Don’t Alienate Your Customers with Fraud False Positives
One of the major challenges for any eCommerce merchant is figuring out how to detect and stop credit card fraud. Fraud is a common and widespread problem, and it inexorably leads towards angry customers, time-consuming disputes, and costly chargebacks. Merchants have many tools available to help them root out potential fraud, but these tools often mistakenly identify legitimate transactions as being possibly fraudulent.
False positives may not be as directly harmful as actual fraud, but they can still cause problems for merchants. What are the dangers of fraud false positives, and how can merchants improve their accuracy when it comes to identifying fraudulent transactions?
Online fraud is a problem that keeps growing every year, and fraudsters are constantly revising and evolving their tactics to get around the defenses that cybersecurity experts are able to come up with. As the COVID-19 pandemic shifted an even greater share of economic activity toward the eCommerce sector over the past year, online fraud has only grown more prevalent.
This kind of fraud is extremely damaging to merchants, who are financially liable for the chargebacks that result from credit card abuse. Worse yet, merchants may be penalized or blacklisted by their banks and payment processors if their fraud situation gets out of control.
Fraudsters try their best to slip under the radar by mimicking real customers, which makes it more difficult to detect and block them with only basic identity verification methods. Increasingly, fraud detection relies on artificial intelligence, machine learning, and sophisticated algorithms that look for subtle signs of fraud in transaction data. These fraud detection tools may be highly effective, but they are not infallible, and false positives are a frequent occurrence.
What are Fraud False Positives?
Many current fraud detection tools operate by assessing the risk that any given transaction may be fraudulent. They assign a score based on how many possible fraud indicators can be found in the transaction data and block transactions that meet or exceed a certain score value.
These risk scoring tools look at things like the device fingerprint, geolocation, IP address, transaction frequency, order history, and various other elements of the transaction data. They may also apply proprietary analytics, often informed by machine learning and AI technologies, in order to arrive at a final score. Often, the merchant will have some control over the settings and thresholds, allowing them to determine a risk tolerance level that makes sense for them.
Transactions blocked by an anti-fraud tool may be held for manual review and approval, or simply rejected outright.
Invariably, no matter how your tools are configured or which methodologies are used, a fraud detection system will erroneously identify a legitimate transaction as possible or likely fraud. This is a fraud false positive.
How Do Fraud False Positives Impact Customers?
When a legitimate transaction is falsely flagged as fraud, there are two possible outcomes. If the transaction is withheld for manual review, the merchant has a second chance to look it over and use their best judgment to determine whether or not the transaction seems safe enough to process.
Ideally, the merchant will recognize that the transaction is most likely valid and will allow it to go through, but humans are just as capable of arriving at the wrong conclusions as algorithms are.
If a merchant rejects a false positive after manual review, or if their ordering system is set up to automatically reject transactions that exceed a particular risk score threshold, then a legitimate sale has been lost, and it is likely that the customer will be quite unhappy.
With nearly one in three consumers embracing a one-strike-and-you’re-out approach to bad merchant experiences, the true cost of a sale lost to a fraud false positive can be considerable.
Even the orders that make it past manual review can be detrimental to the customer experience. Manual review takes time, which can mean delays in processing and shipping the order. It also often involves the merchant directly contacting the customer to verify their information, which some customers may find intrusive.
Manual review may be necessary to keep your overall false positive rejection rate down, but it is not without its own costs.
How Can Merchants Improve their False Positive Rate?
It’s not easy to tell how many valid orders are being rejected due to fraud false positives. Fraudsters never follow up with you later to confirm that they were, in fact, trying to defraud you, and customers insulted by an erroneous rejection will often just silently take their business elsewhere.
You can obtain valuable insights from customers who do contact you to complain about being rejected by your fraud filters, as well as from transactions that are held for manual review and later found to be legitimate. You might learn, for example, that multiple false positives are coming from a particular geographic region or IP block, and can adjust the configuration of your anti-fraud tool accordingly.
Beyond that, getting a clear picture of your fraud false positive rate may require high-level auditing and analysis of your transaction data, anti-fraud activities, and manual review processes. There are also third-party vendors who can test your fraud filters in simulated environments.
For many merchants, anti-fraud tools that use risk scoring methodologies to detect and block credit card fraud are a necessity. With online fraud rates as high as they are, these merchants would be overwhelmed with chargebacks if they didn’t use some sort of automated filtering system to stop fraudsters.
The problem is that it’s not always easy to draw clear lines between customers and fraudsters, so merchants are stuck in a delicate balancing act trying to keep their filters restrictive enough to stop fraud but not so restrictive that it affects too many of their real customers.
Efficient manual review processes are an important part of maintaining this balance. By moving quickly to verify and approve false positives you can minimize the delays and bad experiences for customers whose orders were blocked in error. Later on, you can analyze the information from the false positives to inform your fraud filter settings and improve the efficiency and accuracy of your manual review procedures.