7 Ways Automated Fraud Protection Can Fail You
Automated fraud protection systems – such as the fraud filters that are often integrated within e-commerce platforms – have historically been the most popular ways for e-commerce merchants to prevent potentially fraudulent orders from processing. After all, fraud filters are a relatively inexpensive solution that can quickly analyze a vast amount of transactional data.
But times are changing, and so are the most effective approaches to fraud prevention. The algorithms used in fraud filters and other automated fraud prevention systems may no longer be enough to provide the level of security a growing e-commerce retailer needs.
Why is this approach, which has been used for so long, decreasing in effectiveness?
The Major Drawbacks of Automated Fraud Protection
There are many reasons automated solutions can cost e-retailers profits and customers, including:
1. They Have High Rates of False Declines
One of the biggest problems merchants experience with these automatic systems is ensuring legitimate transactions are approved while fraudulent ones are declined: LexisNexis reports that fraud filters have a false positive rate of approximately 25%, costing merchants $118 billion in 2015.
2. They Use Incomplete Data Sets
These systems generally make decisions based on limited amounts of transactional information, so they can’t account for situational changes. For example, if a business establishes a rule that automatically declines transactions when the billing address is different from the shipping address, they may inadvertently reject an order a customer is buying as a gift for a family member.
3. They’re Rarely Customized
Fraud filters can’t pick up on the fraud approaches that are unique to individual businesses or industries. Instead, they typically apply the same set of rules to all businesses.
4. They’re Ineffective if Improperly Layered
If fraud filters aren’t optimized, they can limit sales and increase false declines. And if merchants add one filter on top of another filter without thinking about the order in which the filters will be applies, certain rules may override other rules — reducing or even eliminating their overall effectiveness and either allowing fraudulent orders to pass through or incorrectly flagging legitimate orders as fraudulent.
5. They May Be Resource-Intensive
Although these solutions may be automated in the way they evaluate transactions, they still require constant maintenance and dedicated human staff to ensure the various rules are up to date and being applied properly.
6. They May Be Too Strict
While merchants might think that the strictest possible security measures provide the best protection, they’re only half right. Strict rules may miss fewer fraudulent transactions, but they’re also more likely to falsely decline legitimate transactions. Strict systems can also slow down order response times, increasing customer frustration and checkout abandonment.
7. They Aren’t Self-Learning
This is perhaps most important but least understood by merchants: some automated fraud systems don’t use past experience to improve future decisions. Once a rule is set, it’s set. If a fraudster understands the rules a merchant has in place, it’s easy for the fraudster to exploit them.
Fraudsters are always increasing the sophistication of their attack methods. A common tactic is to make subtle changes to existing threats and malware. These tiny changes to coding produce new threats that are virtually undetectable by software — including automated fraud protection methods. So while automated systems may be effective at picking up major fraud trends, they’re less effective at detecting emerging fraud strategies and small-scale attacks.
Even if the automated systems could pick up on the slight changes and new patterns, they’re not sophisticated enough to incorporate these new fraud strategies into their algorithms on their own and use them to identify and prevent future attacks.
So … if automated systems aren’t the end-all-be-all many merchants believe them to be, what should merchants do instead?
A Better Approach to Fraud Protection
With so many parts of e-commerce business going high-tech, it’s no surprise that merchants are looking for a technology-based solution to screen their transactions. Unfortunately, with a simple automated fraud protection system, merchants often find themselves vulnerable to lost profits due to false declines, fraudulent transactions and chargebacks.
Instead, merchants should consider a hybrid solution: expert human analysis combined with advanced artificial intelligence (AI). The AI system continually analyzes and processes high-level features from raw data. Then the analysts add value by adding data points, transaction results and incident details into the AI system, making it smarter and more effective.
Together, the hybrid system becomes nearly foolproof.
At ClearSale, we combine data analytics, statistical intelligence and human expertise to help e-commerce merchants improve fraud protection while maximizing sales. Learn more about the different fraud protection tools available to you, and then contact us to find out how our unique approach can benefit your business.