The holidays are an intense season for anyone in e-commerce, but those in charge of reviewing incoming orders for fraud really have their work cut out. Not only is there a dramatic increase in the overall orders, but Riskified’s data from previous holiday seasons shows that fraud on entry increases significantly as well.  Whether this increase in fraud is driven by fraudsters taking advantage of dramatically increased sales or by consumers committing “friendly fraud” due to stringent returns policies, handling these orders correctly is a challenge.

A properly prepped fraud team, armed with the right tools, can help merchants recapture revenue from transactions they might ordinarily decline.

Following are four steps merchants can take to ensure optimal performance during the holidays:


1. Prepare your team for the spike in order volume

The biggest challenge risk management teams face during the holidays is the sudden, dramatic spike of incoming orders. It is imperative that orders be processed expeditiously, otherwise the back-log could have customers waiting several days for order confirmation.

Therefore, it is important to analyze the make-up of your team to understand its capacity and ensure it can handle the projected incoming volume.

Establishing rules or guidelines for which orders analysts can approve independently and when they must seek a second opinion will help achieve optimal performance even during the busiest of times.

The most common mistake made in this context is setting limits only to which orders can be approved by certain analysts while allowing everyone, including inexperienced team members, to decline orders of all types and values. A junior team member will naturally be apprehensive about approving a high value transaction, and turning away a good customer can have a significant negative impact on your bottom line – especially when taking into account the customer’s lifetime value.



2. Use smart technology to route fewer orders for manual review

Leveraging technology is critical for maximizing efficiency. Setting up your systems to auto approve legitimate transactions is the easiest way to lighten the load on your team.

Behavior analysis is key to identifying legitimate transactions, but it can be difficult for a machine to interpret. However, there are clues you can program into your systems to help them distinguish friend from fraudster. For instance, did the buyer use a promo code, compare features of similar products or read the warranty and/or returns policy? Were they referred to your store through an email marketing campaign? Fraudsters using stolen credentials are neither going to engage in such on-site activities nor be included in your lead database.

The unique characteristics of holiday shopping must also be taken into consideration when preparing your systems for this period. Consumers wait all year for the holiday sales, when they splurge on high ticket items and purchase gifts for family and friends across the country (or the globe), often at the last minute. Thus, expedited next-day shipping, high order values, and a shipping address that does not match the billing address, all of which might be considered to be fraud indicators, may be perfectly legitimate in the context of holiday shopping. Creating new rules and adjusting existing models to reflect the unique patterns of the holidays is highly recommended.

If you don’t have a models- or rules-based system, you can create a batch query that runs every few hours on incoming orders. Even printing out a help-sheet that your team can keep by their work stations will help ensure they don’t invest more time than necessary on easily approvable orders.


3. Improve the efficiency and accuracy of the manual review process

The problem most merchants face is telling the difference between potentially fraudulent and blatantly fraudulent orders. Caught between a rock and a hard place, some merchants feel declining a questionable transaction is safer and less costly than approving a bad order. Safer – yes; less costly – not necessarily. By nature, we are risk averse, and more likely to remember a chargeback than an order we rejected. But if merchants stop to consider it, they may find they lose more revenue due to fear of fraud than they do to actual fraud.

To avoid false positive declines, your system should only auto approve, never auto decline. Naturally, this could lead to a large number of transactions being routed to manual review, so it’s imperative that analysts are empowered to review these transactions as accurately and quickly as possible. The more time spent on any particular order reduces team efficiency and increases consumer frustration.

A useful process is to analyze past orders that took a long time to review, and to investigate what caused the delay. Specifically, look into cases where the review time was long and the transaction was ultimately approved. Figure out where the team got “stuck” and which data was used to eventually approve the order. Would displaying data differently or flagging certain information have helped? Perhaps online tools, such as Whitepages, Emailage or Maxmind can assist in dealing with such transactions going forward?

Previously declined orders should also be reviewed to see which transactions could have been approved. Knowing why a transaction was declined can expedite the processing or approval of similar orders. At Riskified, we find that on average, 66% of orders typically rejected by merchants can and should be approved, which equates to a big boost to a merchant’s top- and bottom-line.

If analysts do decline an order, it’s important they note their reason for declining and communicate it with customer service, who will likely field any follow-up contact with the buyer. In certain cases, a declined transaction could be approved with just a few additional pieces of information.


4. Have a strategy for international orders

International markets can have a dramatic impact on a merchant’s business growth.  However, reviewing international orders manually can take an exorbitant amount of time due to language barriers, unfamiliar data sources and use of alternative payment methods. Over the holidays, many international customers will want to take advantage of the attractive sales and promotions, and you may see an increase in non-domestic orders.

Now is the time to review prior international orders and segment approved, declined and fraud-related chargebacks by country or region to identify the most important non-domestic markets for your business, as well as your weak points. To better handle such orders going forward, familiarize yourself with local data sources in these markets (local white pages, local social media networks, etc.).

Many merchants have higher decline rates for international markets because they believe these orders are not worth the time and/or risk to approve. In this instance, you might want to consider partnering with an outside company that specializes in validating high-risk orders.

There’s a lot of truth to the business axiom that ‘you have to spend money in order to make money,’ but in Retail, merchants seem resigned that they have to lose money in order to make money – either through fraud or fear of fraud. Ramping up your risk and fraud capabilities now – whether by fortifying your existing infrastructure and/or by outsourcing it to experts – can deliver significant ROI.

Not only will it protect your organization from fraud, but it can also potentially convert more borderline transactions that might have otherwise been declined and build customer loyalty as we head into the busiest shopping days of the year.