eCommerce Fraud Detection: Five Recommended Order Hold Rules

Earlier this year we saw record-low reports of fraud rates related to online retail. In January Cybersource reported 1.2% of revenue lost to fraud in 2009, attributing the improvements to the way retailers are managing risk.

With many tools, services, and technologies available for retailers to manage fraud, we wanted to provide some simple tactics to be deployed within the Order Management System. This post provides five great order hold rules for every high volume online retailer.

But first, lets explain what we mean by order holding within the eCommerce Order Management System (OMS) as part of your anti-fraud strategy. We're assuming that the fraudster has successfully passed all "frontend" validation checks, Address Verification System (or AVS), and other "card not present" technologies like Card Security Code (CVV2 for example) or even PIN systems like Verified by Visa. In our example, we basically have valid payment details and the order will be shipped unless we stop it within the OMS. So how does a merchant protect themselves without manually checking every single order? The answer: creating custom business rules within the OMS to trap and hold orders that meet certain criteria.

Within the OrderDynamics eBusiness Manager, we provide a comprehensive Fraud Detection engine that allows merchants to create custom detection rules. Each rule is defined using over 25 parameters and can be chained using AND/OR operators to form a "Fraud Detection Filter". All "payment authorized" orders are required to pass through these filters prior to being cleared for fulfillment. Orders that are caught by the filter are placed in a special queue called "Pending Fraud Validation" requiring a manual review by the customer operations team. As the merchant's business changes or they want to change their approach to fraud, they can easily modify their filters (rules and chains) over time.

Five Recommended Order Hold Rules

Rule #1: Bill To Address does not match Ship To Address

This can be very effective but may also flag too many orders for fraud review depending on your business. For example, if gifting is a big part of your business this rule may cast the net too wide. However, these rules can chained together so you can easily tighten the scope using AND/OR operators with other rules.

Rule #2: Overall Order Value Greater than $X

This allows retailers manually review orders of a certain value that they may not be comfortable being completely automated. Customer Service Tip: Retailers should take the necessary steps to ensure their high-value customers are getting "white glove" service. This rule will give them the immediate visibility into their largest customers.

Rule #3: Customer Account Charge Backs Exceed 1

Merchants should automatically hold orders for manual processing if the account has a history of charge backs.

Rule #4: IP Addresses or Billing and Shipping Addresses Out of Country

There are fraud hot spots around the world. Flagging shoppers from uncommon shipping destinations is a great way to avoid fraud.

Rule #5: Volume of Products, Categories, or Brands

With popular, and expensive, brands and products (for example iPhone 4G by Apple) available for online ordering, it is often a good idea to flag orders with high quantities of these brands, products, or categories from a single customer.

By chaining these 5 rules within the Order Management System (OMS), merchants can create comprehensive, business-specific Fraud Detection Filters sure to reduce fraud without costly manual effort.

For more information about OrderDynamics, contact Michael Turcsanyi.