The Australian retail eCommerce market is estimated to be growing at 20.2% YoY. But could it grow even faster?

Fear of fraud is one of the biggest factors stunting further growth for Australian merchants. Fraud teams over decline good orders by leaning on outdated rules to detect fraud. In addition, they often resort to protective measures like 3D Secure, two-factor authentication and other sophisticated security protocols to prevent fraud and abuse. Unfortunately, these measures introduce friction into the shopping journey, ultimately resulting in dissatisfied customers who take their business elsewhere. 

For digital and online teams, the trick is to find the optimal point between approval/acceptance rates and chargeback rates. Traditionally, Australian merchants have taken a relatively risk-averse approach, choosing (sometimes unknowingly) to keep chargeback rates down, even if this means preventing some good customers from checking out. 

The argument we often hear played out between fraud prevention and online sales teams across the country (and globally) goes something like this:

Online Sales Team: Our KPIs are to increase online sales and provide a great customer experience to ensure repeat visits – it is all about customer lifetime value. We spend thousands on marketing and customer acquisition and then when the moment of truth comes, and the customer is finally going to make a purchase, you, the fraud review team, are not allowing them to checkout because you think it looks too risky?!”

Fraud Prevention Team:That’s our job. Our target is to reduce fraud as much as possible. So even if the order mostly looks ok, we will decline it if there are any risky elements. We can’t afford to accidentally approve a fraudulent $1,000 order and end up with a chargeback for $1000 plus the associated fees – it just isn’t worth it. Sorry. We are just doing our job.”

Let’s delve deeper into this imaginary, but very familiar debate. In Australia, many companies work with traditional fraud solution vendors who will design rules to detect fraud. For example, a rule might be designed to say that if the billing address of the order does not match the shipping address of the item, the order should be declined. The problem with these rules-based decision tools is the temptation to continue adding more and more rules in the hope of preventing fraud. However, these rules end up hurting the purchase funnel because they overdecline good orders. This makes the checkout process difficult for the customer and creates a negative shopping experience. Despite this challenge, the fraud prevention team has the final say in approving transactions, so it becomes a greater problem. Unfortunately, the risk/reward tradeoff here swings towards preventing risk instead of reaping the reward.

In one recent conversation with a large retailer, the fraud team proudly told us about having put in place over 600 rules to prevent fraud. These rules are reactive, require constant tweaking and often act as a barrier to a customer completing their purchase. Despite the increased investments that Australian businesses are making in marketing, customer acquisition, and online user experience, many companies end up with a much smaller pool of customers who are eligible to make a purchase.

Moreover, many of the traditional solutions still require employees to manually review orders. We recently learned that manual order reviews were the underlying issue for a company that had been the subject of intense media scrutiny for their slow delivery times across Australia. This meant that over weekends and peak times, order volumes overwhelmed the manual review team and orders sat for hours in a queue, thus delaying their fulfilment, shipping and delivery to customers.

These traditional rules-based setups can have a material negative effect on a merchant’s bottom line. The revenue left on the table (for larger Australian businesses we estimate figures in the tens of millions of dollars), the frustrated customer who doesn’t return to purchase and the heavy costs of managing fraud are hurting too many Australian companies.

Thankfully, Australian businesses are starting to shift away from the rules-based risk-averse fraud tools and adopting machine learning solutions designed to optimise performance in their place. In 2012, the global fraud review sector was disrupted with the entry of machine learning solutions to the market that more accurately and proactively detect fraud. Whether it be industry retail leaders in Australia, fashion labels or even video game merchants, the switch to machine learning solutions that can approve more good customers is on. The risk/reward tradeoff is starting to swing back towards reward.