Why sorry? Because although the data indicated significant growth in online retail sales AND corresponding significant growth in fraud attempts, my message about how retailers can prevent fraud just did not seem to reach as many as I’d hoped.
What we saw was an impressive 45% transaction growth during the period 1 November 2017 – 31 January 2018, compared to the same period the previous year. The Sunday prior to Cyber Monday (78%), Black Friday (77%) and Cyber Monday itself (74%) experienced the largest transaction volume uplift. Surprisingly, Click Frenzy (an Australian initiative) experienced a miniscule increase of less than 2% over the previous year.
Although some key dates did see a fall in fraud rates, what we witnessed overall was a huge increase in fraud over the holiday season, with a whopping increase of 71% YOY. Peak day for fraud was Christmas Day itself, supporting what I like to refer to as the ‘Melbourne Cup effect.’
In Australia, the Melbourne Cup is held annually and is regarded as the nation’s premier thoroughbred horse race. It is known to be the richest ‘2-mile’ handicap in the world, with a purse of more than AUD $6.2M in 2017. The race is run at 3pm on the first Tuesday in November and is well known locally as ‘the race that stops the nation.’ In the State of Victoria, where the Cup takes place, it is even a public holiday – and most Australians stop to watch it. Many once-a-year punters place their annual bet on this race.
The ‘Melbourne Cup effect’ is a theory that if you are going to commit a crime, doing it during the Melbourne Cup race would be a good time, as (almost) everyone across the country is distracted by it. The same could be said for Christmas Day too. Whilst many are distracted with family get-togethers, others are shopping online with their freshly unwrapped Christmas gift cards, so there is a lot of ‘noise’ with genuine sales and, conversely, a lot of opportunity for fraud to be perpetrated.
So here we are again, we have experienced another year of aggressive fraud growth during the 2017 holiday shopping season. As we work through disputed transactions and chargebacks, and take time to review the fraud losses, it is also time to start preparations for this year. With this knowledge, retailers can arm themselves in advance and tackle the problem head on. After all, knowledge is power!
The key is closely examining the fraud transaction data; both financial and non-financial. This data can be utilised to build profiles of the fraudulent and suspect activity. Data that is essential to profile includes card issuing country, shipping country, billing country as well as device intelligence – IP address, geo-location, Apple vs Android, operating system, browser language, ese of TOR exit nodes, and device time on file.
Adaptive machine learning profiles that can action activity in real time are best practice in this space, because they evolve with the seasons, tracking genuine customer spending as well as changes to fraudsters’ modus operandi. Profiling in this way applies a risk assessment of activity to highlight high-risk transactions, and conversely, quickly identify genuine customers. This real-time profile is used to direct actionable intelligence strategies, such as denying transactions (declines) and applying authentication to only those transactions with a high risk of fraud, leaving the genuine customers to get on with their purchases.
Using this intelligence to identify and action high-risk transactions in real time is a sure bet for reducing fraud losses and importantly, reducing friction for genuine customers.