A false-positive is when a test identifies an issue incorrectly. The phrase false positive used to be predominantly used in the medical world — if someone tested positive for a condition they didn’t have, that was a false positive. Now, however, the phrase is often heard in the banking industry. It refers to situations where fraud detection software flags a legitimate transaction as fraudulent.
This issue can compromise the customer experience and erode consumer trust in your bank. To protect your financial institution, you need to actively look for ways to reduce false positives.
How False Positives Degrade the Customer Experience
Consumers expect their financial institutions to protect them from fraud. If someone in another country tries to make a purchase with their card, they expect their bank to deny the transaction. If someone tries to cash a forged check from their account at the bank, they expect the bank to refuse to cash the check.
However, if the customer is using their own card to make a purchase in a foreign country, they expect the transaction to be accepted. If they write a check on a day when their hands are a little shaky, they expect the check to go through. Unfortunately, however, banks can’t always tell the difference between real and legitimate transactions — especially when there’s not a lot of difference between the two.
As a result, banks invariably end up flagging some legitimate transactions as fraudulent. When this happens, customers typically don’t think, “My bank is doing a great job at fraud detection. I’m so glad they’re protecting me.” Instead, they get annoyed about the friction.
Customers want a smooth experience. They want to be able to use their cards online and in person whenever they want, wherever they are. If they are unable to, they feel dissatisfied. In a worst-case scenario, they may close their account and look for a new bank. The more hassle the customer experiences, the more likely this becomes.
The Risk of False Positives With Rule-Based Fraud Detection
Rule-based fraud detection has historically been the most common way to detect bank fraud, but unfortunately, it has a very high rate of false positives. This type of fraud detection relies on preset parameters, and anything that falls outside of the parameters is labeled as fraud.
For example, a rule-based system may flag any transactions outside of the country as fraudulent unless the customer contacts the bank about their travel plans. Similarly, a rule-based system may automatically flag any transactions over a certain value as potentially fraudulent.
In some cases, the rules may be slightly more complex. For instance, a bank may be able to set different spending limits on new accounts than on seasoned accounts. It may provide more scrutiny to low check numbers than to high check numbers.
Unfortunately, however, regardless of the sophistication of the rules, rule-based fraud detection is very susceptible to false positives. To protect themselves from fraud without compromising the customer experience, banks need to take a more adaptive and dynamic approach to fraud detection.
How to Reduce False Positives
To reduce false positives, you need to automate your fraud detection processes. Automated tools can scan more transactions at a faster rate than humans. They can also look for minute discrepancies that cannot be detected by the human eye. For instance, an automated solution will detect forged signatures much more effectively than an employee.
Ideally, your automated solution should utilize the following:
Artificial intelligence — AI-driven fraud detection solutions don’t have to rely on rule sets. Instead, they learn intricate details about different types of fraud. Then, they scan multiple data points to look for signs of fraud. They assign a risk score to different transactions, and if the score exceeds a certain level, they flag the transaction for manual review or customer verification.
Behavioral analysis — In addition to learning the general signs of fraud, your fraud detection solution should also learn about the unique behaviors of your customers. While certain activity on one account may signal fraud, the same activity on another account may be business as usual. If you have a solution that can learn about your customers’ behaviors, you will reduce the risk of false positives.
Easy customer verification — If your system flags a transaction as fraudulent, you want to be able to easily contact your customer. A simple text or call should alert them that you have denied the transaction. If the process is legitimate, they should be able to verify that they initiated it with a tap on their phone. They shouldn’t have to call you or jump through hoops. If a customer can easily resolve a false positive, the impact will be a lot less severe on your bank.
Machine learning — Even the very best fraud detection programs occasionally generate a false positive. Machine learning helps to stop this from becoming a pattern. If your software flags a legitimate transaction as fraudulent and you note that it was legitimate, the software learns from the experience. It looks at all of the context surrounding the issue, and the next time it sees the same issue, it uses the knowledge it gained during the last interaction. There are many different types of machine learning that can help improve fraud detection processes.
Real-time data — Utilizing real-time data in fraud detection isn’t just about scanning transactions as they occur. It’s also about taking into account recent and previous activity to draw conclusions about transactions in real time.
To give you a basic example, imagine that a customer is making a transaction for an unusual amount in a place where they have never used their bank card before. Your system flags the transaction as fraudulent, but when your customer verifies that they were really using the card, your bank approves the purchase. Only a few minutes later, your customer tries to make another purchase — it’s also an unusual amount in a strange location. However, because your fraud prevention solution works in real-time, it already knows that your customer is traveling, and it allows the transaction to go through smoothly.
Get Help Reducing False Positives
Financial institutions must be diligent against fraud. They need to protect their assets, and they also need to stay compliant with industry regulations. However, overly strict, rule-based processes often label legitimate transactions as fraudulent, and this compromises the customer experience.
What is your bank’s false positive rate? Is it higher than the industry average? If so, you’re likely frustrating customers or even driving them away. We can help. At SQN Banking Systems, we have a variety of fraud detection and prevention services solutions that can help you reduce fraud without generating unnecessary false positives.