The simple truth about scaling fraud management

Fraud detection doesn’t scale, fraud prevention does. It’s not complex, so let’s make it simple.


Risk in general and fraud risk as well, consists of 2 parts and is usually represented with a simple formula:

Risk = Chance x Impact

Fraud detection is all about reducing the impact of fraud. That means fraud already happened. You detect fraud, preferably as early as possible, so the financial 💸 and/or reputational damage is minimised.

Fraud prevention is all about reducing the chance of fraud. Fraud didn’t actually happen yet, but you are doing everything you can to prevent fraud. You want to know who or what you are dealing with before a final transaction happens (#kyc etc.)

Scaling Fraud management

Let’s say you want to grow as a business (that’s a choice by the way) and let’s say you want to double your revenue by doubling the amount or orders. If you keep your fraud management at the same level (both prevention and detection) the amount of fraud will also double. That means your fraud operations (detection) needs to double as well, including customer support, dealing with the police 👮‍♀️ etc, and so on. Furthermore, if you make your detection even better by innovating, you will probably find even more fraud. That means, at least temporarily, the damages grow 📈 even more percentage wise. So, although you can do fraud detection smarter, scaling is hard.

Fraud prevention does scale. When you put more innovation effort on how to prevent fraud from happening, you make sure the chance of fraud is lower. And as a consequence, the need for fraud detection lowers as well in the long run. The more you prevent, the less you need to detect. 

⚖️ We all know this is a 🐈 and 🐁 game, so you cannot only do one thing. Both prevention and detection need to be on par, especially because detection can give learnings for your prevention efforts. But, if you want to scale fraud management, double down on prevention.

#innovation #management #fraud #fraudprevention #frauddetection #kyc