Interview #2: Spotlight on Polish fraud detector, Nethone

Nethone is a product partner on the 404Partners network. As a machine learning-based fraud prevention Saas company, Nethone lets online merchants and financial institutions to understand their end-users holistically. With its proprietary online user profiling and ML technologies, Nethone effectively detects and prevents payment fraud and account take-overs.

When and how did Nethone first come about?

Nethone’s beginnings go back to 2015 when a group of data scientists, developers and experienced industry managers within the Polish IT group Daftcode decided to test their idea of creating a profiling tool to detect and prevent instances of fraud. We trusted we could make this work backed by machine learning (ML) models as opposed to an ineffective rules-based fraud management format. Our hard work paid off and in 2016 Nethone was born along with our proprietary Know Your Users solution. We have been continually improving our solution and growing our company ever since.

What are the tactics you use to identify potential fraudsters?

Our solution is to focus on the payment process from start to finish and analyse in real time the interactions between a user and the service provider. We use digital fingerprinting along with behavioral biometrics backed up by ML models to determine the user’s hardware and software setup and how they behave during each browsing session when making a payment. Over 5,000 pieces of data are analysed that can provide a clear insight as to whether or not the user behind the transaction is genuine or a cybercriminal hiding behind a smokescreen.

What financial losses can you predict for this year due to fraud?

In a global context, fraud is on the rise. It is estimated that eCommerce losses for 2021 will be $20B - an increase of 14% from 2020’s $17.5B. Although efforts are being made to combat fraud with ever more advanced technologies and political and financial initiatives, the professionalisation of fraud has made it easier to commit cybercrime. This is particularly relevant now as eCommerce has seen a rise in the number of new online shoppers who are not fully aware of the dangers of the internet and how to best stay secure during each browsing experience.

What are the types of attributes that you collect from users?

Our analysis takes place in real-time, during which we can establish if a user is making a coordinated effort to hide their true identity, location and even true intentions. For example, an average user with genuine privacy concerns may use a VPN as an additional layer of security, however, a fraudster will do as much as possible to cover their tracks, from spoofing their location, the hardware they are using, the type of browser they are using - put simply, we actively search for suspicious and irregular behaviours.

How did the covid-19 pandemic affect e-commerce and what would that mean for fraud prevention?

The effects of the ongoing COVID-19 pandemic have particularly had a huge impact on the threat of fraud as millions more people around the world opt to shop online - in many cases for the first time ever - as brick and mortar retail stores temporarily closed due to lockdown measures. Many new online shoppers were, and to an extent still are, unaware of the full scale of the online security threats. Despite cybercrime tools becoming more advanced and readily available, it’s fraud based on tried and tested social engineering techniques that has been particularly effective against this new wave of internet users. Fraud prevention is likewise continually advancing, however, effectively communicating and educating internet users about how to best stay safe online is also a key part of the fight against cybercrime.

How is Nethone beneficial for the user experience?

Our advanced fraud solution not only protects online users and their accounts, it provides eCommerce merchants with the ability to ensure a smooth payment process without having to use invasive tools to verify every single user. The result is that all necessary customer authentication processes take place in the background, without the need to use arguably unsafe SMS pass codes or fill out annoying CAPTCHAs (a simple slip up can cause a transaction to be blocked) that can frustrate customers seeking a fast payment process.

How do you adapt your features in a rapidly changing industry?

We have a dedicated team of data scientists, developers and intelligence specialists who continually analyse fraud threats in the online domain and in the payments process. This knowledge and expertise is used to continuously improve our fraud solution in order to match the tools and techniques deployed by cybercriminals. Likewise, our ML models are able to continuously improve on all past analyses, but with an extra helping hand from our experts, the effectiveness of the models increases in order to prevent fraud.

20 views0 comments