Using Data to Fight Mobile Fraud in 2018

January 4, 2018         By: Kevin Lee

Do you remember a time without smartphones? When up-to-the-moment traffic, breaking news, or an update from a friend wasn’t just a tap away?  If you’re like 98% of U.S. Millennials, you own a mobile device, and it’s becoming more and more indispensable to how you find information, communicate, and shop.

Consumers aren’t the only ones who’ve recognized the importance of mobile. Fraudsters have too – and it’s costing merchants dearly. Research from LexisNexis shows that every dollar lost to fraud costs m-commerce merchants $2.83.

How do you protect yourself from mobile fraud? Start by taking a look at the data you may already be collecting from your users. Here are five types of mobile-specific data you can leverage to fight fraud:

 

 

  • Number of apps on a phone

The average smartphone has as many as 26 apps downloaded – however, fraudsters usually have far fewer.

 

 

  • Whether the phone is jailbroken or rooted

Jailbreaking an Apple device involves removing limitations that the manufacturer put in place, allowing fraudsters to use unapproved software and perform non-recommended tasks. Rooting is similar, permitting fraudsters (or other hackers) to bypass Android’s security architecture. Just because a device is jailbroken or rooted doesn’t necessarily mean that a fraudster is using it, but it’s a good clue.

 

 

  • Type of phone

Most consumers want the latest hardware available. But fraudsters generally use older and less sophisticated phones that are cheap and easy to discard if they need to.

 

 

  • App version used

In many cases, customers fall prey to fraud because they’ve failed to update the apps on their phones. Similarly, fraudsters rely on older versions of your app to exploit security loopholes.

 

 

  • User behavior and biometrics

Biometrics can be powerful clues when you’re on the hunt for fraudsters. Is the user swiping or typing? Are they typing rhythmically or erratically? This data might seem trivial or anecdotal, but it can actually help you figure out whether you’re dealing with a human or a bot. Humans prefer swiping to typing, while bots type more evenly and systematically than humans.

 

Look at the full picture to identify mobile fraud

While these data points are all good indicators that a fraudster might be interacting with your business, they’re clues – not guarantees. But how do you piece together this data to form a coherent fraud story? When it comes to building your mobile fraud-fighting strategy, make sure you’re taking advantage of a technology that can look beyond individual data points to identify patterns that point to fraud.

Machine learning-based fraud solutions can ingest data collected across all stages of the customer journey, irrespective of device. A user may download your app, but have a history of browsing your site and interacting with your brand. Or they may have previous interactions on other websites or apps that can be used to generate a profile of the user, even if they are unfamiliar to your company. It’s this combination of big data and technology that provides the key to successful fraud prevention via the mobile channel.


About the Author:

Kevin Lee, Trust and Safety Architect at Sift Science, builds high performing teams and systems to combat malicious behavior. He has lead various risk, chargeback, collections, spam and trust and safety initiatives for organizations including Facebook, Square and Google.