PetSmart is Using AI and Machine Learning to Conquer Digital Fraud
brought to you by WBR Insights
A new report from ThreatMetrix analyzing 17.6 billion digital transactions in 2018 has revealed a sharp increase in ecommerce fraud across all touchpoints. It's estimated that 25% of new ecommerce accounts are fraudulent, which represents a 130% increase from the same period in 2017. Compounding matters is the increasingly global nature of ecommerce, with 54% of digital transactions now being cross-border. International transactions are 69% more likely to be fraudulent.
This is leading online retailers to come up with new ways to combat these threats. Fraudulent transactions cost businesses millions of dollars every year and can result in companies shutting their doors forever.
PetSmart is aiming to combat this kind of fraud by using AI and machine learning technology to identify fraudulent orders before they are shipped, a move it believes could save $12 million per year.
The technology is developed by AI specialists Kount, and is designed to not only prevent losses at the business level but also gather evidence which can help with the prosecution of those responsible.
(Video source: youtube.com)
Kount's impressive technology works on two levels. First, the unsupervised machine learning process analyzes bad actors from all over the world and connects them together to build models of fraudulent behavior as they emerge, which can then be employed by businesses to guide policy creation.
On the other side, supervised machine learning feeds transaction data through hundreds of models to assess the relative risk of transactions, logins, account creations, and more, offering several layers of protection which could otherwise be missed. The whole process takes about 250 milliseconds, so it doesn't inconvenience genuine customers at all.
"Fraud is an ever-evolving problem, so businesses must ensure their prevention technology evolves right alongside to maximize protection," said Chief Customer Experience Officer at Kount, Rich Stuppy. "In order to keep ahead of sophisticated fraud techniques, such as account takeover and transaction fraud, while simultaneously helping online merchants maximize order acceptance, our solution is designed to be dynamic, providing both protection and opportunities to increase sales and revenue. These efforts continue our long history of using AI and machine learning."
An Increasingly Crowded Market
PetSmart is already starting to see great results with Kount's anti-fraud technology, but they are far from the only fish in the pond. As the technology employed by cybercriminals becomes ever more sophisticated, many organizations are springing up with their own solutions to the problem.
Companies such as Accertify, CyberSource, and Easy Solutions all have their own proprietary platforms which can be employed by ecommerce retailers to combat fraud, using their own combinations of rules, modeling software, location-based algorithms, and more, to investigate and stop fraudulent transactions.
This works out well for retailers, as competition in the marketplace will lead these companies to develop ever more sophisticated platforms to stay ahead of the curve. Customers put faith in the companies they deal with to treat their personal information with the respect and care it deserves, and, although fraud is a global issue, they will hold brands personally responsible for any breaches.
United Kingdom-based credit rating agency Equifax was recently fined PS500,000 (about $663,000) for failing to protect the personal data of 15 million Britons, the highest fine possible under UK law.
While a large company such as Equifax or PetSmart may be able to absorb such a fine without too much worry, it could be devastating for a smaller business. Failing to take fraud prevention seriously can be extremely damaging not just for brand reputation, customer retention, and future sales, but also for the immediate survival of the business.
"By tracking online fraudsters, PetSmart has been able to cancel nearly $4 million in fraudulent orders," said Manager of Ecommerce Investigations of PetSmart, Chad Evans. "After labor loss, shipping costs, merchandise costs, chargebacks, fees, and fines are added, that could translate to $3 to $3.50 for every dollar in fraudulent charges, for an annual total of about $12 million. For 2018, the figure is already at $1.5 million for the fiscal year, which began on February 1, so there might be more fraudulent orders caught this year than last year."
Download the agenda today for more information and insights.