Artificial intelligence at PayPal – two unique use cases

0

The company that would later become PayPal Holdings first entered the electronic payments space in 1999, a year after its founding as Confinity. Confinity merged with Elon Musk’s x.com in 2000 and was rebranded to PayPal the following year. The company went public in 2002, just prior to its acquisition by eBay, making it became “The site’s official payment processor.” eBay spun off PayPal as a separate company in 2015.

PayPal today processes more than 35,000 transactions per minute. Last year, PayPal processed approximately $1.2 trillion in total payment volume (TPV). On your website, the company claims to have grown to 416 million active consumer and merchant accounts and employs 27,700 people. PayPal has been trading on the Nasdaq since January 2022 with a market capitalization of APPROX $210 billion. For the year ended December 31, 2020, PayPal reported Total revenue of $21.5 billion.

In this article, we look at how PayPal has explored AI applications for its business and industry through two unique use cases:

  • Payment Authorization Rate — PayPal uses machine learning models to improve authorization rates of valid transactions on its platform by predicting and addressing issuer rejections that interrupt users’ payment requests.
  • fraud prevention — PayPal uses machine learning and graphics technologies to create connections and evaluate relationships within its data that help detect and stop fraud.

We’ll first examine how PayPal has turned to machine learning technology to process valid transactions more efficiently by anticipating and resolving card declines that slow down online purchases.

Improve payment authorization rate

Fake credit and debit card declines cost real money. ONE Study 2020 Research conducted by checkout.com in partnership with Oxford Economics found that false card rejections cost merchants around $20.3 billion in 2019 in the UK, US, France and Germany. More than sixty percent of that amount – $12.7 billion – went to competing sites. The rest, $7.6 billion, was simply lost.

In e-commerce, card authorizations unlock revenue for online retailers. Even small strides in preventing false rejections and ensuring efficient approvals can translate into significant revenue increases and cost savings.

As a processor of more than $1 trillion in payments over the past year, PayPal was looking for a solution that could improve the efficiency and effectiveness of its payment authorization process.

Source: PayPal Technology Blog

The solution developed by PayPal, the payment processor allegations, offers trader approval ratings above the industry average by leveraging a combination of the company’s capabilities and expertise in AI and machine learning, extensive datasets, network tokenization, issuer and network partnerships, and various funding tools.

In the machine learning component of its solution, PayPal has developed algorithms that:

  • predict issuer rejection, through Create Models that take into account, for example, recent issuer rejection rates and patterns, as well as approval and rejection history on different days and times.
  • Fix declined transactionsby asking users to try different payment methods, provide another form of authentication like CVV, or fund their account.
  • Design effective repetition strategies, which help determine the best way to retry the payment, taking into account factors such as the card used, the parameters of the transaction, and the optimal time to retry.
  • Detect fraud in real timeby identifying warning patterns, such as carding attacks and draw on insights from PayPal’s vast, two-page data sets.

When deciding which model to apply to the data, PayPal claims to have taken into account both tree-based algorithms such as Random Forest and more complex models like neural networks before deciding to go into production with tree-based production Gradient increase machine (GBM) model that delivered the best results in the test.

From 2017 to 2020, PayPal claims to have improved global authorization rates for its brand processing by over 300 basis points. says his SVP of Omni Payments, Jim Magats, in a company newsroom article. He goes on to say that new users increased their authorization rates by 600bp after signing up with PayPal.

Research of PayPal also points out that its payment authorization results have improved by 30bps based on its intelligent retry strategies alone, and its rejection prediction algorithms and resulting remediation actions have improved authorization rates by as much as 240bps at some merchants.

fraud prevention

Payment fraud is on the rise. In 2020 almost one in four fraud checkers reported see a significant increase in the risk of payment fraud. And 47 percent of auditors predicted that the risk of payment fraud would increase significantly over the next year.

The growth in e-commerce sales is one factor driving the rise in payment fraud. According to eMarketer According to research, US retail e-commerce sales will grow from $598 billion in 2019 to more than $1 trillion by the end of 2022.

As a global provider of payment services, PayPal is looking for new solutions to meet its fraud prevention needs. While these solutions must protect against the financial and reputational damage that payment fraud can bring, they must also avoid the additional costs and lost business that can result from unnecessarily slowing down the transaction process through unreasonable rejections, human intervention, and unnecessary customer checks and controls.

To combat payment fraud, PayPal has developed PayPal Fraud Protection, an “adaptive machine learning solution this helps merchants protect themselves from increasing fraud.” PayPal allegations that the solution’s strength in fraud detection rests on the company’s vast data stores generated by its 2-side network, the name PayPal uses to indicate the two sides of each transaction. (PayPal processes payments and can act as a payment method through its PayPal wallet service.)

With the data generated of its 15 billion payment transactions generated by millions of accounts in 2020 alone, PayPal allegations that adding his digital wallet data “helps [PayPal] to detect anomalies or suspicious patterns… [that are] incredibly effective in fighting fraud.”

On his technology blog PayPal admits that the number of relationships in his data is “too large to iterate through and analyze when they should be stored in a relational database”. As a result, the company turned to graphical analysis, the allows PayPal to “create a payment chart that depicts transactions between buyers and sellers.”

PayPal claims that its real-time graph database allows the company to connect different relationships in (near) real time, which supports its fraud detection activities.

“Graph technologies have proven to be very effective in fraud detection and prevention,” concludes PayPal in a December 2021 post on his technology blog. In fact, PayPal’s fraud protection service goals to:

  • Reduce losses for merchants and buyers
  • Reduce false alarms and unnecessary customer inconvenience
  • Minimize fraud prevention costs

As a percentage of the total payment value, or TPV, PayPal transaction losses (including fraud losses, chargebacks, and protection program costs) have lost weight from 0.18% in 2018 to 0.12% in 2020. According to LexisNexis data reported in one intel Case Study PayPal’s Fraud Rate Is “Significantly Below the Industry Average of 1.86% [but] is still over $1 billion a year in losses for the company.”

Share.

Comments are closed.