Screening transactions for fraud requires businesses to find the right balance. Should they screen each and every transaction that passes through their e-commerce portals? Or should they flag only those that appear fraudulent?
The former may be more time-consuming and expensive, but it does a better job of making sure fraudulent transactions can’t slip through. The latter requires less manpower and may be adequately managed by standard fraud filters, but it risks seemingly innocuous (but actually fraudulent) orders getting through security channels.
In today’s online marketplace, most transactions are legitimate. But because not all of them are, it’s critical to have protections in place. And that means finding the right balance between ease and efficiency on one hand, and protection of profits and brand reputation on the other.
Let’s look at the pros and cons of each approach.
The Risks of Selectively Screening Transactions
When businesses opt to use artificial intelligence (AI) strategies to “cherry pick” only the transactions that appear fraudulent, they assume several risks.
- Difficulty discerning fraud from not fraud. Inexperienced merchants may misjudge fraudulent transactions that appear legitimate.
- Rejecting too many transactions. A lower fraud rate may actually reflect a high turndown rate that can negatively affect a merchant’s reputation.
- Changing face of fraud. Traditional AI-based strategies may be unable to detect emerging fraud patterns. When fraudsters evolve their tricks and AI rules can’t keep up with the changes, the merchant’s fraud management is rendered ineffective.
- Playing the system. Fraudsters are very good at flying under a fraud filter’s radar. If they discover a merchant has a $1,000 threshold for fraud screening, they’ll keep their purchases under $999 so they can be automatically approved.
- Complicated fraud filter management. Merchants often use multiple rules in their fraud filter to flag as many fraudulent transactions as possible. But counterintuitively, this can increase risk exposure. If the merchant isn’t careful about this order in which rules are applied (which rule is applied first, second, etc.), it’s possible that some rules may cancel out others, which can leave the merchant less protected than before.
- Missing the big picture. Handling transactions individually may cause merchants to overlook the bigger picture. For example, the purchase of one PC delivered to Cherry Lane seems legitimate. But if the merchant is not screening every transaction, he might not realize that 10 PCs were delivered to different addresses on Cherry Lane — all around the same time — which could indicate a fraud attack.
When the focus is on individual cases and not the larger pattern, preventing large-scale fraud attacks becomes difficult.
The Advantages of Screening Every Transaction
If hand-selecting transactions to be screened isn’t the answer, the answer may be to screen every transaction.
Why is this advantageous?
- Low value can still mean high risk. Fraudsters often make low-value purchases just to see if a stolen credit card number works. If it does, that opens the door to larger fraudulent purchases. Screening every transaction lets you see the small transactions that are testing the fraud waters — transactions you might otherwise ignore.
- Screening more transactions helps you develop the bigger picture. As you screen more transactions, you increase the amount of stored transaction data you have to draw on, so more accurate decisions can be made across the board. Consider our example above with the PCs shipped to Cherry Lane; each individual sale may look innocent, and the true fraud pattern only emerges when the transactions are viewed together in a broader context.
- Questionable transactions can be more closely evaluated. Some transactions will always fall into a gray area, where it’s simply hard to tell if they’re fraudulent or not. Fraud management systems that rely solely on AI will automatically decline all of these transactions. But if you have a system to research and analyze these transactions, you’ll most likely find that some of the gray area sales are good and should be approved.
- Prevents large-scale fraud attacks. Professional fraudsters often work together to coordinate attacks on a company in a very short time period. If a company individually assesses orders, it’s like pouring cold water on a fire; attacks can be more quickly identified, stopped and prevented.
While AI has made great strides in fraud management, businesses can’t rely on it alone to determine which orders are truly fraudulent and which ones aren’t. Instead, businesses that complement AI with a team of human analysts find that the team helps the AI solution become smarter and more effective, allowing the right transactions to be approved and increasing client satisfaction and retention.
Developing a Comprehensive Fraud Management Solution
As fraudsters continually improve their strategies, it’s important for e-commerce merchants to develop a comprehensive approach to preventing fraud that includes:
- Advanced technology — to quickly gather data
- Statistical intelligence — to determine which data patterns are suspicious and worth further review
- Sophisticated human analysis — to help businesses see the overall pattern (not just the details) and improve order acceptance
This approach helps make Clearsale different.
Technology, though impressive, isn’t enough to combat fraud on its own. When merchants take a holistic view of their sales landscape – which includes both technology and human review – they can increase fraud identification accuracy and keep up with a rapidly changing market.Clearsale combines big data analytics, statistical intelligence and human expertise to offer a precise balance of fraud protection and maximized sales. Contact our fraud prevention analysts today to learn more about our approach.