It's a good observation, but I was trying to work out what this actually says. I read this as being fraud accounts for 0.1% of the value of all handled transac

It's a good observation, but I was trying to work out what this actually says. I read this as being fraud accounts for 0.1% of the value of all handled transactions. The transaction fees of a payment processor seem to vary between about 1% and 2.5%(although this may differ for CNP scenarios). Let's assume 2.5%, so fraud accounts for 4% (0.1%/2.5%) of the payment processor's fees. Intuitively, this doesn't feel too bad. However, I'm not actually sure what that says about how much the payment processor spends on _preventing_ fraud (and so how this impacts their margins). I think all we can really say is that they're likely to be at or near an equilibrium where if they aimed for less fraud it would have a negative impact on their business. This might not even be due to having to spend more money. For example, as you make your detection sensitivity higher, you're more likely to have false positives and reject valid transactions. In turn, this is likely to drive customers away from your payment processor towards your competitors. Maybe a certain level of fraud is just a cost of doing business? Just some thoughts...