It’s been nearly five months since the prime minister announced a state of lockdown in the UK due to the coronavirus. In this time, the retail sector as we know it has changed dramatically.
Without much warning, retailers have had to close brick and mortar stores from the start of the lockdown in late March to opening them back up again following the easing of restrictions in June. In these two and a bit months, consumers in their droves have sought out e-commerce sites in order to purchase items that they previously would have brought in store. The Office of National Statistics has found that online shopping as a percentage of total retail sales hit a record peak in May following lockdown.
The switch from physical stores to online purchases has continued since shops were allowed to reopen in June. According to research, fewer than two in ten (16%) UK consumers are intending to return to their old shopping habits post-COVID, with online shopping being predicted to account for over half (51%) of all spend moving forward.
The rapid move to online commerce during COVID-19, while great for revenues, has unintended consequences for retailers and consumers. With more people shopping online, money is coming from different places and therefore creating the perfect conditions for fraud. In other words, traditional fraud detection that relies on previous transactional data will become more prone to false positives and false negatives as old patterns in customer behaviour no longer apply.
Online retailers prevent fraud by analysing a number of different data points. Current methods of fraud detection rely on measuring past behaviours against internal risk thresholds or against a baseline of individualised normal behaviours determined by machine learning. While both methods do result in a minority of false positives and false negatives, the system slowly becomes more accurate at detecting fraud over time.
With e-commerce purchases on the rise, the fear is that retailers are unable either adapt their risk thresholds or baseline behaviours for machine learning models. As a result, retailers could let fraudsters steal information and exploit customers or over-correct and prevent customers from making genuine purchases. For example, new customers purchasing large items may be flagged as fraudulent, while fraudsters might go undetected due to a surge in unusual activity that obscures their purchase.
The third way
The answer in combating fraud relies on creating another way. Rather than deploying overcorrecting draconian measures or letting fraud run rampant, retailers must measure transactions against new types of data such as intent.
Intent data such as location, search history and spending helps determine whenever a purchase is fraudulent or genuine as it provides context behind each transaction. For example, if someone is spending time searching for a hotel in Cornwall, then an e-commerce site can identify and approve related purchases such as surfing boards and wetsuits. It can also flag a purchase in a specific location that doesn’t match with their intended destination.
Retailers should also partner with data consortiums in order to obtain intent data that can help prove or disprove assertions about a customer. What’s more, data partners allow online retailers to safely and securely combine insights to create a more rounded and accurate picture of customer behaviours that can help create more confidence and reduce fraud.
E-commerce sites should welcome, and also be vigilant of, the surge in online transactions during the pandemic. While the increased number of online sales will increase the bottom line of many retailers, it will also coincide with fundamental challenges to their fraud prevention capabilities. If retailers can’t manage to prevent fraud on their platform then they will find that customers will choose to spend their money elsewhere on sites that are able to prevent fraud and reduce friction in the purchase process for customers.
Nguyen Nguyen, Vice President, partner development & technical services at ADARA