Preemptive fraud management practices for merchants

Reference Link: Press Here

Anka Yip, Payment Asia, talks about fraud management in ecommerce and explains how the company supports merchants in managing risk while expanding into China

Ecommerce is no more merely the privilege enjoyed by big merchants and enterprises that have massive infrastructures and resources. Small businesses carrying a smartphone or tablet are already enough to join the endless network of ecommerce. The growth of ecommerce, however, arouses the concern of fraudulent transactions, implying that effective and qualified fraud management is also a must in minimizing the lost suffered by merchants or businesses. Considering the special situation in the Chinese market, Payment Asia (PA) shifts the ‘fraudgate’ from standing at the rear to staying at the forefront by preemptively screening out suspected transactions.

The management of fraud

Fraud is commonly seen in the business world, and it is unavoidable in the Internet as well. From phishing, identity theft, and pagejacking to merchant identity fraud, perpetrators would not waste any chances to take advantage of the stolen money, personal property and sensitive information. The exponential growth of ecommerce could transfer all the black and white transaction records into computer servers, but it also provides room for hackers and cybercriminals to deprive the information via the Internet. What businesses could do is take precautions and keep monitoring online fraudulent activities. Thus, fraud management comes to exist and becomes the core of ecommerce solutions.

The 2017 report of LexisNexis True Cost of Fraud evaluated the effect of fraudulence activities across private businesses and government agencies in the US, and especially in ecommerce, financial services, and digital lending.

Merchants and enterprises had to spend USD 2.66 and around 2% of revenues on average for every dollar of fraud, and the fraud cost rose to USD 3.48 if the transactions are about digital goods sales or conducted primarily through remote channels, accounting to about 2.5% of revenues on average. The report simply shows fraud is a lucrative business.

The two generations of fraud monitoring 

PA is a veteran in dealing with fraudulent transactions in the internet world, where the legal framework of commercial and financial criminal procedures is still very challenging. When facing online fraud, merchants may often choose to surrender, as they do not find it worthwhile to spend unscalable time and legal cost to report and claim the loss. PA is then determined to help the vulnerable merchants by identifying potential perpetrators in preliminary stages to lower the risk of fraud.

In the beginning, PA implemented a fraud scoring system and offered some basic rulesets based on the database provided by merchants. Rulesets changed from time to time. Our data analysts successfully and continuously identified the characteristics of known fraud cases and updated our rulesets. We might then block the high-risk fraud transactions based on the scores recorded by our system, and later on follow up with merchants to find out more about their involved clients. This is the first generation of PA’s fraud management practices for merchants.

However, individual identities or information might be easily obtained by some clusters or groups. The traditional monitoring is not enough anymore to tackle potential fraud. Therefore, PA is developing the second-generation technology to expand the coverage of monitoring. We run all transactions in our models and connect the fraud cases with cluster analysis to improve the scoring mechanism.

The role of AI in detecting fraud

To process a huge amount of transaction data and to pinpoint potential fraud, PA adopts artificial intelligence to conduct big data analysis of past and current transaction records, as well as to summarize the possible fraud patterns and highlight them as parameters. The more parameters are put in the AI-driven machine learning, the higher the threshold of fraudulence is against cybercriminals. At the same time, business market changes every second, and so does the way the corresponding cybercriminal acts. PA would adjust the weight and choices of parameters based on the trend of fraudulent activities to ensure our machine learning principles and practices are sticking to the dynamics of ecommerce market and its fraudsters.

Until now, PA has screened hundreds of thousands of transactions. The frequency of suspected fraudulent activities is effectively lowered from several cases per day to one case per several days. PA’s fraud screening helps merchants to conduct KYC compliance that would reduce the risk of losing money and credentials, as well as the threat of leaking sensitive information.

From quality to loyalty 

Quality brings you loyalty. It is always the key to success in the business world by offering qualified goods and services to customers, and it is how business confidence is built. As fraudulent transactions harm merchants’ credentials and revenue, eventually disrupting the market order, PA provides comprehensive ecommerce solutions and goes hand in hand with merchants to keep scam activities under control and maintain the market for the benefit of all lawful parties. Given the potential of the Chinese ecommerce market, PA will never stop adopting the latest technology and combine it with our built-in market sense to secure and drive up merchants’ businesses in China.