What 2 Men And a Runner Taught Me About Data Science and Fraud Modeling

What 2 Men And a Runner Taught Me About Data Science and Fraud ModelingPiyanka JainBlockedUnblockFollowFollowingFeb 20As I was wrapping up my work last evening, darkness started to loom outside.

I decided to go for a short run on a neighboring track, hoping to do eight rounds.

The run turned out to have a twist and taught me an interesting lesson.

On the track, a few folks could be seen strolling, and stretching.

I started my run.

While I was on my sixth round, I saw two large men strolling right on the tracks, blocking it.

I was at the far side of the track and no one could be seen around.

Although I conjured up some courage and passed the men, I felt a bit threatened.

Their presence made me feel the need to short-circuit my run and make an exit from the park.

But just about then I saw a guy with a dog I had passed earlier, standing beside the pull-up bars.

An impulse crept in, and I stopped beside the man to ask him if he would stay there for the next five minutes.

He replied with an affirmation but initiated an interrogation.

– “Why?” he asked.

As I tried to come up with an answer to explain my predicament, the two burly men passed by, and my eyes involuntarily followed them.

– I mumbled, “Because, eh….

”– “Okay.

I will,” he said, stopping me midway after having understood my anxiousness.

Receiving an assurance, I finished the last two tracks in peace.

I thanked the man with the dog and started my journey back home.

It was then that the stark similarity between the incident and a fraud model we were building for a client struck me.

Similar to my hunch back at the park, the statistical model tags every account with a probability of whether the person concerned is likely to commit a fraud or not.

And just like my hunch, which may have been a false alarm, fraud tagging too may or may not be absolutely correct.

Although it is true that fraud models built using statistical methods have much lower odds, they still have false positives — the model falsely tags somebody as a fraudster who may not be one.

To throw light on the matter, let us consider banks.

They use a classic way to deal with probable fraudsters: by keeping them out.

By doing so, they are saying ‘No’ to potential and good customers as well — a fact they are well acquainted with.

It leads to a catch-22 situation where to save losses, they get deprived of revenue too.

This is quite akin to me thinking of short-circuiting my run because of a possibility of a threat.

But what if the companies did what I finally decided to put into action — tag the fraudster, give them limited functionality (similar to the guy and his dog keeping an eye on me, thereby lowering the risk), and see if they actually turn out to be fraudsters or not.

This would ensure that ‘good’ people can keep coming in, while losses incurred due to fraud is reduced.

In fact, many banks and Fintech companies have now started to adopt this method of managing risks and frauds — giving all qualified customers limited functionality and access, and letting them prove their credit-worthiness or band worthy-ness in order to get additional functionality over time.

At Aryng, our data science SWAT team is always pushing the envelope on using data to boost revenue and reduce losses, while finding new ways of delighting end users.

If you need our help with fraud or other high-value customer analytics projects, please contact us for a FREE consult.

Thank you for reading my post.

I am passionate about using data to build better products and create amazing customer experiences.

I write about my learnings at LinkedIn and at Forbes.

To read my future posts, simply join my network here or click ‘Follow’ or join me on Twitter.

About Piyanka JainA highly-regarded industry thought leader in data analytics, Piyanka Jain is an internationally acclaimed best-selling author and a frequent keynote speaker on using data-driven decision-making for competitive advantage at both corporate leadership summit as well as business conferences.

At Aryng, she leads her SWAT Data Science team to solve complex business problems, develop organizational Data Literacy, and deliver rapid ROI using machine learning, deep learning, and AI.

Her client list includes companies like Google, Box, Here, Applied Materials, Abbott Labs, GE.

As a highly-regarded industry thought leader in analytics, she writes for publications including Forbes, Harvard Business Review, and InsideHR.


. More details

Leave a Reply