Online payment fraud is when a fraudster uses stolen card details to make purchases online. Fraudsters can buy large numbers of stolen card details on the dark web and then attempt to use them to make purchases.
Fraudsters can use payment fraud to get products for free for themselves or they can monetise payment fraud by selling the goods for a profit.
You can read more about online payment fraud in our Insights Guide.
To prevent payment fraud, you send us details about orders using our API. We will decide if we think the order is fraudulent and return a recommendation as to whether the you should accept the order.
Ravelin produces these recommendation using a combination of machine learning, graph networks and rules.
Ravelin builds machine learning models, based on your specific data, to stop payment fraud. Machine learning is the best way to stop attacks; models are adaptable, scalable and less easy to circumvent. Models also allow you to optimize conversion.
Every business is different and each custom machine learning model will care about different things.
Our payment fraud models take into account hundreds of customer features, for example:
Transaction and order history
Does the customer usually order from this restaurant? Do they usually spend this much?
Location information
How does the customer’s billing address differ from their delivery address?
Device information
Have they used this device to make an order before? Has this device been used by other customers?
Behavioural data
What did the customer do on the checkout page?
Customer specific data
What email domain did the customer use? How old is the account?
Ravelin continuously monitors the performance of machine learning models, retraining them, adjusting thresholds and identifying and adding important features.
You can learn more about machine learning for fraud detection in our Insights Guide.
We monitor customer connections and how close customers are to known fraudsters or chargebacks. This allows us to identify fraud in many ways, for example:
Networks growing bigger quickly
There are some cases of small networks of genuine users - a family sharing a device or a team using a corporate credit card.
But these networks remain static and rarely grow any bigger, or if they do it happens slowly. A fast growing network is almost
always due to fraud.
Lots of widely shared cards, devices or email addresses
It’s very rare for genuine customers to share a device, card or email address. We’ve seen fraud networks with over 800
accounts sharing a single payment method, and networks showing account takeover where over 10,000 customers appear to be
sharing one single device.
Lots of chargebacks in the network
We allow our clients to disregard any genuine chargebacks when they upload their data to Ravelin Connect, so we use a
chargeback node as an indicator of fraud. This means if there are any chargebacks in a network, all the network’s users are
fraudsters.
Machine learning models and graph networks are mutually reinforcing. For example, you can teach your machine learning model to flag large networks for review and to block payments from networks which have grown super quickly, to prevent a fraudster from using multiple accounts to order goods.
You can learn more about graph databases in our Insights Guide.
Rules can be an effective way to either enforce business logic at checkout and to add payment fraud safeguards.
For example, rules can allow you to:
Enforce business policy
If there are markets you don’t operate in, you can block all orders from IP addresses in certain countries.
Act fast to stop an attack
Fraud analysts can use rules to quickly stop a fraud attack whilst it’s happening.
Proactively block new fraud trends
Fraud analysts could use rules to protect against emerging fraud trends,
before the machine learning model adapts to this fraudster behavior.
Allow good customers
Rules can be used to allow and not just to prevent.
Rules can be used in combination with machine learning, for example, if you want to send customers using a new device to 3DS where the machine learning score is below a certain threshold.
Ravelin continuously monitors the performance of rules and suggests any improvements.
Each of the integration steps is shown below. We have a guides for each integration step. You should work through each of these guides to complete your integration.
Getting started
Learn the basics of how to request a payment fraud recommendation and view the result in the Ravelin dashboard.
Integration Process
Learn how we will work with you closely throughout your payment fraud integration project.
Choose a payment flow
Decide at what point during your payment flow that you are going to ask for payment fraud recommendations.
Send device information
Send us customer device information by integrating our JavaScript and mobile libraries into your website and mobile apps.
Request recommendations
Gather customer, payment method, order and transaction data and request payment fraud recommendations at payment method registration and checkout.
Send updates
After acting on our recommendations, provide order and transaction updates.
Send disputes
Send us your disputes and chargebacks so we can learn about the type of fraud you’re experiencing and provide better recommendations.
Test your integration
Test you are correctly handing our recommendations.
Error handling
Ensure you are correctly handling errors so that your integration is reliable and resilient.
Going live
Decide on how you are going to enable payment fraud recommendations on live traffic.
Was this page helpful?