I'm thinking about a SaaS with pay-as-you-go billing based on amount of computational resources consumed by a user at given month.<p>However if my service charges customers at end of each month via recurrent credit card payment nobody prevents some of customers to use at registration credit cards without sufficient amount of money, so effectively such customer won't pay while my service should pay to cloud provided (e.g. AWS) because additional computational resources are already consumed by this customer.<p>Even if my service has quota for resources that can be consumed by new users that can be increased by connecting to support nothing prevents such malicious customer to have a few accounts (each of them is connected to new online virtual VISA card) and use my service from multiple accounts and then don't pay for used resources.<p>How common such problem is and how does small businesses deal with it? I think that it is not a problem for large companies or VC-backed startups, but such malicious customers can destroy a bootstrapped startup.
Is there a reason you can't have your users load their account with a certain amount and then have their usage charge against that pre-loaded amount? That's what they do at Twilio. As for fraud, there are a ton of 3rd party systems out there that can do fraud detection for you. Sift Science being one of them (<a href="https://siftscience.com/" rel="nofollow">https://siftscience.com/</a>).