Besides the high pressure to innovate – such as to prevent bot purchases and avoid the general danger of the misuse of credit card data, the Covid-19 restrictions have resulted in high order volumes and new shipping processes, which in turn lead to increased fraud and identity theft in online retail. The entire e-commerce sector faces the challenge of dealing with fraudulent orders in order to minimize the associated costs and maximize the purchase experience for genuine customers.
Our customers in e-commerce use a wide range of technical solutions to prevent fraud. AI-based engines work effectively when there is a high volume of data. Database reconciliation and identity checks protect end customers from the consequences of data theft. Credit checks can obtain data for offering the customer the most appropriate payment methods already during the shopping process. In short: a highly coordinated interplay of providers from the payment universe is key to preventing fraud and achieving optimized conversion. Technical automation allows up to 90 % of all orders to be checked for fraud, while the leading e-commerce merchants count on case management for the remaining 10 to 15 %. This way, the checkout process remains quick and convenient, and customers are not deterred from purchasing by additional hurdles. What’s more, no delays arise before the customer can receive their desired goods at home.
Case Management – what do that involve?
Our colleagues specialized in data analysis check orders at an individual level based on a thousand statistical decision points. Let’s consider an example. Imagine I am currently on vacation and using a new smartphone. After making new friends, I decide to take up a new hobby and would like to buy my own surfboard straight away. For me, this would represent a completely untypical order for a high-priced sports good, a new delivery address different from where the order was made, as well as a new phone. These are all criteria that could potentially indicate a case of fraud. My order would therefore be suspicious. But if the shop would reject me as a customer for this reason, I would no longer order any new equipment for this new hobby there in the future. This is a classic case that requires a manual order review. Naturally, not all the 1000 decision points mentioned above would be applied to my case. In this hypothetical scenario, a data analyst would review the address and payment option: a hotel by the sea, goods receipt and room number of the hotel as the delivery address and a reliable credit card that I have been using for years. I would receive my surfboard and would hardly have noticed anything about this check, since there was no delay in the order process. Our employees take care of this process for our clients with an accuracy of over 98 %. This security also has a positive effect on the brands of our merchant customers: that’s because they would like to shop at a store that does not allow any fraudulent orders. If fraudsters had the ability to pay with stolen data, this would have negative implications for the brand image.