In this edition:

• More data isn’t always better
• New content from Incognia: Fraud-as-a-Service and GenAI, promo abuse, user verification
  

More data isn’t always better

By André Ferraz, CEO and Co-Founder at Incognia

In fraud prevention, the problem is usually not a lack of data.

We have so many signals available to us, they’re coming out of our ears. 

But there are two common data problems I see in fraud prevention:

1. You can’t rely on the data you have because it’s not tamper-resistant 
2. You have lots of data, but it’s low quality data

Let’s start with the first one.

If you can’t rely on the data you have, you might not even know the scale of your fraud problems.

Your platform could be barreling towards an iceberg of different fraud types, but you would only see what’s poking above the surface:

The tampered data that gets reported back to you. 

Take device fingerprinting as an example.

If your device fingerprinting solution isn’t tamper-proof, fraudsters can use methods like emulators or app tampering to escape detection.

You might think that your device fingerprint provider is keeping your platform safe from multi-accounting. But without the right tampering detection, fraudsters could be making thousands of accounts right under your nose without you noticing.

If their solution doesn’t have strong tamper detection, the data you’re getting back from it isn’t trustworthy. 

Then you’ve got the second problem.

Too much low-quality data.

Some fraud vendors pitch you on the huge amount of data they can provide you.

But at the end of the day, that dump truck of data might not be that useful.

The balance of quality vs. quantity is critical when it comes to fraud prevention data.

You don’t need a million signals if they’re not actually helping you prevent fraud. 

So instead of how many different signals a vendor can provide, think about the uniqueness of their signals.

Are these signals something you could get somewhere else? Or replicate in-house relatively easily? 

A vendor’s goal should be to help you prevent fraud as efficiently as possible and streamline your operations, not to add to the noise. 

And ideally, you should look for vendors that collect unique data that you can’t get on your own.

Data is everything in modern fraud prevention, but not all data is created equally.

In the end, what data is most important for fraud prevention?

In my experience, the most important signals are related to accurately identifying the device and its location, along with network information.

So next time you’re thinking about new signals to add to your stack, ask yourself and the vendors: 

• Is it tamper-resistant and reliable? 
• Is it unique, a signal I can’t get anywhere else?
• Does it help me do more with less? 

You don’t need all the data in the world—only the data that counts.

To see more on this topic, check out my video:

Quality Over Quantity: Improving fraud prevention with high-precision data

New content from Incognia

We regularly share insightful content about fraud prevention on our website, on YouTube, and on LinkedIn. Here are a few recent highlights for you:

Arūnas Umbrasas from Vinted on the impact of GenAI and Fraud-as-a-Service on user verification:

Jaanus Uudmäe from Bolt on what makes a platform more vulnerable to promotion abuse:

Our recent webinar with Reverb and Vinted about seller and buyer verification in P2P marketplaces:

Incognia, a digital identity company, detects fake account creation and account takeover attempts for gig economy, marketplace, and financial technology applications. Benefits of using Incognia’s location-based digital identity include reduced false positives and a low friction user experience.