Arvato informs about fraud prevention in E-Commerce

Many fraud attempts in e-commerce only become apparent after an order has been completed. Fraudsters increasingly use methods such as artificially generated bulk orders, stolen identities, and automated bots to bypass traditional fraud checks. The international supply chain and e-commerce service provider Arvato explains in a statement how analyzing logistical data can help curb order, return, and service fraud.

Multi-level verification processes

According to Arvato, order fraud often relies on identity theft or automated attacks that trigger a large number of orders in a short period. In such cases, frontend checks are under immense time pressure, as decisions must be made quickly. Often, there is a lack of contextual information, such as the plausibility of the delivery address.

To address this issue, Arvato has implemented a multi-level verification process in the backend. This process begins immediately after checkout. Initially, rule-based controls are employed to identify known anomalies such as entries on blacklists or unusual order frequencies. In the next step, machine learning models analyze the data for hidden correlations. For example, velocity checks can be used to detect sudden increases in order volumes in specific regions or unusual clusters of deliveries to hard-to-verify addresses such as hotels or dormitories.

www.arvato.com