What Should You Look For in A Fraud Management System?

Your fraud management system needs several key features to be truly efficient. Let’s see what they are.

Businesses of all sizes now have access to a plethora of fraud management systems, which is why the market growth is so impressive. But it doesn’t mean they’re all equal. Some are legacy solutions that are costly and use antiquated tech. Others are more modern and agile, but might try to sell you tools you don’t really need.

So in an effort to help you find the right tool to protect your business, here is what we believe are key features for a great fraud management system.

How to Pick the Right Fraud Management System?

Your fraud management system must be fluid across departments so you’ll need to be able to create multiple logins for your team, with various permission levels. Having a centralized GUI (graphical user interface) where you can control, manage and communicate ideas about how to fight fraud utilising each department.

Depending on your business requirements, your fraud department might need a full end-to-end system or could work better with a multi-layered approach built up of differing products. 

The 7 key features of a Fraud Management System 

Team Management Features

Chances are that you’re not trying to mitigate risk by yourself. If you are part of a fraud detection team, you’ll need to be able to all access the platform, without stepping on each other’s toes.

So you’ll need to be able to create multiple logins for your team, with various permission levels.

Another area to mention is the GUI. If your fraud prevention tool comes with its own dashboard, you’ll want to be able to customize it and add notes for other team members. It should be a centralized place where you can control, manage, and communicate ideas about how to fight fraud for every department.

Integration Flexibility

Some businesses will need a full end-to-end fraud prevention system. Others will only need certain modules to create multi-layered security that accomplishes the following:

• Enrich data: some fraud managers only need to access extra databases to build a more complete profile of their users
• Meet scalability issues: if you have a previous fraud prevention system that hasn’t scaled as fast as your user base, you can add a third party system to help
• Patch holes in the line of defense: No need to build a brand new fraud management system if your old one is 90% efficient. Sometimes all you need is a 3rd party module with specific goals, such as phone or email lookup, device fingerprinting or IP analysis to bring it up to 100%
• Accelerate manual reviews: manual reviews are the most resource-heavy part of management, both in terms of time and effort. Adding an extra module could speed it up, if it is integrated properly.

The key point is that you needn’t reinvent the wheel. A good fraud management system should give you plenty of flexibility on how it works with your business, which is why SEON offers:

• REST APIs for developers
• GUI
• JavaScript snippet or mobile SDKs for our Device Fingerprinting tool
• Complete custom integration
• And even a one-click Chrome Extension for data enrichment

And as you can see with our integration workflow, even deploying a full end-to-end solution is surprisingly quick, allowing you to start flagging risky users in record time.

Data Enrichment

Modern online businesses don’t have the luxury of long onboarding processes and lengthy credit checks as it increases churn. This is true of any vertical, whether you’re a cryptocurrency exchange, an iGaming operator, or a loan provider.

So how do you get a complete picture of your users with minimum amounts of effort on their part? With data enrichment. Quite simply, it’s a process that will take user-submitted data points, link them to external databases, and give you a better idea of who you’re letting join your platform.

• Have they been blacklisted before?
• Are there discrepancies between datapoints, like a billing address in one country and shipping address in another?
• Are they using suspicious tools like VPNs or emulators?

As we’ll see below, data enrichment is what makes prevention faster, frictionless, and adaptive.

Predictive Risk Scoring

A key component of your prevention tool will be risk scores. These are calculated using a set of rules, for instance:

• Users who attempt too many login attempts in a row increase risk
• Visitors using TOR and virtual machines which increase risk
• User behaviour and previous transactions within the site 

For these rules to function, however, you need two things: a lot of user data, and the ability to test the results.

Regarding the first part, you should be able to leverage data enrichment in order to gain a better picture of who your users are, even with few data points.

It’s also possible to start with preset rules. Just be careful that your fraud management system has rules that were built for businesses similar to yours. If they are too niche, or not tailored for your vertical, you could be wasting a lot of time trying to recalibrate them.

Custom Rules, Fields and Parameters

Of course, you also want the ability to create custom rules. These are essential for improving fraud prevention efforts, as they let you:

• Adapt to new attack vectors
• Uncover fraudsters who have learned the latest ways to imitate legitimate customers
• Detect fraud attacks that are very specific and don’t follow a larger pattern

But adapting your line of defense is only really possible if you can work with data that’s relevant to your business. Which is why, once again, flexibility from your fraud management system is key: you’ll want to be able to work with any kind of data, from cryptocurrencies to shoe size.

Machine Learning Engine

Manually creating risk-predicting rules and testing them works wonders for most businesses. But how do you ensure you can scale your efforts when attacks increase in frequency and sophistication?

The answer: machine learning. While it is a bit of a buzzword in the tech world (and especially fraud tech), all you need to understand is that:

• AML engine uses an algorithm that trains itself to update the rules or suggests new rules
• You should be able to read the rule results in a way that makes sense
• That means using a whitebox solution, that makes it easier to oversee the AI with human intelligence
• Alternatively, you should be able to talk directly to a data scientist to support your use of the engine

At SEON, we have worked hard to create a powerful ML engine that suggests rules in a human-readable form via decision trees. This gives you complete power over the algorithms, while enjoying the benefits of a powerful system that can process huge amounts of data in real-time.

Real Time Results and Dynamic Friction

In fraud prevention, time is of the essence. If a fraudster can slip through the net and process a transaction with a stolen credit card number, it’s already too late. You risk losing goods, paying hefty chargeback rates, failing to meet anti-fraud regulations and even paying fines for it.

This is why SEON delivers results in a few hundred ms, so you’re never left playing catch up with the bad guys.

But it’s not just about flagging fraudsters fast. You also want your legitimate users to go through the net without delays. In short, you don’t want your fraud prevention system to be too stringent with good users. So how do you ensure only the risky actions are verified? With dynamic friction.

The best way to think about it is about two kinds of KYC (Know Your Customer) processes: light ones and heavy ones.

With dynamic friction, you can allow every user to enter the minimum amount of information (light KYC). But your fraud prevention system is still working behind the scenes to calculate risk. Thanks to data enrichment, you can get a good idea of whether the customers are who they say they are.

And if you have any doubt? The system will let you know, so that you may trigger heavier KYC processes, such as:

• 2FA
• ID scan or selfie with it
• Official documentation upload

Because you’ve only asked the riskiest users to jump through these hoops, you can discourage fraudsters, and let others know that it’s for their safety. Those with a clean track record, meanwhile, will be able to onboard your platform, pay, or browse without any delays.

8 Popular Fraud Management Systems Alternatives 

There are a range of systems available with varying expertise and focuses to factor when making a decision along other with important variables such as integration, costs and time. See below some other products on the market today:

Ekata

Ekata provides fraud prevention and identity verification products with graph and network visualisation capabilities to help companies with handling masses amounts. Recently acquired by MasterCard, the company work with AliPay, Microsoft, Stripe, and AirBnb.

Riskified

Riskified specialises in eCommerce security with one of its key differentiators being their chargeback guarantee which removes a major pain point for the merchant. Clients include Gymshark, REVOLVE, Wish, Canada Goose and Peloton.

Ravelin

Ravelin is a UK-based fraud detection and prevention platform with a blackbox machine learning solution that aims to mitigate risk with minimal input from the merchants themselves, aside from the initial rule suggestions, they also offer graph visualisation.

Kount

Kount holds data of over 32 billion annual interactions across 250 countries, 75+ industries, 50+ payment processors and card networks. The company focuses on supporting retail companies with both offline and online presences.

Signifyd

Signifyd provides an end-to-end fraud prevention platform with a 100% chargeback protection model in place to increase automation and remove fraud liability for the merchant. They also offer integrations available across Salesforce, Shopify and Magento.  

Iovation / TransUnion 

Iovation is a device-based fraud prevention and multi-factor consumer authentication solution acquired by TransUnion in 2018. The company has a database with over 3 billion devices and 30 million reports of fraudulent activities. 

Sift

A former YCombinator accelerator beneficiary, Sift offers both a complete ‘Digital Trust & Safety’ suite utilising blackbox machine learning to streamline operations to remove the pressure on human resources. The firm’s main industries are fintech, retail, food and beverage.

Feedzai

Feedzai offers both a complete platform as well as individual solutions to retail banks and other industries. Its main service scans transactions as fast as 3 milliseconds in real-time to spot potential fraudulent activities.