Refund Abuse Integration

Overview

What is refund abuse?

Refund abuse occurs when a customer uses your returns policy so much that it becomes unprofitable.

One of the most common types of refund fraud is when a customer receives an order, but claims it was never delivered and requests a refund. Determined refund fraudsters can even make a business out of exploiting returns policies.

Ravelin can identify behaviour that looks like refund abuse and provide recommendations on whether you should issue refunds.

You can learn more about refund abuse, and how to prevent it, in our Insights Guide.

How Ravelin protects against refund abuse

To prevent refund abuse, you send us details about refund requests using our API. We will evaluate whether the refund request is an instance of refund abuse, and return a recommendation as to whether the you should issue the refund.

These recommendations are generated using a combination of methods which are explained in further detail below.

  • Rules and refund features
    Using our advanced rules engine you can enforce your refund policy and also implement safeguards against refund abuse. We calculate many features which you can use in rules to help you identify refund abuse, for example, percentage of orders refunded, rate of refund requests and refund requests by address, device or payment method.

  • Network analysis
    Refund abusers often create multiple accounts to bypass refund policies. Our graph network can connect customer accounts by device, payment method or email address. If one account is identified as committing refund abuse, we can prevent refunds from connected accounts.

  • Machine learning
    Over time we collect evidence of refund abuse, and once we have enough data we will produce machine learning models capable of identifying refund abuse.

We continuously monitors the performance of our refund abuse recommendations and will suggest any improvements.

Next steps

Learn how to request refund abuse recommendations

Feedback

© Ravelin Technology Ltd. All rights reserved