Riskified is a software-as-a-service company that offers chargeback prevention and fraud management for mid-market and enterprise level businesses, specifically online retailers. The company boasts clients such as Aldo, Skullcandy, Burlington Coat Factory, Mattel, Swarovski, Prada, and more. With fraudsters increasingly turning to the internet as a result of increased in-person security due to chip cards, it’s more important than ever for ecommerce businesses to protect themselves.
Industries Served
Riskified works with a variety of online retailers, including businesses in fashion, ticketing / events, travel, gift card sales, electronics, and luxury goods. Businesses that have international clientele and those with digital items in particular may benefit from services like Riskified, as foreign transactions and items that are quickly re-sellable are particular targets when it comes to fraud.Fraud Review
Riskified offers an end-to-end solution that help reduce ecommerce fraud, but also prioritizes not affecting customer experience. The fraud review process utilizes machine learning and comes with instant decisions, as Riskified commits to a “frictionless” fraud review process. Continual order analysis helps refine fraud detection to ensure not only fraud prevention but also preventing false declines that were actually valid orders. After all, you don’t want to lose out on true sales from a false anti-fraud response.Fraud Detection Process
Of course, fraud prevention companies don’t typically go into great detail about how their systems work, as that would make for a less effective system. However, Riskified does provide this image showing the steps:
As you can see, fraud detection is a multi-step process. As the purchase is made, the process begins, undergoing several steps before arriving at the ‘decision’ regarding the transaction.
The company provides some detail about each step, saying that it first engages in device and browser fingerprinting, behavioral analytics, and proxy detection. It states that it then augments that information with additional details from in-house sources, social media, and third parties. It explains the “linking” step as utilizing linking technology to look for connections between the current order and all of the orders placed through the Riskified system. Lastly, the machine learning component analyzes all of the information gathered to make a determination regarding the order’s legitimacy. An “approve” means less chance of fraud, while a “decline” indicates a fraudulent transaction.
