Features Analytics participates together with The Fraud Practice in the "Enabling Custom Modeling & Analytics for the Modern eCommerce Merchant" US webinar

(Nivelles, Belgium / Sarasota, Florida – December 4, 2014) – Features Analytics participated at the webinar “Enabling Custom Modeling & Analytics for the Modern eCommerce Merchant” organized by The Fraud Practice. The Fraud Practice hosted a webinar to discuss the increased accessibility and use of custom modeling in eCommerce risk management. Cristina Soviany has explained to Ecommerce fraud department and IT heads how custom modeling and eyeDES® technology can respond to their needs.

The webinar was accompanied by The Fraud Practice's latest free white paper titled "Enabling Custom Modeling & Analytics for the Modern eCommerce Merchant." You can download the paper here.

David Montague, an industry renowned expert, discussed the particularities of fraud scoring techniques and custom modeling techniques. Cristina Soviany explained to the webinar audience the advantages of using custom modeling in solving today’s challenges.


Fighting online fraud in today’s complex and competitive environment is becoming increasingly difficult, particularly for merchants. While merchants are focusing on increasing the sales, they also need to keep the risk of fraud loss under control and to ensure the protection of their brand. Merchants are often using fraud detection techniques which are not effective enough. Therefore they are absorbing too much risk and having too much cost to deal with the fraud. In the recent years, great progress has been made in the field of advanced data analytics and modeling technologies. As a result, today Merchants can have access to state of the art custom models, great technology that provides higher accuracy at an affordable cost.

About Custom Models

Using custom models to score transactions or orders is extremely effective. While fraud scoring models are built on aggregated data coming from other merchants, custom models are different, they are built on customer data. Using custom models, the organization is able to address its specific business issues and risk channels, and this is in a much more effective fashion. Custom models can provide higher detection rates and reduce the number of insults. Those have a high impact on your sales conversion and generate additional manual review costs. Statistical custom models can be used standalone or in combination with other fraud detection techniques that you are using in the organization. Custom models should be built by experts. Experts have the required background and knowledge to employ state of the art predictive modeling technologies. They can provide latest ideas to build a fully optimized statistical model that will fit organization’s business demands and KPI’s.

About Features Analytics

Cristina Soviany, Features Analytics CEO, has stated that to put effective custom modeling in place an organization will need 3 key components:
  • First, one needs access to state of the art predictive modeling tools that experts or your analytics department will use to design the models.
  • Secondly, the organization needs a platform on which to deploy and execute the models, preferably able to deliver scores in real time.
  • Third, the organization needs a team of modeling experts that can apply advanced modeling technologies and ensure the necessary support for implementation and retrain of the models.

Cristina Soviany, Features Analytics CEO, quotes

"At Features Analytics, we provide eyeDES. This is a fully integrated platform where our team of experts designs the custom models for you. As soon as they are designed, the models are immediately deployed on the execution platform that we also provide. Additionally, the platform comes with a dashboard, a real custom business intelligence tool for model performance and fraud trends monitoring of your business data."

"Features Analytics team of experts builds the custom models using an innovative statistical pattern recognition technology. By using this technology, our experts have the ability to understand your business data, identify business segments or data channels that need specific models to deliver different KPI’s. The models can be immediately deployed on eyeDES® execution platform which delivers scores in real time."

"We are very proud of having developed a platform that streamlines the implementation of models, meaning that models, after they have been designed by our experts, they can be immediately executed on the platform, with no additional coding effort. This cuts down significantly the model development time and thus it reduces the overall cost.

"Additionally, besides the scores, reason codes are also delivered in real time. The reason codes stand for useful explanations that, together with the score, can be used by your existing fraud scoring techniques or team of fraud analysts to streamline the manual review process. For example, they can use the score in combination with one or more reasons to approve the order or handle better the call with the organization Customers / Consumers."

"Our eyeDES® fully integrated platform comes with a dashboard via which you have access to custom analytics and business intelligence on your data. You can continuously monitor your data fraud trends and the models performance."

"Features Analytics experts are ready to retrain the models on your demand. In our case, the integration is easy and you have a choice. eyeDES® platform can be installed on your premises or you can subscribe to eyeDES® services via cloud."

"In today’s complex environment, in order to effectively address the increasing challenges, adopting a custom modeling solution is a realistic approach and will represent a key advantage for your business."


If you want to learn more about custom modeling please download the white paper at this link.