Retailers around the world are currently experiencing their busiest time of the year in the run-up to the festive season, with post-Christmas sales also right around the corner. However, it is not just consumers hunting for bargains or engaging in last-minute panic buying that will be turning to online stores this season.
It is also a prime time of year for fraudsters, who will hope their transaction will go unnoticed in among the Christmas rush. If they are not careful, retailers can find themselves seriously out of pocket if they fall victim to such crimes – so many companies are turning to big data analytics to help clamp down on this problem.
One such organisation is eBay Enterprise, which provides omnichannel fulfillment services for hundreds of brand-name merchants – and as such, has to absorb all fraud-related losses incurred by its clients. Therefore, it is vital that the company is able to spot fraudulent transactions and block them before they are finalised.
It was noted by Datanami that in 2014, the company stopped $55 million worth of these purchases, and that number is expected to rise this year. The key to this success is how the firm uses big data to analyse buying activity and identify patterns.
Tony Ippolito, strategic risk and technology manager for eBay Enterprise, told the publication that the more data the company has available to it, the better.
Identifying a fraudulent transaction typically involves cross checking information provided by the customer against a wide range of fields – including names, email addresses, billing addresses and shipping addresses – to find inconsistencies. In eBay Enterprise's case, this involves running every order through an Oracle database containing around 1.3 billion entries.
"We also collect as much information as we can about the product, the kind of item, and the amount of the order," Mr Ippolito said. "We do device fingerprinting, we collect IP address, and then we do geolocation lookups."
Common red flags include long-distance orders for high value goods such as electronic gadgets, video games and designer clothing, which can be easily resold on the black market. Expedited shipping requests are another key signifier of fraudsters, as they are keen to get items in their hands before their crimes are spotted.
To counter this, eBay Enterprise runs each transaction through a big data analytics system that is equipped with around 600 rules, as well as a machine learning algorithm that uses more than 20 models to match incoming transactions against known fraud patterns.
Mr Ippolito said: "It's a lot of data collection and aggregation, seeing trends and applying that across the board to make sure we're not missing anything."
This is only possible with an effective big data system that is able to process millions of incoming transactions and accurately assess their likelihood of being fraudulent in real-time, as any delays or mistakes will either lead to genuine customers being inconvenienced, or enabling a fraudulent transaction to be completed.
Such systems also have to be regularly tweaked and updated to keep up as fraudsters change their tactics in response to these tools. For instance, if the company implements a rule that requires it to 'queue', or manually review, every transaction over $50, then the fraudsters will move their target to orders under $40.
"When you close off a certain area, you have to be aware of what the next logical step for them is," Mr Ippolito says. "If you shut down overnight shipping, then they'll move into third-day shipment. It's a lot more nuanced than that, but that's the general idea."