Maxpay launches fraud prevention solution Covery
March, 2017 – International payment processor, Maxpay, announced the launch of their fraud prevention solution Covery. Covery helps merchants to protect their online business from various types of fraud and reduces operational overhead by up to 95%.
When a merchant has a weak anti-fraud system he doesn’t only lose profit, he also should be ready to deal with the following issues:
- Fraud alerts and chargebacks
- Penalties and sanctions from Visa/MasterСard
- Manual reviews of transactions
- Lower conversion
- Customer complaints
Covery brings together event chain analysis, feature engineering, rule management and machine learning to obtain the most accurate results. The solution is fully customizable and suitable for any industry, merchant business model or traffic source.
Covery currently analyzes over 5 000 000 actions daily in both high and low risk segment. The system only needs 0.5 seconds to complete an analysis of a customer and make a decision. Covery is scalable, so it can easily increase capacity as needed.
How it works:
Covery starts from data collection.
How the system collects data:
The majority of anti-fraud solutions only collect and analyze specific pieces of the customer’s information. They do not cross-reference the results. To prevent e.g. fraudulent transactions Covery starts gathering information about a customer’s actions as early as their first registration on the merchant’s website and analyzes the hole event chain. The analytical scheme which is applied by default to each website consists of the following steps: 1) Registration (to reveal fake accounts) – 2) Confirmation (fake accounts prevention) – 3) Login (to reveal hacked accounts) – 4) Payment (fraud prevention) – 5) Pay out – (fraud prevention). Moreover, the chain is easily customizable so the merchant can add any event or parameter in it, required by the type of their business.
What data the system collects:
Covery’s approach is based on feature engineering. The system collects and then analyzes not only the customer’s basic parameters such as e.g the card type and geolocation, but also any other aspect, including custom, aggregated and complex characteristics. Covery is able to use any information and logics used by merchants, even historical data, that can be easily uploaded right into the system and come into account during the decision making process.
The complete customizability allows creating unique features and analytical scenarios, taking into account the type of the merchant’s business and predict the customer’s behavior.
Step 1 – Trust list check
The system verifies users through Trust list, a global database of fraudulent and reliable identifiers to stop known fraudsters and enhance the experience for legitimate users. Covery is integrated with the key globally-known vendors of data to make process even more accurate.
Step 2 – Rule-based risk score
The second step of the check uses rule-based scenarios. At the end of the process the system evaluates each customer and assigns them a score. Covery uses a scoring system that goes beyond just negative scores. It assigns each customer a score on a scale from -100 to 100, where -100 is a completely trusted customer and 100 is fraudulent customer. The system then saves the data so that it can be sorted. This allows the merchant to determine which factors made contributed to labeling the customer as fraudulent, or to know the merchant’s trusted customers so that the merchant can take a more individual approach with them.
Rule management system
Covery allows merchants to create flexible if-then rules. It allows the system to reproduce any user behavior patterns and even incorporate the merchant’s in-house data.
The system uses flexible scoring models, created using ‘if-then’ rules. The models are also easily customizable for any business model and reﬂects human intelligence.
Machine learning risk score
In addition to the existing tools, Covery also uses the power of artiﬁcial intelligence. The ensemble methods use multiple machine learning algorithms to dive even deeper into data arrays and prevent fraud.
The system applies Trust list and Rule-based scenarios by default. The merchant can use any combination of the above mentioned steps to obtain the most accurate result.
Step 3 – Decision
Decision-making agent analyze the data obtained from all the system fundamentals, including risk policies, and makes an automated decision whether an event should be accepted, rejected or whether it requires individual attention.
“The total loss from online fraud today is over $100 billion. Since this is a global industry trend, Maxpay focuses on customer safety to reduce the amount lost to fraud and scams, while avoiding chargebacks. We have developed a fraud prevention solution, which uses more than 1000 attributes to identify fraudulent behavior,” says Artem Timoshenko, CEO Maxpay