27 December 2023
7 min read
Creating complex rules can be a streamlined and intuitive process. Let's dive in to see how Nethone’s rule engine features can help you to:
Our decision engine stands as a hybrid and flexible tool leveraging both your data as well as insights provided by our profiling solution - thousands of unique user attributes and detects hundreds of browser and mobile-specific risk signals. By leveraging a large pool of data from our network as well as from yours, you can fully customize your decision logic to fit your needs. This can be achieved by selecting or creating rules and models and building your own set-up. We are about to show how our rule creation features enable greater precision in detecting risky users and allow you to build complex expressions easily.
The Rule feature improves the rule engine capabilities, letting you finely specify the data you want to include for fraud assessment. You can connect multiple expressions using logical operators, such as AND and OR, with an unlimited level of nesting. Whether you need to combine multiple conditions to trigger fraud alerts or fine-tune your screening process, you can craft in-depth rule sets that align precisely with your fraud detection requirements.
Let's say you'd like to block transactions that match a group of behaviors:
Nethone's advanced rule builder enables you to formulate a complex condition such as:
“Flag orders that are over a certain amount AND are either made by an account that is newly created OR has placed several purchases in a short time. Also, make a separate alert for any orders placed at odd hours that have shipping addresses different from their billing addresses.”
This allows you to pinpoint specific scenarios that are more likely to be fraudulent based on patterns you've noticed in the past.
The Rule provides an extensive range of configuration options called expressions. Think of expressions as customized statements based on different criteria or attributes.
These attributes can range from simple fixed value aspects like the user's geographical location or purchase amount to more variables like the time or frequency (velocity) of transactions in a given period. Other attributes can include 'distance' (comparing geolocation distances) and 'text similarity' (matching patterns in entered data such as email or address).
With hundreds of attributes available, you have a vast array of possible permutations, particularly when combining multiple attributes or using attributes with filters. This creates the potential for tens of thousands of unique options when constructing a statement.
Our rule engine provides a powerful combination of automated rule-based procedures and human-driven fraud intelligence, ensuring you stay one step ahead in the fight against fraud
What’s more, you will soon be able to leverage your historical data to test and validate your rules before they influence the final recommendations. Evaluating the rules against past data ensures a more reliable and accurate system while mitigating the risk of incorrectly blocking or flagging legitimate transactions.
When creating rules, it's essential to consider historical data too. The initial rule creation should be based on a selected set of data, and then further refined against another distinct set of historical data.
A common pitfall often observed is that the rule has been trained and tested against the same period and/or data. This practice can result in biased outcomes that don't account for potential variability in other periods. We recommend testing the rule against an entire year's worth of data where feasible. This approach provides a clearer view, especially for addressing specific new fraud trends where historical patterns may not be representative.
On top of all, to make the best out of a rule engine means quickly adjusting rules in response to changing fraud patterns. Our rule engine is designed with the flexibility to modify and update rules to accommodate emerging patterns and trends in fraudulent behavior. You can tweak the rules' variables, expressions, or thresholds, ensuring they remain effective against the prevailing types of fraud.
This agility allows you to bridge the gap between detecting a new type of fraudulent pattern and implementing a rule to prevent it. You maintain control over the dynamic fraud landscape without becoming overwhelmed by its changes. Take, for instance, this habit where fraudsters make several small purchases to stay under the radar before making a significantly larger one. With our adaptive rule engine, you can quickly set a rule to track such unusual purchasing behavior.
This ability to quickly adapt to changing fraud trends, combined with the methods to create in-depth rules and execute testing without influencing transactions, ensures your fraud prevention system stays tuned and capable of detecting even the most creative fraudsters.
You can take advantage of an intuitive user portal designed to serve fraud professionals regardless of their level of experience. The dashboard offers a customizable decision engine - where you can create virtually any rule on data processed within our system - and a complete real-time 360° view of user behavior