Scoring Engine and Sandbox

Overview

The Scoring Engine is the heart and soul of fighting fraud with SEON. However, it's often best not to start testing things live from day one. Luckily, our sandbox function will allow you to test everything out before flipping the switch on any new rules or your complete integration.

 

Understanding the Scoring Engine

The Scoring Engine is the heart and soul of SEON's risk prevention toolset. It's the system that takes all the information our data enrichment solutions unearth and decides which transactions are good and which are bad.

In SEON, each transaction is given a fraud score and a state. The fraud score indicates the level of risk associated with a transaction. The higher the score, the more likely a transaction is fraudulent.

A transaction can have three states: APPROVE, REVIEW and DECLINE. This state will change when the fraud score exceeds the pre-configured thresholds of each state. You can also create rules that directly change the state of transactions that trigger them.

Rule types

SEON separates rules into three categories: default, custom, and machine learning.

Default rules are curated by our team and are the foundations of your SEON integration. While you cannot edit the rules themselves, you can quickly turn them on and off or change their score values in the Scoring Engine.

Custom rules make SEON infinitely customizable as a fraud prevention solution. You can start creating rules on day one, either on your own or with the help of our seasoned Customer Success team. Custom rules range from simple statements to complex velocity and regular aggregate rules.

Machine learning rules are created by SEON's whitebox machine learning solution that trains itself on your data to provide insights into your specific use case and counter the types of fraud your business faces every day.

Rule categories

Rule categories act like folders for your rules. Collect rules designed to catch certain types of fraudulent activity or flag specific transaction characteristics. Rule category settings also control how a rule category will affect a transaction's overall fraud score or final state.

You can use rule categories to calculate a category score. SEON will return this category score in your API response and in the Admin Panel. Category scores can also trigger category states, which can be used to automate fraud prevention in your systems.

Creating your unique rule set

We strive to make SEON as customizable as possible, and your freedom in creating rules to meet your needs is at the core of these efforts. Of course, freedom can be daunting at first, so our Customer Success team will help you from step one.

We'll assist you with planning and creating custom rules for your SEON account from the day you start your journey with SEON. Our fraud experts will leverage their broad industry knowledge and global experience in fighting fraud with SEON to adapt their best practices to your use case.

Rules created by humans are only one piece of the puzzle. Our Machine Learning tools provide the others. Of course, machine learning algorithms need data to learn from. While our Blackbox model will work from day one thanks to the base model, our Whitebox will only work with your data. That's why we recommend uploading your historical transaction data to SEON.

Working in the SEON Sandbox

The SEON Sandbox is a testing ground with no consequences. Everything you do in the Sandbox Mode is completely separate from your live environment. As a result, the Sandbox is where you can try new rules, play around with different rule categories and find the best way to fight fraud with no risk.

Sending data to the Sandbox

Because the Sandbox is entirely separate from your production environment, you'll have to populate it with data specifically.

The easiest way to fill your Sandbox Database is to send API requests authenticated using your Sandbox License Key. You can also populate it manually by completing manual checks in the sandbox.

When integrating the SEON API, your teams can configure your systems to send data to the Sandbox automatically. If you are actively testing various solutions, it can be best to duplicate your entire production database in the Sandbox to ensure accurate results.

Working in the Sandbox

  1. Click your user name in the top right corner of the SEON window
  2. Find and switch the Sandbox environment toggle.
  3. A yellow bar will appear at the top of your tab and stay there as long as you remain in the sandbox.

What happens in the Sandbox, stays in the Sandbox. The changes you make here won't affect your production integration. Have a complex velocity rule you want to try out or a new rule category to test so fraudsters can't abuse an upcoming promotion? This is the place.

When you're confident that everything is working as you'd like, you can copy rules from the Sandbox to your production environment using the multi-select feature.