How EQ8 Scoring Works

How EQ8 Scoring Works

Getting started in eCommerce is intimidating. Many business owners entering the market understand their products and customers, but these two things may have very little or nothing to do with the internet. Getting a website built, learning how to take payments, and managing orders can sometimes feel overwhelming.

As if understanding the internet wasn’t hard enough, there are also internet criminals and fraudsters to be wary of. As a new business owner, you don’t really want to spend time understanding the complexity of internet fraud on top of everything else.

That’s good! At NS8, we don’t want you to either – that’s our job.

Central to our ability to simplify eCommerce fraud prevention is the EQ8 Score. The EQ8 Score is a number calculated between 0 and 1000 that rates the potential risk of a website visitor. This is accomplished by watching the visitor from the time they enter your website to the time they leave.

As soon as a visitor interacts with your website, the analysis begins. When you need to make a decision, a real-time score is given. Highly suspicious traffic is given a low score and good traffic is assigned a high score. Once an EQ8 Score is received, you can act on the visitor’s score risk and segment traffic.

Okay, But How Does NS8 Score Traffic?

This is complicated. There are hundreds of variables that define what data is used dynamically as the EQ8 Scoring Engine learns about the visitor. In particular, different data sets are used to score traffic based on the country the user is located in, and as this data is applied, the score will be adjusted.

A simple example could be as follows: a visitor’s computer is indicating that it is a Windows 10 machine; however, the web browser they are using says it’s Apple’s Safari. Since Apple does not make Safari available for Windows 10 users, this would be viewed as suspicious. While this alone does not necessarily make this session fraud, it would then be combined with all scorable factors, like geolocation, behavior data, and internet address, to determine the visitors EQ8 Score.

Confused yet? Again, don’t worry, you don’t have to understand it all. We have some really smart people figuring this all out with the assistance of artificial intelligence. Let’s look at it from another perspective, one where Java is a drink and not a programming language.


It’s Story Time

Meet Marty. Marty runs a tavern. As a diligent barkeep, Marty pays attention to everything he can about his customers. He must keep out a variety of riffraff: hooligans, thieves, underage drinkers, and robots. Especially robots.

Looking for all these troublemakers while pouring pints of cold refreshing beer was a problem for Marty, so he devised some security systems.

He has cameras looking at the street outside his bar and the parking lot. Using these cameras, he can see what vehicles people arrive in. He can observe if they are acting suspiciously by putting on disguises in the parking lot. He can watch what direction they are arriving from. He can take note of the vehicle they arrive in.

In other words, noting the activity as a patron arrives, Marty can begin to identify risks as a guest approaches.

Once a patron gets to the door, they have to get past Marty’s bouncer. The bouncer checks their identification to make sure it’s valid, but also uses the opportunity to gauge a customer’s behavior. If a visitor’s ID is questionable, the bouncer is prepared to ask additional questions to establish validity. If the person is on a watch list as a known hooligan, they will get turned away.

So, the people the bouncer approves are allowed in the door, the robots and troublemakers can go elsewhere.

Regulars that Marty trusts are not much of an issue, nor are guests who do not exhibit any suspicious behavior. Still, he keeps an eye out to make sure a pesky robot isn’t impersonating someone else. If the person is a newcomer, Marty will use all the information he’s gathered to determine how to proceed. If they are walking and talking like a robot, they probably are.

Before handing over a cold pint, Marty also is going to check into how he’s getting paid. He’s going to make sure the payment is not counterfeit and that a payment card has funds. Only when all the data Marty has gathered adds up will he hand over a pint.

Oh, and if Marty is still suspicious, he can always go and ask for additional data. All of the data that Marty compiled and analyzed is similar to the NS8’s EQ8 Scoring Engine.

Wait, How Is EQ8 Scoring Like Bartending?

One of the most important points to note is that Marty the Bartender did not just use one factor to identify suspicious activity. He looked at the whole picture from the time the customer arrived up to the time they paid to determine how to react. Likewise, the EQ8 Score is a comprehensive score and not just a simple checklist. NS8 uses dynamic data sources to compare the entire interaction with a visitor to calculate a score.

Another important distinction is that Marty’s decisions were not based on being a startup bartender. Like Marty, NS8 is not new at fraud prevention. Our technology is built on over ten years of experience and compiled data in web analytics. The whole reason our founders got into fraud detection was to solve problems for businesses they were already involved in.

Also, like Marty, our customers are not locked into a set of rules and black lists without control over the outcome. Using NS8’s tools, our customers can choose how to react to an EQ8 visitor score. This allows a system tailored to your fraud risk tolerance and customer profiles.

Perhaps the most important similarity is that the technology behind the EQ8 Scoring Engine is not static. Using machine learning and other data science techniques, our team is making sure that we stay ahead of the bad guys.

What Can an EQ8 Score Be Used For?

EQ8 scoring has many applications in the world of eCommerce. Merchants can use an EQ8 Score to fight advertising fraud, including retargeting fraud, as well as transaction fraud. Another application for a merchant would be to use data gathered in the scoring process to target and segment customers.

For example, a software store may choose to offer a promotion on Apple software to customers using a Mac or offer PC products to Windows users. Many factors are identified that can be used to upsell, target, and market to customers while fighting fraud.

Post Author: John Brown

John Brown

Solutions Consultant at NS8. After years in the payment industry, John now lends his expertise to NS8 clients by helping them implement effective fraud protection.