Will Analytics and Technology Put an End to Credit Card Fraud?

If you havent noticed the change, youre living in a cave. Now, businesses are charged with helping banks and credit card processors fight credit card fraud. The stakes are high and the battle against thieves is on a field that includes big data, machine learning, and hardware.

Around half of the worlds credit card fraud happens in the United States. Because of that stunning statistic, in October of 2015 the Federal Reserve Bank ordered US merchants to adopt EMV (Europay Mastercard Visa) readers by the end of 2016. If merchants dont get an EMV chip card reader (and many places I frequent still have not), they face the liability shift.

The banks are saying, “chip cards are safer than magnetic strip cards because theyre harder to counterfeit. Well issue the cards and its not our fault if you dont get the technology necessary to run them.” The shift means merchants will be liable for fraudulent transactions if the customer has a chip card but the merchant doesnt have the reader.

For whatever reason, the US has been behind Europe and the rest of the world on this. By 2013, nearly 97% of transactions were EMV in Europe alone. Since half of the worlds credit card fraud is on Americas shoulders, we can safely assume the lag on EMV adoption is at least partially to blame.

EMV may cut down on counterfeit cards, but theres still the matter of stolen cards and stolen data. Cybercriminals are worldwide, and the amount of data generated by American credit card transactions is a low-hanging fruit.

• According to research from Ohio University, the US racked up a $7 billion dollar credit card fraud bill in 2013 alone.
• There was a 59% increase in the number of data breaches from 2013 to 2014.
• And according to an infographic on cybercriminals from the University of Cincinnati, financial access accounts for the most data breach cases, at 16%.

Businesses are using analytics in attempts to stymie theft. Credit card processors such as Visa are using big data and machine learning to analyze the huge stream of transactional data in real time.

Algorithms determine the probability of fraud by analyzing your purchasing habits and comparing each transaction with what preceded it. Each new algorithm builds on the last. If a transaction looks strange, such as several cash advances in one day when youve never done that sort of thing before, those advances may be denied.

But what if youre traveling in Timbuktu, your car breaks down, and you need emergency funds? Your card gets denied. Its one of the most frustrating feelings. You get a call from a fraud detection unit.  

Interestingly, digital wallet technology such as Apple Pay lets you store your credit card information on your phone and pay via Radio Frequency Near Field Communication (NFC). Combine this with the fact that your smartphone can access location data via GPS (if you allow it to).

The potential for payment and fraud detection is to revolve around one device. If someone gets a hold of your credit card information and attempts to make a payment, the credit card company would automatically know because your phone is transmitting a signal from somewhere else.    

Because of big data in the cloud, machine learning algorithms, and new payment technologies such as digital wallet, the future of credit cards and fraud detection may be entirely location and device-based. But what about online transactions?

Ecommerce is huge and presents its own set of data security challenges. Because of this, analysts predict cyber security expenditures will hit $170 billion by 2020, while the cost of cybercrime will hit $2 trillion.     

This speaks to the power of hackers in an age where its incredibly hard to protect data. Although firms will spend billions on security, the cost of crimes will be in the trillions.

Algorithms can recognize if an online credit card payment is coming from a different IP address than you normally use. But a sophisticated hacker can find ways around this. Companies with gigantic troves of credit card data will have to invest heavily in security protocols for the cloud, such as multifactor authentication and encryption.

Big data is a double-edged sword. On one hand, algorithms analyze it to determine the probability of fraudthe more data there is on your habits, the more accurate the analysis. On the other, reams of credit card data are a treasure trove for cybercriminals.

The solution to credit card fraud is to make sure every access point is secure. Its possible, but with so many organizations in possession of credit card data, and so many varying levels of security, its hard to foresee an end to fraud anytime soon.