent research into how online businesses weigh the importance of customer security vs. satisfaction shows that, despite considering both to be critical, only some companies are backing up that belief with platforms that don’t pit these goals against each other.
“Every company understands the importance of creating an amazing experience, but that doesn’t necessarily mean that they’re doing it,” said Payne. “We surveyed 500 online businesses in the U.S. and Canada last year, and we found that three-quarters of them are prioritizing a frictionless experience…but at the same time, nearly two-thirds of them say that their online fraud solutions actually introduce the maximum possible friction.”
More friction might seem like an obvious way to reduce risk, but multiple security gates often slow down the entire customer journey, and strict risk thresholds frequently block trusted users and legitimate transactions. The disconnect goes further: despite clear flaws, many of those same survey respondents said they plan to continue investing in the same tools, the same strategies, and will consequently have to hire more trust and safety experts to fight increased fraud. These businesses are essentially doubling down on an approach that treats every customer with suspicion, something Gartner essentially describes as treating your customers like criminals—risky until proven loyal. For companies that choose to invest in the status quo time and time again, outpacing fraudsters and fueling growth might be inch-by-inch, if not impossible.
Digging deeper into Sift’s data, says Payne, it’s clear that advanced machine learning solutions provide unparalleled accuracy, speed, and sophistication, enabling fintech merchants to catch more fraud at scale. In industries now blighted by the impact of COVID-19—like fintech, where this year’s traffic and transaction volumes have gone through major peaks and valleys—merchants can’t afford to bank on solutions that only get half the job done.
The Sift Score: Surfacing Meaningful Signals and Shifts
Data is at the core of how Sift and Flinks evaluate trust. Financial institutions use Flinks’ APIs and data tools to connect, enrich, and translate data into action.
“When it comes to machine learning-based systems,” said Gibson, “Data, not algorithms actually make the difference in accuracy.” Aggregating multiple types of data, he says, and applying them to a set of customers, greatly increases the accuracy of a holistic risk assessment.
Sift uses supervised machine learning to ingest, connect, and analyze billions of events across e-commerce, giving trust and safety teams a precise view into the amount of risk associated with a login or other event. Using Sift’s Dynamic Friction, they can then choose to automatically apply appropriate security gates to every event, ensuring that legitimate users get fast-tracked—while fraudsters get tracked down and taken out.
“With that [Sift score], you can actually customize the action you take to maximize both the customer experience and the level of protection that they’re experiencing,” said Payne. “So, from minimizing friction, to forcing a password reset or two-factor authentication, the power is in [the merchant’s] hands to actually apply just the right amount of protection for just the right level of risk.”
Because different types of fraud occur at different points in the user journey, it’s key to understand that Sift Scores aren’t fixed; they’re not an assessment of individual customers, nor are they based on a single user action. Sift assesses transactional, identity, and behavioral signals, and aggregates them with signals surfaced from our global merchant data network, such as whether a vendor has seen a certain behavior, email address, or type of fraud before.
“We’re not just looking at the login. Behind one single score at one particular point, you’re talking about hundreds and even thousands of different signals that we’re evaluating from device fingerprint to IP address, velocity scroll, failed login attempts,” said Payne. “A hundred thousand different signals that are being evaluated to create one final score that’s being updated all the way across [the user] journey.”
A Sift Score captures the whole picture, distilling these thousands of signals into a single metric, providing a consolidated view of how likely or unlikely it is that fraudulent activity will occur—and depending on the merchant and what they want to know about a specific event, that score can change.
Watch the full Friction, Fraud, and Fintech: Delivering Speed and Security with Digital Trust & Safety webinar for more insights into online fraud’s impact on financial technology services, and what fintech merchants need to know to gain traction and users without losing revenue to fraudsters. You’ll also learn:
- How Sift and Flinks use data to accurately assess risk and prevent fraud in fintech
- The tools and tactics that enable e-commerce merchants to deliver frictionless, secure experiences to customers
- The shortcomings and true cost of prioritizing speed over security
- How different fraud vectors, like account takeover, content abuse, and payment fraud affect different stages of the customer journey
The post [Webinar] Friction, Fraud, and Fintech: Delivering speed and security with Digital Trust & Safety appeared first on Sift Blog.
Source: Sift Science