Predicting whether a user will commit fraud is tough. Maybe just let them do it?

View in browser

The Signal by Incognia
NL---The-Signal,-Nov-13-2024

Just let them commit fraud

By André Ferraz, CEO and Co-Founder at Incognia

 

Predicting whether a user will commit fraud can be really tough.

 

You may be better off just letting them do it.

 

Why do I say that?

 

Here’s a strategy that could be more effective:

 

When you detect that a user has committed fraud, put all your focus on never letting them do it again.

 

I’ll share an example to illustrate.

 

Recently a ride-hailing platform shared with us that they had a problem with offline rides.

 

Offline rides happen when drivers and riders work together to cut the platform out of their share of profit for the ride.

 

For example, a rider might request a ride, and then once their driver arrives, cancel the ride in the app. They then make a verbal agreement with the driver in person to do the ride for cash instead. This allows the rider to pay less and the driver to earn more, but the platform doesn’t profit at all from the ride that it facilitated.

 

So this company knew their users were doing these offline rides, and they were asking us how they could use Incognia’s data to predict this kind of behavior so they could stop it before the ride happens.

 

We were frank with them: There’s not a good way to predict this kind of behavior.

 

Instead, we suggested this:

  • Monitor which drivers were frequently canceling rides, especially the ones who were canceling right when they arrived at the pickup spot
  • Put a policy in place that drivers engaging in this behavior will be banned
  • Implement a strong device ID and location intelligence solution like Incognia to ensure those banned drivers stay banned.

We told them that if they did those things, their offline rides problem would likely go away.

 

It’s a classic case of working smarter, not harder.

 

Some types of fraud are very difficult to predict before they happen.

 

You could spend a lot of time and money trying to do that, only to fail.

 

In some cases, you’re better off detecting when a user first commits that type of fraud (or noting when they break the policy).

 

And then putting all your focus and investment toward ensuring they can never do it again on your platform.

 

div_5

Other things you should check out

 

Blog

Stopping Ban Evasion When Fraudsters Get Creative With Obfuscation 

This post examines how fraudsters get away with ban evasion and the tactics fraud prevention can use to stop them.

 

Webinar Recording

Multi-Accounting: The Hidden Gateway to Marketplace Fraud

Experts from Just Eat Takeaway, Glovo and Incognia explore the risks of multi-accounting and how marketplace platforms can fight back.

 

Conference Presentation

Guardians of Growth: Fighting Promotion Abuse and Redefining Fraud Prevention

An expert panel from Swiggy, iFood, and Incognia discusses the delicate balance between preventing fraud and using promotions to boost growth.

      
Incognia Logo

Incognia is the innovator of next-generation identity solutions that enable secure and seamless digital experiences. For more information, visit Incognia.com.

Sign up for a demo →
LinkedIn
YouTube

Incognia, 333 West San Carlos Street, Suite 600, San Jose, CA

Unsubscribe Manage preferences