Online merchants need to be ready for the unexpected costs of false declines

Global eCommerce is surging with more businesses relying heavily on digital channels to reach their customers. In Asia Pacific, businesses recorded 37.6% growth in online orders in 2020, and this is expected to nearly double to reach $2 trillion by 2025.

Mobile-first Southeast Asia has seen traffic for online shopping increase by leaps and bounds in the past year, with Singapore leading the trend with a surge of 35% compared to 2019, followed by the Philippines (21%), Vietnam (19%), Malaysia (17%), Thailand (15%) and Indonesia (6%).

Yet, with opportunity come risks. Against this backdrop, it is found that 27% of online sales end up being fraudulent, and at least 86% of global consumers have fallen victim to credit/debit card fraud, identity theft, or a data breach in the last year.

More importantly, while losses due to eCommerce fraud are projected to reach $6.4 billion this year, losses due to false declines are estimated to reach $443 billion - nearly 70 times more than losses from fraud itself.

Despite this growing fraud risk, the latest Global Identity and Fraud Report revealed that there was a 16% decline in business intention to increase fraud management budgets between June 2020 to January 2021.

Businesses are missing out on growing revenue when they do not manage false declines.

Why false declines are costly

A false decline happens when a legitimate transaction is flagged as fraudulent and rejected. This can happen under several situations - when a user’s IP address is in a different country than their billing address; there can also be varying filters such as daily velocity filters which limit the number of transactions one IP address can make each day, shipping and billing address mismatch filters, as well as high ticket purchase filters that are triggered by transactions exceeding a certain value threshold.

Many businesses hold to the unrealistic goal of achieving zero fraud. There is one easy way to accomplish that: decline all transactions despite insufficient data to justify otherwise. But if merchants intend to build a profitable business, fraud prevention cannot come at the cost of suppressing revenue and alienating potential customers.

In particular, it was found that new customers are five to seven times more likely to be declined because merchants struggle to differentiate these legitimate sales from fraud attempts.

With businesses pivoting in a short time to eCommerce during the pandemic, the technology gaps and data silos may result in insufficient information to decide on the legitimacy of transactions. This leads to immediate lost revenue in the long term as declined users are going to be frustrated, and that frustration can seriously damage a brand. Customers may abandon their cart, which is a lost sale.

They can abandon your company all together and turn to online forums or review sites to complain, and who can blame them? Other potential customers may read these reviews and decide to buy from another business. This leads to many lost sales targets, all of which cuts down your revenue.

While the market for fraud-prevention solutions and technologies offers some options, the same cannot be said for false declines detection. False declines do not show in your accounts like chargebacks do.

Without a conscious effort to track them, the problem can grow into a full-blown profit leak.

Steps to mitigate fraud risks

So what should businesses do?

First, they should consider fraud recognition rather than fraud detection. Fraud recognition involves building deep learning capabilities that allow a system to learn and analyse, the way a human would, to determine whether a transaction is suspicious.

If we block all transactions, then we are not only driving away legitimate customers, we are giving anti-fraud tools less data to work with to increase its accuracy. Real fraudulent transactions can then slip through the net.

Rather than overly tightening acceptance parameters or choosing a set-and-forget fraud fighting tool, effective anti-fraud systems must be given the flexibility to learn and adapt continuously to new fraud vectors - this is the key to fraud protection and revenue generation.

Fraud prevention cannot come at the cost of suppressing revenue

- Shabab Muhaddes, General Manager of Vesta, Asia Pacific

Second, they can consider solutions that couple artificial intelligence and machine learning with manual reviews. Threats evolve and solutions need to learn from data such as tagged approvals and declines and identify patterns to increase the approval rate of transactions.

At the same time, expert data scientists can also be on hand to train and refine the fraud system to stop threats and optimise payment acceptance. This may sound counterintuitive in an era of deep machine learning and automation. But these automated tools need to be fed data to be effective and nothing replaces human interaction to understand the complexity of customer behaviour.

Lastly, if a particular order seems suspicious, get in touch with customers directly before immediately flagging it. Not only will this give the customers a chance to prove their legitimacy, they will appreciate you taking the time to check on them, which helps in establishing a long-lasting relationship.

At the end of the day, there will always be more legitimate online transactions than fraudulent ones. Tackling fraud head-on is necessary to address the bigger need of returning prosperity to the online marketplace.

It’s getting more and more competitive out there and the customer experience is an important battleground among merchants to win and retain customers. False declines have become a part of that overall shopping experience.

Either merchants address this problem or suffer the consequences of losing their competitive edge, brand reputation, their customers and revenue.

Shabab Muhaddes is General Manager of Vesta, Asia Pacific.

Research sources