4 ms·

Consumer identity fraud isn't the only type of fraud problem that needs to be solved from the payment processor POV - there's also friendly fraud (when the cust

Consumer identity fraud isn't the only type of fraud problem that needs to be solved from the payment processor POV - there's also friendly fraud (when the customer is actually the owner of the credit card but claims fraud anyway), merchant fraud (merchants setting up bad sites and trying to steal funds), and collusion between merchant and customer. It seems like Bolt is focused on solving the consumer identity fraud problem for merchants, but this biz models will 100% make them a huge target for fraudulent merchants to collude with customers to steal funds. I guess I don't see how even an additional low single digit % fee will make up for false negatives. Assuming the company keeps .5% of the standard payment processing fee + takes an additional ~3% in fees on top of that, a $1000 false negative would require $1000/(.035) = ~$28.5k in additional processing volume to breakeven. This doesn't even take into the account the fact Bolt will eat the chargeback fee passed on from the network, so merchants with high volume/low average order value (think digital goods) will be hugely expensive for Bolt to service given they're making pennies per transaction but potentially paying 10x+ that per chargeback.

rbres 9y ago
Hey dawhizkid - totally right. Some comments on that in this thread: https://news.ycombinator.com/item?id=16216682 https://news.ycombinator.com/item?id=16216682 https://news.ycombinator.com/item?id=16216936 https://news.ycombinator.com/item?id=16216936 We vet our merchants before onboarding and make sure to work with high-integrity companies. We monitor orders not just for identity fraud, but for merchant / collusion fraud. That being said, we can certainly make mistakes. We also end up in the red some months with some clients. The good thing is that when we make a mistake, we pay the cost, not the merchant (which is contrary to the current state of the industry). We have made reasonable profit per client even with our costs, but the real winners are our clients driving millions in newfound revenue.
- dawhizkid 9y ago
  What's the source of truth for deciding whether it was friendly fraud or not? Is it the chargeback reason code? As far I know there isn't a specific reason code for "friendly fraud" (i.e. banks are passing along chargebacks telling you as a merchant that they think it's not actual fraud) and more often than not, especially having worked on this problem at scale in the past, friendly fraud chargebacks just come back with a "this is fraud" reason code.
  - AdieuToLogic 9y ago
    In my "prior life", I can say that processors provide generic reason codes in order to mask how the determination was reached. The idea being that it helps mitigate forms of phishing attacks.
AdieuToLogic 9y ago
Other insidious forms of fraud include, but are not limited to, "settlement attacks" and "refund attacks." Note that both do not require merchant knowledge or collusion.

rbres 9y ago

Hey dawhizkid - totally right. Some comments on that in this thread: https://news.ycombinator.com/item?id=16216682 https://news.ycombinator.com/item?id=16216682 https://news.ycombinator.com/item?id=16216936 https://news.ycombinator.com/item?id=16216936 We vet our merchants before onboarding and make sure to work with high-integrity companies. We monitor orders not just for identity fraud, but for merchant / collusion fraud. That being said, we can certainly make mistakes. We also end up in the red some months with some clients. The good thing is that when we make a mistake, we pay the cost, not the merchant (which is contrary to the current state of the industry). We have made reasonable profit per client even with our costs, but the real winners are our clients driving millions in newfound revenue.

dawhizkid 9y ago
What's the source of truth for deciding whether it was friendly fraud or not? Is it the chargeback reason code? As far I know there isn't a specific reason code for "friendly fraud" (i.e. banks are passing along chargebacks telling you as a merchant that they think it's not actual fraud) and more often than not, especially having worked on this problem at scale in the past, friendly fraud chargebacks just come back with a "this is fraud" reason code.
- AdieuToLogic 9y ago
  In my "prior life", I can say that processors provide generic reason codes in order to mask how the determination was reached. The idea being that it helps mitigate forms of phishing attacks.

AdieuToLogic 9y ago

Other insidious forms of fraud include, but are not limited to, "settlement attacks" and "refund attacks." Note that both do not require merchant knowledge or collusion.