How AI and machine learning boost conversions—and revenues
eCommerce continues to grow at impressive rates around the world. Worldpay’s 2018 Global Payments Report projects that the global eCommerce market will surpass $3.6 trillion in 2019, growing to a remarkable $4.6 trillion as soon as 2022.That’s great news for eCommerce merchants everywhere.
That enthusiasm is tempered by inefficiencies around conversion. Shopping cart abandonment, fraud, and false declines present complex challenges that erode profits.
These complex conversion challenges exist on a continuum and can be most effectively managed in an integrated way. The transformative power of machine learning and AI are helping propel payments data to the forefront of the conversion discussion.
ONE—Global complexity demands next-generation conversion solutions
The global payments ecosystem is an incredibly complex heterogeneous mix of merchants, merchant acquirers, card networks, issuing banks and issuing processors. That universal breadth creates inefficiencies that are felt in higher fraud rates, lower authorization approval rates and checkout friction that drives cart abandonment.
The system is driven by data, data that tells us a story about the value as well as the legitimacy of every transaction. A static approach to submitting payments transactions is increasingly insufficient in the context of global eCommerce. The complexity and diversity of the payments ecosystem demand dynamic intelligence and tools that adapt quickly to new information and changing conditions in essentially real time. And since eCommerce is borderless, the tools need to leverage transaction data at global scale. With this much change and scale, the ability to identify patterns and adjust demands machine learning and AI.
TWO—Data is critical to creating less friction and converting more sales
Merchants are constantly reassessing how they can improve the consumer experience to reduce card abandonment and increase conversion rates. Reducing friction requires continuous adaptation to the rapidly shifting eCommerce landscape, to ensure that customers complete their transactions.
Global eCommerce merchants are constantly innovating to attract new customers and grow their markets beyond their core demographics. When penetrating new geographies there are many assumptions made about buyer preferences, and in particular, their preferred payment method.
In eCommerce that alternative product or service is just a click or tap away. Reducing friction is essential. Sophisticated merchants conduct A/B testing on every word that’s on the page, every color, every placement of every item in the shopping experience. Ultimately, the data can reveal customer preferences in each market and result in higher conversion rates.
THREE—AI fuels efficient account updating and recycling
With intelligent account updating and authorization retry/recycling, AI and machine learning are helping merchants reduce friction, boost authorization rates and optimize revenue.
Recurring payments can be the optimal business model for many businesses, whether selling services or physical goods. Not only do recurring payments provide revenue predictability, they represent the best return on the investment in customer acquisition as they maximize customer lifetime value.
The challenge with recurring payments is that “leakage” can occur. Recurring payment transactions are unattended and fail for any number of reasons: account credentials expire or change, a balance may have insufficient funds, a network connection may not be available, etc. These things happen, but the impact is huge if you lose not just that payment—but the customer relationship and all subsequent expected payments.
Account updater services lift recurring revenues just by keeping card information current. Most merchants obtain updates in advance of an expected payment transaction. However, if the transaction is declined you shouldn’t give up. Intelligent account updater systems respond to the decline by checking for more current information, and then applying that update with a retry. Interestingly, the retry pattern – number of days after the decline, number of times to retry, time of day to retry, etc. -- matters a great deal. Different retry patterns yield different approval rates based on the combination of initial decline reason, issuing bank, card product, etc. This is where data science and machine learning come in as the optimal approach is to develop algorithms for managing account updating and the recycling of initially declined transactions.
The biggest no-brainer in eCommerce payments
Account updating services are widely seen as a sure-fire way to increase recurring revenue, and now they are finding utility beyond the recurring business model. For merchants who stored card information on file for their consumers who very frequent but nonrecurring purchases, card changes can result in payment friction and potential cart abandonment. The introduction of real-time account updating services put the power of dynamic learning to work in the service of reducing checkout friction for this growing segment.
Take Gaming, which features many in-game purchases. Multiple purchases can trigger overly-aggressive fraud filters to cause declines. That adds unnecessary and destructive friction at the most important customer touchpoint.
Payments intelligence improves conversion outcomes for ride-hailing services, food delivery and any business where customers make frequent transactions. The lift in approval rates account updating services offer makes it the biggest no-brainer in eCommerce payments.
AI and machine learning are being actively applied to the inefficiencies that matter to merchants most. Combined with real-time data sets of sufficient scale, AI and machine learning are helping enterprise and eCommerce merchants achieve greater reach, drive incremental revenue and become more responsive to their customers.