Manual fraud system outpaced by company growth
Fraud is a serious problem for any business, but crypto exchanges have become a distinct target for attacks in recent years. Banxa has encountered multiple fraud scenarios, including fake profile creation, card fraud, scams, and chargebacks. Initially, Banxa set up their own fraud function from scratch, handling everything manually when volumes were manageable. The team built fraud rules into this system to manually review orders using basic variables for customers, demographics, and payment methods. However, as Banxa began to grow, this basic model became too limited for their needs. It introduced unwanted friction for trusted customers and became riskier when incorporating multiple variables and increased velocity. So when Banxa’s volume spiked 30x, their fraud rate rose alongside it.
The team knew they needed to implement something quickly to support their scaling business, which is where Sift came in. The fraud operations team worked with a consultant specializing in fraud and payments to find the right tool for their needs, focusing on a criteria of cost, usability, and ability to scale. After a thorough analysis of online fraud detection platforms, the team found Sift to have the right capabilities and ease of use they were looking for.
Sift’s automation and machine learning means that we’re detecting fraudsters before they’re even able to transact with us.
Rohan, Manager of Fraud and Scams
Leveraging Sift Payment Protection to shut down fraud
Banxa built their fraud operations around Sift and the platform has become a primary component of their fraud control model. The team leverages Sift Payment Protection to measure daily fraud and chargeback rates, identifying and responding to any fraud patterns associated with certain partners or transaction methods.
Sift also fits nicely within Banxa’s value chain, both for order creation and payment risk. To protect order creation, the team leverages Sift’s network of device intelligence and user session tracking to prevent suspicious users from creating orders. And to protect against payment fraud, they utilize Sift’s machine learning and data science capabilities to construct highly effective and automated detection strategies. With these capabilities, Banxa can automatically accept, hold, verify, and decline orders based on routing and rules. The team has even created a multitude of different watch decisions to segment the various types of user behavior on a spectrum from trusted to risky. They use the explore function to look at common attributes of fraud cases and pinpoint exactly when a user goes from trusted to risky. This is especially useful for complex scam activity, as it can often look like regular customer behavior.
Route workflow metrics allow the team to analyze friction points that are and aren’t working and quickly make changes. Instead of setting and forgetting rules, Banxa is able to continuously review these metrics to determine how they can further reduce friction for trusted customers.
The network visualization in Sift is also extremely useful for Banxa to see how network attacks with hundreds of compromised cards are all linked together from an IP address or other signal. Sift needles out users that are stealing credit card information and sharing it across multiple customer profiles—a process they would have had to identify one by one before Sift. Being able to easily digest this relevant data allows the team to establish an efficient operational rhythm and take action quickly. Finally, Sift allows them to automatically detect and shut down card fraud, which is lower value but easier to stop.
Every tool has ML now, so it really comes down to how comfortable you are with the tool and how the team works with you. We really appreciate the people who work with us on this problem, that shouldn’t be overlooked.
Igor, Head of Fraud
Preventing fraud and enabling growth
Since implementing Sift, Banxa has been able to identify and auto-block fraudulent activity before any damage could ensue, resulting in reduced and stabilized fraud and chargeback rates. Once hamstrung by manual effort and limited on how they could respond to attacks, the team can now scale and quickly squash major attacks with ease. After initially dealing with an influx of scams that impacted trusted customers, Sift has helped Banxa significantly reduce these scam events.
Tapping into Sift has allowed the fraud team to feel in control and in the driver’s seat, and systematically respond to fraud when it does pop up. By maintaining consistent fraud and chargeback rates, Banxa can now accurately predict margins and costs. Sift has also helped the team stay lean and efficient, as it would have taken 5x more resources to keep up with the volume without Sift. The automated efforts and operational efficiency has helped the business scale by keeping up with a 30x increase in volume.
Having Sift in place means the team can shift gears and focus more on end-to-end conversion, which is crucial for an on/off ramp like Banxa. The team now spends more time fine-tuning and optimizing the customer journey, paving the way for a seamless user experience and opportunities for growth. Sift enables Banxa to expand to parts of crypto that others may steer away from, including Web3. Banxa is now able to provide more services to help their Web3 partners operate in a safe and secure ecosystem, and they know they can rely on Sift.
Sift has helped us stabilize our fraud and chargeback rates so we feel in control and in the driver’s seat. We don’t have to look over our shoulder all the time. Being confident in our predicted fraud rates means we can be confident in our margins and managing costs, that value can’t be overstated.
Igor, Head of Fraud
Secure your business from login to chargeback
Stop fraud, break down data silos, and lower friction with Sift.
- Achieve up to 285% ROI
- Increase user acceptance rates up to 99%
- Drop time spent on manual review up to 80%